Transcript Model schematic

Climate Prediction and
Agriculture
Lessons Learned and Future Challenges
from an
Agriculture Development Perspective
Jock Anderson
Why this outsider speaker?
 Queensland farmer/drought manager
 Decision analysis background
 Early interest in climate
 Risk management as a way of life
 Decades on agricultural development
 Impact assessment as major hobby
 Including contemporary IFPRI work
 Past endeavor on CGIAR, Bank-supported
research & extension
 Impact of impact studies?
Semantic Issues Persist
 Weather, Climate, Climate Change

Timescales critical but open to opinion

But let me commend the paper of Holger Meinke!
 “Forecast” covers many interpretations

Categoric vs Probabilistic

Concrete/specific vs descriptive
 Not that this is the only field with such
semantic issues, e.g., “Risk”
 Uncertainty and Climate Change

John Zillman, Warwick McKibbin, Aynsley Kellow
www.ASSA.edu.au Policy Paper #3
Ex Ante or Ex Post
A
(prior beliefs)
Receive Forecast
Signal
Realized Climate
Outcome
B
C
(update beliefs)
(observe outcome of event)
(also observe agent’s actions)
(model possible response)
Ex Ante
Ex Post
Modeled behavior
Measured behavior
Simulated Benefit
Realized Benefit
Measuring Forecast Value
 Information has value when it can
influence behavior
 It usually also has a cost
 So, whether it has +ve net value is an
empirical question
Evidence on this has been sparse in
this Workshop: should be key item!
Indeed, has CLIMAG been worthy?
One user-friendly Bayesian manual
COPING WITH RISK IN
AGRICULTURE
Second Edition
J. Brian Hardaker, Ruud B.M. Huirne,
Jock R. Anderson and Gudbrand Lien
CABI Publishing, Wallingford
2004
Forecasting in an Uncertain World
 Priors represent uncertainty held
before a forecast
 Forecast information is captured in
likelihood probabilities
 Posterior probabilities come from
combining these
 Such revision cycles can be treated
sequentially, dynamically
Towards an analytic approach

Take a multi-enterprise production function

Often estimated pragmatically, simplistically, badly

But if done right, provides a framework worthy of our
attention
Ag.
Output
Qt  f ( X t , Z t , K t ,U t )
Conventional
Inputs (e.g. land)
Unconventional Inputs Technical
knowledge (e.g.
(e.g. infrastructure)
R&D investment)
Uncontrollable
factors (e.g.
weather)
Mark’s Pragmatic Reduced Form
 Relationship tying farm profits (P) to
climate information (K) and other onfarm characteristics
Pt  f ( X t , Zt , Kt ,Ut )
Conventional
Inputs (e.g. land)
Unconventional Inputs Climate Information
(knowledge sources)
(e.g. infrastructure)
Uncontrollable
factors (e.g.
weather)
Behavioral Factors
 Representing preferences is a possibly
significant challenge…Risk-averse?
 Ability of farmers to adjust should be
accounted: Representing constraints?
 Farmers and others are all swimming in
the stormy seas of risk, with and without
formal climate forecasts… Are such
forecasts a marginal part of the picture?
All easier said than done
 Estimation is “demanding”
 Of conceptualization, incl dynamics &
participatory insights
 Of data, especially in LDCs
 Of “estimational” /“modeling” skills
 Of optimization skills
 Of interpretation skills
Challenges of Assessment
 Many challenges, even if one can borrow
or adapt existing models, such as the
now-popular crop growth models
 Recall Mark noting that “Dis-entangling
the underlying structural relationships is
non-trivial”!
 So, much research, intrinsically multidisciplinary, is seemingly needed
Ex Post Assessment in Ag Research
 Mark spoke on this extensive (competitor)
literature…and I can not get into it here,
except to raise it as a “problem”
 But some of the research products that
will have potentially high payoffs in
responding to climate predictions present
new evaluation tasks (e.g., short-cycle
varieties that can “escape” or better
“endure” some droughts)
Wider Cogent Research Themes
 Understanding the mechanisms diverse
rural communities use for
 Managing risk e.g., borrowing, selling,..
 Coping with risk e.g., calling on rellies
 Shifting from risk e.g., migrating
 Agro-meteorologists may not have spent
much time grappling with rural financial
systems, futures markets etc. but maybe
they will have to? Or work more with….
Some Policy Dimensions
 A few selective aspects of farmer risk
management to illustrate a widened agenda
 Property rights (especially land)
 Other enabling aspects such as PSD (incl
index insurance), investment climate,
 Emergency policy and intervention history,
safety net processes, etc.
 Climate policy? Informed by climate
research? Understanding & prediction!
Risk transfer for
market premium
Reinsurance and Capital markets
GIIF
EC Co-financing
to cover
Transaction
Costs
Cofinances
premium
Government
pays true risk
cost Premium
(Re)insurance
contract based on
risk Index
Government
Bank
Primary
Insurer
Borrowers
Payout
according to
index trigger