Transcript Slide 1

Applications to Agriculture and Other Sectors:
The Role of Weather Index Insurance
Dr. Jerry Skees
H.B. Price Professor, University of Kentucky, and
President, GlobalAgRisk, Inc.
US CLIVAR/NCAR ASP Researcher Colloquium
Statistical Assessment of Extreme Weather Phenomena under Climate Change
NCAR Foothills Lab, Boulder, Colorado, USA
June 13-17, 2011
1
Special Thanks to GlobalAgRisk Team!
GlobalAgRisk, Inc.

Mission
Improve access to financial services and the value chain for the rural poor through
innovative approaches for transferring weather risk
 Activities
 Research and development tied to
University of Kentucky research program
 Technical capacity building
 Educational outreach
 Supported by
 Select Country Work
 Peru – El Niño/Flood
 Mongolia – Livestock
 Vietnam – Flood/Drought
 Indonesia – Earthquakes
 Mali – Drought
 Multinational donors
 Morocco – Drought
 Governments
 Mexico – Drought
 Nongovernment organizations
 Romania – Drought
 Ethiopia – Drought
2
Current Support for GlobalAgRisk, Inc.
 Bill and Melinda Gates Foundation (Peru)
 Ford Foundation (Vietnam and Indonesia)
 Gov’t of Mongolia via Swiss Trust Fund
 UNDP (Peru)
 GiZ (Peru)
 Risk Management Agency of USDA
 Actuary and Underwriting Reviews
3
State of Knowledge Reports from GlobalAgRisk
Supported by Bill and Melinda Gates Foundation
Innovation in Catastrophic Weather Insurance to Improve
the Livelihoods of Rural Households
 March 2010
 “Data Requirements for the Design of Weather Index
Insurance.”
 March 2011
 “Market Development for Weather Index Insurance Key
Considerations for Sustainability and Scale Up.”
 Under Revision:
 “Legal Considerations for Index Insurance”
 Forthcoming a book
4
Somebody Always Pays for Catastrophic Risk
Who? How?
Society needs to understand the cost of natural disaster risk
Someone always pays
 The poor pay through direct losses and long-term economic
impacts
 Financial institutions restrict services as they learn that the
correlated losses of many of their borrowers and savers create
significant banking problems
 Governments — Disaster relief and recovery expenses,
infrastructure investments, subsidized agricultural insurance
 Donors forgive debt and divert funds for recovery
Need incentives for proper risk management and mitigation
5
Designing Sustainable and Scalable Weather
Index Insurance Programs Is Challenging
 Products must be developed in context
 Costly (technical support, capacity building, R&D)
 Not easily replicable
 Basis risk
 Tradeoff between transaction costs and basis risk
 Limited or no data to develop products
 High delivery costs
 Small transactions/Small market volume
 Nascent legal and regulatory systems
6
Lessons Learned
Risk Assessment
 Understanding vulnerability necessitates examining a specific
context:
 The risk, geography, demographics, risk management strategies,
government policies, etc.
 Risk assessment requires a systems approach
 Proper risk assessment benefits decision makers and can guide
policy decisions
 Cognitive failure – decision makers often underestimate extreme
risk of natural disasters
7
Adaptation
Multilateral Organizations Call for Insurance
 Decision makers suggest insurance can play an important role in
adaptation to climate change
 Article 4.8 of the United Nations Framework Convention on
Climate Change (UNFCCC)
 Article 3.14 of the Kyoto Protocol
 The Hyogo Framework
 Agricultural insurance, in particular, gains increased attention
because agriculture is the primary livelihood of households,
especially the poor, in developing countries
8
Insurance and Adaptation
 Insurers assess specific risks
 Insurers price the risk
– a critical aspect of risk assessment
 Insurance can encourage adaptation through pricing
– if insurance costs that much, maybe I should do something
different!
 Access and price of insurance can be directly linked to adaptation
strategies
 Fire insurance: extinguishers in the house?
 Proximity to a fire hydrant
 Improved building codes
 Insurance reduces variability in wealth/revenues
9
Pricing the Risk
“Fairness” and Perverse Incentives
 If insurance is not priced to reflect differences in relative risk (or
if more subsidies are paid to higher risk areas), perverse
incentives are likely to follow
– if you pay people to take risk, they will take more risk
 The result will be more risk exposure rather than less
 Higher risk activities
 Higher risk areas
 Less investment in risk reduction
10
In the U.S. Subsidies Have Been Used to Lower Loss Ratios
Coverage
Level
1980 Act
1994 Act
2000
ARPA
55
30.0
46.1
64.0
65
30.0
41.7
59.0
75
16.9
23.5
55.0
85
---
13.0
38.0
11
Increased Risk-Taking
 As a result of purchasing insurance, farmers tend to engage in





riskier behavior (plant higher risk crops, monoculture, less
irrigation, etc.)
Difficult for insurer to monitor these changes in behavior
Increased risk-taking increases loss ratios
High loss ratios lead to higher premium rates
Higher premium rates reduce insurance purchasing
Higher premium subsidies are then needed to maintain or
increase insurance purchasing
12
U.S. Agricultural Policy Has Also Favored More
Risk Areas
 If you pay people to take on more risk;
they will take on more risk!
 Program yields for commodity payments are based on planted
acre yields (not harvested acre yields); favored areas with large
abandoned acres
 Ad hoc disaster payments (long history)
 Highly subsidized crop insurance
 Premium subsidizes are a percent of premium: Once again
favoring high risk areas
13
Premium Subsidy per Dollar of Liability
2002
14
Loss Ratios Vary Across the Country
1981–2002
Change in Crop Share
Le g e n d
Hi ghe st Lo ss of C rop S ha re
M od erat e Lo ss o f S r op S h are
S m al l L os s or G ai n o f C ro p S h are
M od erat e G ai n o f C ro p S ha re
Hi ghe st G ain of C rop S ha re
M is s ing D ata
16
Farmer Knowledge and Decision Processes
 Literature shows
 Farmers optimize
 Farmers adapt
 Farmers are good Bayesians
 Farmers know central tendency on yields
 Cognitive failure sets in for catastrophic events
 Challenge for adaptation
 Communicating information about climate change in a
fashion that is useful for decision makers
17
Lessons Learned
Pricing the Risk
 Pricing risk is an important aspect of insurance
 Insurance can encourage households to reduce risks
 Government policies can motivate households to take more risk
at government’s expense
 Agricultural insurance is extremely complex
 Risk classification is difficult
 Risk varies by commodity
 Risk varies by region
 Risk varies by producer
 Difficult to monitor farmer behavior
 Administrative cost is very high
18
Weather Index Insurance: Gaining Popularity in
Developing Countries
 Payouts are based on a measurement of an event correlated
with losses
 Weather event (e.g., rainfall, temperature, or river levels associated
with yields)
 Insurance pays when index crosses a certain threshold
 Payouts do not require inspection of client losses
 Eliminates most information asymmetry problems
 Transfers extreme risk to reinsurers
19
Example of an Excess Rainfall Insurance
Product
 Extreme rainfall in India – payments would occur anytime
rainfall exceeds 2000 mm
 Household might buy US$100 in liability
0
500
1000
2000
3000
4000
20
Payout Structure for Example Excess Rainfall
Contract
Indemnity Payment (in USD 1)
120
100
80
60
40
20
0
0
250
500
750
1000
1250
1500
1750
2000
2250
2500
2750
3000
3250
3500
3750
Rainfall (mm)
 $1 for every 10 mm excess of 2000 mm
 $100 limit at 3000 mm
21
4000
Pricing Index Insurance
Price of
Index Insurance =
Pure Risk
+ CAT and Ambiguity Loads
+ Risk Financing
+ Delivery
+ Education/Marketing
+ R&D
+ Farm level/Asymmetric Information
Problems
 Adverse Selection
 Moral Hazard
 Loss Adjustment
22
Index Insurance Creates Flexibility for Different Users
 Households
 India: Drought coverage to groundnut farmers
 Malawi: Cooperative links insurance to farmer loans for high-yield
seed varieties
 Mongolia: Index-based livestock insurance
 Firms (e.g., Banks/input suppliers)
 Increase access to credit and other services
 Vietnam and Peru: Possible use of index insurance for protecting
credit risk
 Governments/Donors
 Mexico: State governments provide quick drought relief to farmers
in need using a drought index
23
Sahel: Shifts in the Central Tendency
Sahel data
1900 – 2007
800
700
600
500
Sahel
400
Semi-arid region below the Sahara
300
200
100
0
1900
1903
1906
1909
1912
1915
1918
1921
1924
1927
1930
1933
1936
1939
1942
1945
1948
1951
1954
1957
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
2005
• Dynamic climate largely due
to oceanic oscillations
• Unlike the Sahel, climate
change may lead to more
permanent changes
24
200
700
0
600
Insurer would
re-center data around
a current forecast of the
central tendency
1900
1903
1906
1909
1912
1915
1918
1921
1924
1927
1930
1933
1936
1939
1942
1945
1948
1951
1954
1957
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
2005
100
1900
1904
1908
1912
1916
1920
1924
1928
1932
1936
1940
1944
1948
1952
1956
1960
1964
1968
1972
1976
1980
1984
1988
1992
1996
2000
2004
Sahel: Hypothetical Insurance Product
800
700
600
500
Insurer entry point at
1962, 1990, and 2007 to
develop a drought insurance
400
300
800
500
400
300
200
100
0
25
Sahel: Rainfall Distribution Using Data from
1900 to 1961 (Reference Point – 1962)
 Insurance contract for wheat farmers
 Indemnities when rain is below 425 mm
 Central Tendency: 510 mm
 Payout Threshold: 425 mm
 Pure Risk: 2%
Pure risk will likely result in
insurance affordable to
households
26
Sahel: Rainfall Distribution Using Data from
1962 to 1989 (Reference Point – 1990)
 Central tendency is below payout threshold
 Insurance is inappropriate in this setting
 Central Tendency: 328 mm
 Payout Threshold: 425 mm
 Pure Risk: 44%
Pure risk would make
insurance unaffordable for
households
27
Sahel: Rainfall Distribution Using Data from
1990 to 2006 (Reference Point – 2007)
 Increases in rainfall reduced the weather risk
 Insurance would be affordable depending on other costs
(Delivery, ambiguity loading, etc.)
 Central Tendency: 456 mm
 Payout Threshold: 425 mm
 Pure Risk: 6%
Pure risk may result in
affordable insurance for
households
28
Climate Change Increases Ambiguity and
Catastrophe Loads
 Misestimating the central tendency is very costly
Green distribution Insurer forecast in 1962
Blue distribution
Forecasted loss experience based on 1990 Reference
Insurers expecting climate
change greatly increase
ambiguity and catastrophe
loads
29
Adaptation, Policy Interventions, and Climate Change
 Households must adapt or experience increasing farm losses
1. Change farming practices
2. Invest in infrastructure
3. Transition out of farming
 Insurance, by itself, is not a means of adaptation
 Policy interventions should assess the opportunity costs when
supporting insurance programs – other adaptation investments
may be more important
 Insurance can be used to facilitate adaptation
(e.g., linking insurance, credit, and improved seed varieties in Malawi)
30
Premium Subsidies Will likely Slow Adaptation
 Significant caution is needed for insurance premium
subsidies
 If climate change shifts the central tendency, premium
subsidies could slow adaptation
 Households account for subsidies when comparing profits
from farming and other activities
 These subsidies can encourage farmers to take more risk
and delay adapting to climate change
31
How to Improve Access to Catastrophic Weather
Insurance? Our Experience Suggests . . .
 Index insurance is best suited for catastrophic and consequential
losses
 Index insurance that addresses weather risk of firms that serve the
poor (risk aggregators) presents a feasible avenue for market
growth; build a sustainable market first and then move to micro
products
 Household products must find innovative delivery mechanisms to
improve product affordability and offer value to clients (insurancelinked products)
 Solutions that involve public-private partnerships must clearly
delineate the role for markets and the role for government
 Understanding cognitive failure for extreme risk can help
 Risk layering – Putting catastrophic insurance into a broader
conceptual framework
33
Risk Aggregator Products Are Less Costly to
Develop and Implement than Household Products
 Risk aggregator products face lower basis risk
 Risk aggregators effectively diversify much of the idiosyncratic
risks born by their clients
 Data constraints are less binding for risk aggregator products
 It is more cost effective for the insurer to establish a
partnership with a risk aggregator than to market and
distribute products to small holders
 Risk aggregators are more likely to understand hedging and
basis risk
34
El Niño Insurance for Flood
Innovation in Northern Peru
35
Piura and other areas in the North
Severely affected by 1998 El Niño
 Extreme rains (Jan – Apr 1998)
 40x normal rainfall
 Severe floods
 41x normal river volume
 Widespread losses
 Many disrupted markets
 Agricultural production, ↓ 1/3
 Public infrastructure losses
 Cash-flow, debt repayment problems
 Health problems
 Total losses in Piura estimated at USD 200
million
Contract is Written Using NOAA Data
 Nino estimates are derived from satellite data, observations of buoys and
readings of the temperature on the surface and at deeper levels.
 The data are publicly available monthly from NOAA (The U.S. National
Oceanic and Atmospheric Administration)
 http://www.cdc.noaa.gov/Correlation/nina1.data
Strong El Niño in 1982-83 and 1997-98
27
25
24
23
22
2 extreme
events in
the last 32
years
ENSO 1.2 Nov and Dec
26
21
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
1800
Precipitaciones Aeropuerta Piura
2000
1600
1400
1200
1000
800
600
400
200
0
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
38
37
31
Correlation Matrix for top 10 percentile El Niños
Correlation Matrix
South North ENSO1.2ENSO3
South
100% 90% 92% 89%
North
100% 90% 90%
ENSO1.2
100% 100%
ENSO3
100%
40
Example of a Payout from the 1997 Event
- Nino 1.2 (Nov-Dec) temperature = 26.28°C
Minimum payment = 5%
The insured selects the sum insured
Sum insured = 10,000,000 Soles
1998 payment = 76% x 10,000,000 = 760,000 Soles
Primary Goal:
Improve Access to and Terms of Loans
 Capacity building with
 Financial institutions
 Peruvian banking regulator
 Peruvian credit rating agencies
 Sources of social capital flows into Peruvian Institutions
Case to be made
1) Strengthen the resiliency of the financial institution
2) Financial institution can be ready to lend when the
community needs capital the most – post disaster
42
Risk Aggregator Strategy
Natural Disaster Effects on Banking
 Loan portfolio — Systemic repayment problems for borrowers,
problems can remain for years
 Deposits — Depositors withdraw funds
 Costs increase — Costs of funds (e.g., Interbank loans),
administrative costs
 Resulting problems
 Liquidity
 Profitability
 Capital Adequacy
 Lending institutions have many ways of managing these risks (e.g.,
Provisions, restructuring loans, etc.)
43
1997–1998 El Niño Spike and Recovery
10% Spike
3.5-Year Recovery
P
R
O
B
L
E
M
L
O
A
N
S
With this event every 1 in 15 years, 300 basis points must be added
44
Default Risk Significantly Affects Interest Rates!
  p 1  i L  1  r L
i

1 r
1
p
π – Expected profits
p – Exogenous probability of non-default
i – Interest rate
r – Lender’s opportunity costs
L – Amount of funds loaned
Example (No default risk)
r = 10%
p = 100%
Example (10% default risk)
r = 10%
p = 90%
El Nino may add 300 basis points to interest rates
45
45
Historical Pattern of Agricultural Lending in Piura
1994–2006
14%
Percent of Loans to Agriculture
12%
10%
8%
6%
Lenders say they have
“fixed the problem” by not making loans
when they see El Niño coming
4%
El Niño
2%
0%
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
Year
46
El Niño Reduces Capital Adequacy and the
Ability to Leverage and Make Profits
47
Comparison of Sum Insureds using Monte Carlo
% is the sum insured corresponding
to a percent of credit portfolio
e.g.,
Sum insured = 5% of credit portfolio
48
Risk Assessment Includes Evaluating Current
Risk Management Strategies
 Potential strategies for managing these risks and their costs
 Liquidity  Hold higher portion of assets in cash
 Effect — Reduces investment in productive assets
 Profitability  Avoid exposed regions and sectors
 Effect — Limits growth opportunities, especially for untapped
markets
 Capital adequacy  Leverage a lower amount of equity to
provide a “cushion” for the risk
 Effect — Limits growth
49
50
50
Mongolia — Massive Deaths of Animals
 Mongolia: 45 million animals at the start of 2010
 Sheep, goats, cattle and yak, horses, camel
 Value of animals = US $2 Billion
 Some 11 million animals were lost in 2001–2002 due to dzud
(harsh winter weather).. 9.7 million were lost in early 2010!
 Animal husbandry in Mongolia is 20+% of the GDP and over 85%
of all agriculture
 Census is done every year — Mortality data are available by soum
(county) from 1970 onwards
51
Index-based Livestock Insurance — Risk Layering
A New Model for Public-Private Partnerships
Government Catastrophe
Cover (GCC)
100% mortality
GCC — Social Insurance
A layer of very infrequent risk where
decision makers may have a cognitive
failure problem
Paid by government
using a World Bank
contingent loan
30% mortality
LRI — Commercial Insurance
Offered by private companies with
reinsurance from government and now a
global reinsurer
Livestock Risk
Insurance (LRI)
6% mortality
Retained by
Herders
If the government can’t continue to
pay for extreme losses, the
commercial layer can continue
52
Climate Drivers Matter: Arctic Oscillation
3
2
1
0
-1
-2
-3
2009-2010
22 percent
mortality - Feb
value = -4.2
-4
1944-45
33 percent
-5
1899
1909
1919
1929
1939
1949
1959
1969
1979
1989
1999
2009
Underwriting Matters
 Index insurance can indeed address many of the adverse
selection and moral hazard problems associated with
traditional forms of insurance
 However, sitting sales closing dates still maters and the
insured can adversely select using weather forecast
information!
54
Scale Matters
 None of our efforts will succeed unless we design products
that capture the attention of the regulator and the market
from the outset
 Start with biggest risk targeted to risk aggregators
 Rural lenders
 Value Chain
 Farmer Associations
Carefully move to micro products with concept of livelihoods
insurance for consequential losses suffered by small households;
challenges will remain for demand and delivery
55
Visit our Website for Papers, Projects and More
www.globalagrisk.com
56