Use of economic modelling in agricultural policy
Wolfgang Britz, Peter Witzke
Institute for Food and Resource Economics, University Bonn
Context
‰
‰
‰
‰
‰
‰
‰
Types of modelling applications
Types of models used
Institutional solutions to host models
Overview on CAPRI
Examples for applications
D results
Do
lt matter
tt for
f policy?
li ?
Summary and conclusion
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Application types
‰
Scoping: What are possible future developments?
(e.g. study “SCENAR 2020” or IPCC Climate Change scenarios)
‰
E
Ex-ante
i
impact
assessment off planned
l
d policy
li changes,
h
mandatory for EU Commission if larger budget impact
(e.g. recent legal proposal for Common Agricultural Policy - CAP 2014-2020)
‰
Ex-post evaluation of past policies, in the meantime mandatory
(e.g. regular tenders by DG-AGRI)
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Types of models for policy analysis (1)
‰
Focus here on Europe and Common Agricultural Policy (CAP)
‰ CAP changes typically provoke impacts
(at least potentially)
• on prices => equilibrium models [supply=demand=f(prices)]
• on trade =>
> global models
• across EU, but not uniformly => at least detail at country level
⇒ Market entry costs for new models are quite high
(data and parameter requirements due to EU/global coverage, IT
infrastructure)
⇒ Relative small set of models available, not much movement in
market …
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Types of models for policy analysis (2)
‰
Equilibrium models:
• Single market models
– For prices, quantities, policy in single market (or a few)
– e.g. Banana model (G Anania), Dairy model (V Requillart) ..
• Multi
Multi-Commodity
Commodity Models
– Equations for many agricultural markets
– e.g. “AGMEMOD”, “AGLINK”, “CAPRI”, “ESIM”, “GLOBIOM”
• Computable General Equilibrium (CGE) models
– for all markets in whole economy plus budget constraints
– e.g. “GTAP” variants like “MAGNET”, “MIRAGE”, “GLOBE”,
regional CGEs in CAPRI
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Types of models for policy analysis (3)
‰
Equilibrium models possible linked to
• more dis-aggregated supply models
(
(can
b
be iin-built
b ilt as iin CAPRI)
CAPRI), “A
“Aropaj”,
j” “RAUMIS”
“RAUMIS”, “FARMIS” …
• bio-physical models/components
e.g. Dutch model “INTEGRATOR” for nitrogenous emissions,
post model
d l lilink
k off CAPRI to “DNDC“
• land use models:
– CLUE-S (P Verbourg)
– GLOBIOM (P Havlik)
• external studies used for baseline
– CAPRI uses several sources ((AGLINK, GLOBIOM...))
– Large agencies often combine several inhouse models (IFPRI,
FAO, LEI, EU Commission)
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Institutional solutions
‰
Tool may be hosted in
• in administration or government sponsored agency
– ESIM model used by DG-AGRI (in the past)
– Models used by ERS (USA), JRC (Europe), INRA (France), TI (Germany)
• in independent institution,
– long-term contracts or
– Case by case tenders
‰ Viewpoints:
p
• Development versus maintenance
• Control over results
• Response
R
ti
time / ttransaction
ti costs
t
• Synergies / antagonism if several models
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
CAPRI overview
‰
Developed since 1997, operational and maintained since 2000,
financed to a larger extent by EU framework programs
‰ Comparative-static (i.e. no description of path of adjustment),
deterministic
‰ Consistent combination of different sub-models/components:
sub models/components:
• Global trade model
• European supply models
• Regional
R i
l CGE
CGEs
• Post-model processing (environmental indicators, welfare analysis,
CAP budget, energy use indicator)
• Spatial down-scaling to 1x1 km grid cells and link to bio-physical
models
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
CAPRI overview (2)
•
•
Global: 40 trade
blocks and ca. 80
countries
Spatial: “Armington”
assumption
Price
( > goods
(=>
d diff
differ by origin)
i i )
•
Detailed trade policy
S
Supply + Feed demand
l +F dd
d
•
•
•
Non-linear programming models for
280 regions or 1.900 farm types
Profit maximization under constraints
=> detailed policy representation
(premiums, quotas…)
•
… and links to environment
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
CAPRI overview
‰
Data from Eurostat, FAOstat, EU Commission (FADN) …
‰ Technical solution:
• Data transformations and simulation in GAMS
• User interface in Java (steering, tables/graphs/maps)
• Server based software versioning system (SVN) for exchange of
data and computer code in network
‰ Institutional
Instit tional model
• Open source/access, …
• Network of users/developers across Europe, including EU
Commission services
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Examples of CAPRI applications
‰
Research project applications:
• Selection mainly to give a good illustration of model development
• Relevant example may raise awareness of potential users
• But no influence on scenario design: leave or take option for policy
‰ Tendered applications:
• First one for CAPRI 6 years after development started and 3 years
after
ft first
fi t operational
ti
l version
i
• CAP Mid Term Review, sugar market reform, some contributions to
ex post evaluations ..
• Recent years:
– Ongoing CAP reforms (premium, market organizations)
– Free-trade agreements
– Agri-Environmental interactions
– Study on ban of GMOs (genetically modified organisms) by LEI
– Climate related studies
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Example: SCENAR 2020 scoping study
‰
Scenario study => no immediate policy under investigation, only broad
options for EU policy:
• “Baseline”, “Regionalisation”, “Liberalisation”
• Research environment better than administration for scenario
approach
‰ Modelling suite
• LEITAP
LEITAP-IMAGE
IMAGE (global CGE),
CGE) ESIM (ag baseline,
baseline biofuels),
biofuels) CAPRI
(CAP, NUTS2), CLUE-S (complete land use model)
‰ Plus statistical analysis, cluster analysis, SWOT analysis for regions
‰ Results
R
lt ffrom modelling
d lli chain
h i att national
ti
l llevell b
butt rurall problems
bl
(beyond agriculture) need further analysis
• Picked up
p in p
project
j
CAPRI-RD with regional
g
CGEs
• Other methodology also picked up in CAPRI trunk
– Land supply curve
– Implementation
p
of Rural Development
p
p
policies
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Example: OECD 2011 expost study
‰
Tendered by OECD secretariat to contribute to study MORREDU ET.AL.
Evaluation of Agricultural Policy Reforms in the European Union.
OECD,, Paris,, 2011
‰ CAPRI selected as it offered
• rich detail in describing the CAP pillar I instruments
(premiums and market price support)
• full EU coverage
• farm type and regional dis-aggregation
(not available with OECD in-house tools)
‰ What has been the effect of decoupling?
• Overall limited effects on allocation
• But transfer efficiency increased
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Example: DEFRA study on drastic reform
‰
Background:
• UK Department for Environment, Food and Rural Affairs (DEFRA)
in favor of drastic reform for CAP 2013-2020
2013 2020
• But what about land abandonment in marginal areas?
‰ Combination of CAPRI and Land Use Cover Change Model (CLUE-S)
used,
d CAPRI expanded
d d tto simulate
i l t shrinking
h i ki and
d expansion
i off
agricultural area
‰ Main findings:
g
• At country level reduction in land use between 6 (UK) and 15%
(Greece), larger effects for marginal areas
• Price increases for outputs partially offset loss of subsidies
• Environmental impacts overall reduced, but higher intensity on
remaining agricultural land
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Example: JRC study on Mercosur issue
‰
Recent study by EU Joint Research Centre on Free Trade Agreement
with Mercosur
• Combined
C bi d CGE (GLOBE) and
d multi-commodity
lti
dit model
d l (CAPRI)
application
• Explicit representation of trade policies in CAPRI permits realistic
analysis
l i
• Quite detailed scenario definition (tariffs and tariff rate quotas,
sensitive products…)
• Equally interest in regional effects, e.g. will freer trade in beef with
South-America lead to heavy income losses in mountainous
regions specialized in suckler cows?
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Example: MIRAGE-BIOF analysis of ILUC
‰
Background
g
and motivation:
• Searchinger et al 2008, Science: Pay back period of ethanol from
US maize is 167 years, because indirect land use change (ILUC)
effects release huge quantities of carbon
• Stimulated criticism and numerous studies
• But also marks turning point in biofuel enthusiasm
‰ MIRAGE study
d ffor DG Trade
T d (2009-11)
(2009 11) studies
di ILUC effects
ff
off EU
Renewable Energy Directive (“RED“, 10% target for transport fuel)
• ILUC effects had to be investigated
g
according
g to RED
• With several delays finally presented Nov 2011
• State of the art CGE model MIRAGE:
– Endogenous yields,
yields by-products
by products, disaggregated oilseeds
oilseeds,
substuitution of cropland, pasture, forest, income elasticities
updated on baseline, Monte Carlo for parameters...
• => ILUC factors for policy package and by crop (gCO2/MJ)
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Do results matter? (1)
‰
Weak empirical evidence (no systematic with and without observations,
interviews likely to provoke strategic answers …)
‰ Need to look at historical examples
examples, #1: CAP reform process (from price
to coupled to decoupled support) as of 1992
• Standard welfare analysis did back up Commission position
– But
B t this
thi argumentt was already
l d made
d ffor d
decades
d b
by JJosling
li
1973, Koester/Tangermann 1977
• Perhaps more relevant: stabiliser policy unable to meet income
goals
l
• Inviting to use economic model results as a justification for reform
proposals
• But sometimes apparently studies are effective: Choice of Ireland
for complete decoupling in view of farm income benefits
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Do results matter? (2)
‰
Example #2 Sugar market reform 2006
• Several preceeding studies have demonstrated the inefficiency of
the Sugar Common Market Organisation and the need for reform
• Why 2006 after decades of stability?
– Everything But Arms agreement
– WTO panell decision
d i i 2005 ((no C
C-sugar exports),
t )
• But unclear whether Commission has defended its sugar policy as
good as possible in panel
– perhaps
h
already
l d convinced
i
db
by need
d ffor reform?
f
?
• But choice of reduced intervention price clearly informed by earlier
studies
• Could be a general phenomenon: quantitative analysis for choice of
details in implementation
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Do results matter? (3)
‰
Example #3: Increasing weight for environmental issues in CAP :
• Several criticisms of the CAP from academia and non
nongovernmental organisations
• Therefore cross-compliance and decoupling?
• Could also be political vehicle to rescue CAP support
• Some econ arguments are being picked up
– Details of crop diversity requirements
‰ Example #4: Any effect of ILUC studies on policy?
• Yes for Searchinger et al 2007!
• Unclear for MIRAGE study in view of Commission Impact
Assessment (Oct 2012):
– Option D (ILUC Factors) discarded because
g investments in bidiesel industry
y
ƒ Devalution of existing
ƒ RED target of 10% in danger (and a past EU decision cannot be
wrong)
– Preferred option E: Cap for conventional biofuels in meeting
RED target
t
t att 50% => safeguard
f
d ffor pastt investmnts
i
t t (and past
policies)
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
Conclusions (1)
‰
Overall: direct influence of economic modelling
g is open
p or unclear
• Probably highest for fine-tuning of policies
‰ But indirect effect through training of economic thinking for staff
concerned
d with
ith policy
li d
design
i
‰ Modelling may serve to structure the debate on policy issues
• Example:
p Impacts
p
of rural development
p
support
pp
• To quantify regional CGE layer of CAPRI (for example)
• Forces to interpret each measure in terms of the rigorous model
framework
Britz/Witzke : Use of economic modelling in agricultural policy, 18.04.2013, Krško
References for examples
‰
‰
‰
‰
‰
NOWICKI, P., H. VAN MEIJL, A. KNIERIM, M. BANSE, J. HELMING, O. MARGRAF,
B. MATZDORF. R. MNATSAKANIAN, M. REUTTER, I. TERLUIN, K. OVERMARS, D.
VERHOOG
OOG, C.
C WEEGER
G , H.
H WESTHOEK
S O : Scenar
S
2020 - Scenario
S
i study
t d on
agriculture and the rural world, Contract No. 30 - CE - 0040087/00-08.
European Commission, Directorate-General Agriculture and Rural
p
, Brussels.
Development,
RENWICK, A., JANSSON, T., VERBURG, P. H., REVOREDO-GIHA, C., BRITZ, W.,
GOCHT, A. AND MCCRACKEN, D.,: Policy reform and agricultural land
abandonment in the EU, Land Use Policy, 30 (2013) 446 - 457,
doi:10 1016/j landusepol 2012 04 005
doi:10.1016/j.landusepol.2012.04.005
MORREDU, C.(MAIN AUTHOR AND EDITOR), WITH CONTRIBUTIONS BY MARTINI,
R., KIMURA S., BRITZ, W., GOCHT, A., PEREZ, I., HART, K. AND BALDOCK, D.:
a uat o of
o Agricultural
g cu tu a Policy
o cy Reforms
e o s in tthe
e European
u opea U
Union.OECD,
o O C ,
Evaluation
Paris, 2011
A. BURRELL, E. FERRARI, A. GONZÁLEZ MELLADO, M. HIMICS, J. MICHALEK,
S. SHRESTHA AND B. VAN DOORSLAER (AUTHORS), A. BURRELL (EDITOR):
Potential EU
EU-Mercosur
Mercos r Free Trade Agreement
Agreement: Impact Assessment.
Assessment JRC
Reference Reports, EUR 25011 EN, Luxembourg, 2011
LABORDE D (2011), Assessing the Land Use Change Consequences of
European Biofuel Policies. ATLASS Consortium, 111 p.,October 2011.
Britz: IMAP meeting march 2011 - long term projections with CAPRI