General Equilibrium Modelling of Trade
Negotiations Outcomes
Stephen N. Karingi, Trade and International
Negotiations Section, UNECA
Introduction
 Policy makers need options in order to be able to
make decisions.
 These options must be derived in an objective
manner, in order to give decision makers
confidence.
 If options are not derived in an objective and
consistent way, technicians easily lose credibility.
Introduction
 Policy makers get comfort if they know:
 That the methodology used to draw policy options is
theoretically sound.
 That the data used in generating the empirical results is
verifiable.
 That the assumptions used are realistic.
 Different scenarios are offered.
Why are economic
models needed?
Why are economic models needed?
 Different policy questions:
1. What are prospective export markets?
2. What sectors are sensitive in a liberalisation scheme/schedule?
3. Is a country better off with a REC remaining as an FTA rather
than customs union?
4. Has regional integration realised expected trade growth?
5. What are the potential benefits/costs from the EPAs and/or
Doha Round for a country or sector?
6. How long will it take for incomes in a given REC to converge?
7. Does trade liberalisation help reduce poverty – MDG target?
8. Are trade policy options gender neutral?
9. What is the optimal point for trade and environment policies?
Why are economic models needed?
 “Without theory, practice is but routine born out of habit.” Louise
Pasteur.
 “(S)He who loves practice without theory is like the sailor who boards
ship without a rudder and compass and never knows where he may
cast.” Leonardo da Vinci.
 “Being denied a sufficiently secure experimental base, economic
theory has to adhere to the rules of logical discourse and must
renounce the facility of internal inconsistency. A deductive structure
that tolerates a contradiction, does so under the penalty of being
useless, since any statement can be derived flawlessly and
immediately from that contraction. In its mathematical form,
economic theory is open to an efficient scrutiny of logical errors.”
Gerard Debreu, Nobel Prize Winner, 1983.
How do economic models help
improve policymaking?





They provide a theoretically consistent framework
for analyzing trade policy questions.
They provide a handle on complicated questions.
They help give greater intellectual support for a
chosen trade policy.
The use of models can provide a common
“language” for policy discourse or debate.
But models should complement rather than
substitute for policy making.
Models commonly used
for trade policy analysis
Models used for trade policy
analysis
 Trade flow analysis at HS-6 or more tariff line
analysis.
 Econometrics methods e.g. gravity modeling.
 Partial equilibrium modeling e.g. WITS-SMART;
ATPSM.
 Computable general equilibrium model (single
country).
 CGE model (global) e.g. GTAP.
 CGE models combined with micro-simulation
model.
Models used for trade policy analysis
 Simulation models: they help answer “What if” types of questions (+
projections):
 Partial Equilibrium Models e.g. WITS-SMART; ATPSM; TASTE.
 General Equilibrium models e.g. GTAP; MIRAGE; LINKAGE.
 Econometric Models :
 gravity models (reduced form): can be used to establish whether
certain economic variables have an effect on a variable of interest
(Do RTAs or GSP increase trade?)
 Macro-econometric models: tools for projections of aggregates but
no information on the industrial structure of the economy + may
lack micro-foundations.
 Combination of simulation and econometric models.
 Simulation (econometric) models are deterministic (stochastic)
Partial equilibrium framework
Price
Impact of commodity x
market on rest of the
economy can be
neglected
DS
Pw(1+t)
Pw
DD
Commodity x
A General Equilibrium Analysis
spending on goods and services
Savings
Investments
goods and services
Households
Firms
Factor services of production
Factor incomes
REST OF THE WORLD
exports
imports
FDI
When is GE or PE analysis
appropriate?
 Nature of policy change
 Does it cut across many markets or sectors?
 Potential impact of change
 Are there economy-wide impacts?
 Constraints imposed by availability of data and resources:
 PE data and models: free
 CGE data: single country (SAM) could be free, multiple
country (GTAP: from $ 360 to $ 4600)
 CGE models: free (GTAP) but may need software for
mathematical programming to run (LINKAGE, MIRAGE)
A note about CGE and micro-simulation
models
 Top-down CGE models linked to micro-simulation. The
micro-simulation component based on household survey
data.
 The micro-simulation component tends to focus on
some of the following issues:
 Poverty incidence (headcount ratios; poverty depth etc.);
 Inequality measures (Lorenz curves on whose basis Gini
coefficients are derived).
 The micro-simulation allows the endogenous results of
the CGE model to be fed exogenously to poverty and
inequality equations.
Basics of CGE
Modelling
Typologies of CGE Modeling
Static: regions, sectors, factors,
economic agents
+ set of economic behaviors &
relationships
Micro-Simulation Models:
representative
agents hypothesis
“removed”
Dynamic=Static features
+ explicit inter-temporal features
The circular flow model of
economic activity
CGE models: basics
 Computable  numerical solution (empirical
data)
 General  description of the whole economy
 full economic cycle
 all markets
 Equilibrium  demand equals supply
 prices are adjusted to achieve market equilibrium
 general: on all markets simultaneously!
 Model  solvable set of equations
Partial equilibrium: an example

1
2
 objective:
U  C1 C
 market clearance:
Yi  Ci

i  1, 2
,
i
1i
i
 production:
Yi  Ai Li K
 resource constraints:
L1  L2  L
K1  K 2  K
max
General equilibrium: an example

1
2
 objective:
U  C1 C
 market clearance:
Yi  Ci

i  1, 2
,
i
max
1i
i
 production:
Yi  Ai Li K
 resource constraints:
L1  L2  L
K1  K 2  K
 income balance:
p1C1  p2C2  wL  rK
Empirical calibration
 Spend time on finding and organising your data:
 The quality of the results depend on the quality
of the data
 Different sources can not always be quickly
combined
 Always carry out the benchmark check!
 Much data available on internet, e.g. Statistical
Offices;
search and you will find
CGE models: data (I)
 Benchmark data
 base year
 (Social) Accounting Matrix, based on IO-table
 often: values in the IO-table interpreted as quantities
 all benchmark prices are 1
 Essentials of IO-table:
 value of output equals total value of all inputs for each good
 value of supply equals value of total demand for each good
 total value of endowments equals total value of final
expenditures (consumption)
  row total equals column total
A basic input-output table
Agr.
Indus. Serv.
Cons.
Inv.
Gov.
Exp
Total
Agr.
10
30
40
100
0
40
180
400
Indus.
20
60
50
300
200
30
340
1000
Serv.
40
60
10
200
0
20
170
500
Imp
60
120
20
400
100
100
Dep.
10
20
10
40
Wages
200 600
300
1100
Profits
60
70
240
Total
400 1000 500
110
1000 300
190
800
690
CGE models: data (II)
 Data describing the reactions of agents
 often described in terms of CES functions:
 substitution elasticities
 income elasticities
 output elasticities
 Cobb-Douglas, linear and Leontief
functions are special cases of a CES
function
Data and CES function calibration
 Elasticities, benchmark quantities and prices
determine the CES functions (technologies or
preferences)
(i) benchmark demand quantities
 provide an anchor point for isoquants /
indifference curves
(ii) benchmark relative prices
 fix the slope of the curve at that point
(iii) elasticity of substitution
 describes the curvature of the indifference curve
CES function calibration
L
C
L0
0
K0
K
Input-Output economics & SAMs
 Whether neoclassical, structuralist, neo-Keynesian,
or Monetarist, a CGE modeler must respect
accounting identities and equilibrium conditions.
Hence, most applied work is based on a social
accounting matrix to benchmark (calibrate) a model
and to represent relevant accounting identities.
 SAMs capture equilibrium conditions
 Walras’ law applies
Decision Making and Institutions
 Linkages in SAMs are accounted for by modelling
the decision-making process of the firm, the
consumer, as well as other economic agents and
institutions: production and demand structure
 Trade results from that decision-making processes
and their interaction with institutions:
 Production- Exports + Imports=Consumption
Closing the Model
 Need to define a numéraire (Walras law allows to
“drop” one market)
 Assumption about the adjustment mechanism in
factor and commodity markets
 Macro closure
 Macro accounting balance (govt expenditure and deficit;
aggregate saving and investment; balance of trade and real- exchange rate)
 Macro adjustment mechanism (exogenously determined)
Closing the Model
 Johansen closure: investment is exogenous and
consumption is the adjustment variable
 Keynesian closure: nominal wage is fixed and
employment is the adjustment variable (unemployment)
 Kaldorian closure: wages could be less or equal to the
marginal product of labor (exploitation of labor model)
 Classical closure: prices and wages are the adjustment
variables (constant employment) and investment becomes
endogenous and adjusts to total savings available
 Foreign borrowing (Robinson): trade balance is
endogenous, current account and hence net capital
inflows are the adjustment variable
Beyond the Standard Model
 Economies of scale, monopolistic competition and
differentiated products
 Institutional features of a particular economy (e.g. tax
collection costs)
 Specific features of a policy instrument
 Increase effort on estimation of substitution elasticities
 Dynamics to account for dynamic aspects (policy
credibility; capital accumulation; FDI; knowledge
accumulation and spillovers) and adjustment
 Account for the extensive margin of trade (the “smallshares” issue)
CGE Dynamic Models
 Recursive:
 solves annually
 Current economic conditions (e.g. the availability
of capital) are dependent on past outcomes but are
unaffected by forward looking expectations
 Linked with a macro econometric to include
exogenously projected changes in demographic
trends or in technology: baseline scenario
 Impact of policy change is given with respect to
the baseline scenario (sector specific TFP and
real GDP growth are solved endogenously)
CGE Dynamic Models
 Forward looking:
 Infinite lived consumer with financial market
 No extensive baseline scenario: trade
performance-productivity linkage + gvt
investment on infrastructure and TFP linkage +
investment in education and labor productivity
linkage
 Could account for transitionary disequilibrium
states (true adjustment process?)
Micro-Macro Models
 Combination of a Micro Simulation model (based on
Household surveys: fiscal and labor) and a CGE model
 Ideal to assess the impact of macroeconomic (trade) policies
and shocks on poverty/ inequality: MAMS (maquette for
MDG simulation)
 Two types of combination:
 Fully-integrated: the household model built directly into
the CGE : CGE model with heterogeneous agents (high
complexity)
 Sequential (top-down): CGE simulation results are
passed on to an household model (macro and micro need
not to be reconciled but possible lack of coherence)
What is involved in a
policy simulation?
What is involved in a policy
simulation?
Economy before
Policy
trade policy change change
Economy after
trade policy change
Difference between the two is
attributed to policy change
What is needed for a policy
simulation?
Inputs
MODEL / Closure
Outputs
What are the inputs?
 Baseline data:
 trade flows
 levels of protection
 input-output structure: national income aggregates
 Measure of responsiveness of economic agents to
price changes (i.e. elasticities)
 Policy - negotiating scenario
 Sectors (Agriculture, NAMA, etc.)
 Depth of liberalization
What are the Outputs?
 Configuration of the economy after policy change
 Overall income gains/losses from policy change
 Sources of income gain
 Sectoral (agriculture vs. NAMA)
 Policy instrument (market access or domestic support)
 Winners or losers (at the country level)
 Changes in pattern and volume of trade and income
 “Story” to explain how inputs and model combine to
determine the output/outcome
Tracing Differences in Results
 Deterministic – outcome is completely determined by
choice of inputs and model (no “residuals”)
Inputs +
MODEL
Outputs
 Differences in simulation results = differences in choice of
inputs and model/closure
“Story” must explain why the choice of inputs and model is
appropriate/optimal for the policy question of interest
Towards an “objective”
look at trade
liberalization CGE
simulations
Case 1: Global CGE Application on the likely
development impacts of EPAs
What to expect given the provisions on
market access in the interim EPAs
Possible result in SSA incomes for lack
of sufficient asymmetry
GDP Volume (% change)
Asymmetry (40%)
Asymmetry (20%)
-0.12
-0.1
-0.08
-0.06
-0.04
-0.02
% change from baseline
0
0.02
0.04
0.06
Due to focus on trade-rules, welfare
gains for Africa limited
Welfare (million US$)
800
Asymmetry (40%)
700
600
500
400
300
Asymmetry (20%)
200
100
0
Africa to specialise in primary products,
hindering diversification
S S A to specialise in primary commodities
% change in value added from baseline
12.0%
10.0%
8.0%
6.0%
4.0%
2.0%
0.0%
Cereals
Vegetables
Oilseeds
Sugar
Cotton
-2.0%
-4.0%
Asymmetry (20%)
Asymmetry (40%)
oCrops
Livestock
Case for industry modernisation under
the circumstances
Manufacturing sector a concern
Other manufacturing
Light manufacturing
Agro-processing
Natural resources
-15.0%
-10.0%
-5.0%
FTA
Asymmetry (20%)
0.0%
Asymmetry (40%)
5.0%
10.0%
15.0%
Regional integration as Africa’s
development pillar under strain
 Regional integration is to be a
pillar of EPAs.
Trade diversion (US$
million)
 Intra-regional trade shrink
significantly by as much as
18%.
FTA
 Situation likely to be
worsened by kind of
specialization to occur.
-787 (2619)
Asymmetry -532 ((20%)
1752)
Asymmetry -415 ((40%)
1043)
Social & economic expenditures strain
likely due to fiscal loss
Fiscal losses due to EPAs (US$ million)
3,529
2,103
1,038
FTA
Asymmetry (20%)
Asymmetry (40%)
Case 2: Global CGE Application on the Doha Round
Potential implications for African
countries of agriculture negotiations
WTO: Positions and Prospects
US moves on Ag. domestic support
(12 – 18 billions US$ OTDS ?)
Lamy’s Triangle
EU moves on ag. tariff cuts
(G20: 54% cuts)
Advanced DC
move on NAMA
(Swiss 20 ?)
Focus on Domestic Support,
Market Access and NAMA
The Market Access Pillar
 The negotiations here are about tariff reduction formula –
how ambitious?




Tariff peaks dealt with if ambitious.
Preference erosion occurs if too ambitious.
Policy space for African countries could also be lost.
A key issue is to find a balance between enhanced
market access; flexibilities and policy space.
 Sensitive and special products provide the flexibilities.
 Special safeguard mechanisms
Tariff Structure of the European Union (AVE)
300%
Processed
Cereal Grains
Bovine and Pork
Meat
Mushrooms
Garlic
250%
Bananas
Dairies
Sugar
200%
Prepared
Vegetables
Olive Oil
150%
Rice
100%
Source: Mario Jales, ICONE, Brazil
HS TARIFF LINE
52
35
24
23
22
22
22
20
20
20
20
20
19
18
16
15
15
12
12
11
10
08
08
07
07
06
04
04
04
02
02
02
0%
02
50%
01
AVE%
Starch
Wine
10
11
20 1
21
20 0
62
20 1
73
40 6
41
50 0
30
60 0
29
70 0
69
71 0
08
71 0
42
80 0
43
81 0
02
90 0
12
90 2
9
10 4 0
07
11 00
04
12 21
02
12 10
09
12 22
13
14 00
03
15 10
10
15 00
15
16 90
02
17 90
02
18 11
06
20 31
01
20 10
07
20 91
09
22 60
01
22 10
08
23 90
06
70
Agricultural Tariff Structure:
Switzerland, 2001
800%
700%
Source, IFPRI
1465
908
1664
600%
500%
400%
300%
200%
100%
0%
Bound Tariffs in % (AVE)
Tariff Structure of the United States (AVE)
350%
Tobbaco
300%
Grapes
Peanuts Sugar
200%
Peanuts
Products
Dairies
150%
100%
Source: Mario Jales, ICONE, Brazil
HS CHAPTERS
52
51
41
24
23
22
21
21
20
20
20
20
19
18
18
17
16
15
12
12
10
09
08
08
07
07
07
06
04
04
04
04
04
02
0%
02
50%
01
AVE%
250%
Market Access Pillar: Another Dilemma
 Market access versus the question of special products and
sensitive products.
 How far should Africa go? Propose specific numbers.
 What combination of formula coefficients and precise
percentages of special and sensitive products achieves
Africa’s interests?
Doha Round CGE Simulations





Common results:
Multilateral liberalization is beneficial at the
global level
There are potential gains for developing
countries
Developing countries own liberalization is an
important source of their gains
Removing subsidies may damage net food
importer countries
Doha Round CGE Simulations

Results differ among studies on how gains are
redistributed

1. What share of the benefits goes to developing
countries?
2. What share comes from agriculture
liberalization? From NAMA?

The Proposed Cuts in Agriculture
Market Access
Tariff cap
150
Cut = 46.7%
Tariff lines (in %)
130
Tariff cap
Cut = 42.7%
100
Cut: 70%
80
75
Cut = 38%
Cut = 64%
50
Cut = 57%
30
20
Cut =50%
Cut =33.3%
Tariff lines
DEVELOPED
COUNTRIES
(in descending
order)
DEVELOPING
COUNTRIES
A Tiered Formula for reduction of OTDS
subsidies
Overall Trade Distorting Support (OTDS): amber box + de minimis + blue box
Amounts of subsidies
Cut = 80%
60
Cut = 70%
10
Cut = 55%
0
Country A Country B
Country C
Bands
Thresholds
(US$ billion)
Cuts
3
> 60
EU
80%
2
10 – 60
US + Japan
70%
1
0 – 10 +
All DC
55%
Final Bound Total AMS – Amber Box:
A Tiered Formula
AMS (amber box)
Amounts of subsidies
Cut 3 = 70%
40
Cut 2 = 60%
Bands
Thresholds
(US$ billion)
Cuts
3
> 40
EU
70%
2
15 – 40
US + Japan
60%
1
0 – 15 +
All DC
45 %
15
Cut 1 =
45%
0
Country A Country B
Country C
Motivation
 Not enough to look at the tariff structures outcome.
 Economic impacts results useful for following reasons:
 For instance, what should be the view on sensitive products?
 Are other groupings positions in Africa’s offensive interests?
 What are the trade-offs with respect to the pillars and our
coalitions?
Scenario 1 (DV: 0% DVG: 0%)
Tariff band
0-20%
Developed
countries
65%
Developing
countries
20%
20-40%
75%
25%
40-60%
85%
28%
Above 60%
90%
30%
Scenario 2 (DV: 2% DVG: 20%)
Tariff band
0-20%
Developed
countries
65%
Developing
countries
20%
20-40%
75%
25%
40-60%
85%
28%
Above 60%
90%
30%
Scenario 3 (DV: 2% DVG: 20%)
Tariff band
0-20%
Developed
countries
20%
Developing
countries
15%
20-40%
30%
20%
40-60%
35%
25%
Above 60%
42%
30%
Further Assumptions
 Sensitive products: defined on basis of highest
MFN rates.
 Domestic support pillar: Three bands which
applied mainly to EU, US, Japan.
 Export subsidies: Total elimination in our
simulations.
 The domestic support and export subsidies are
same in each scenario.
Welfare Impacts (US$ millions)
Scenario I
Scenario II
Scenario III
SSA
118.59
97.19
88.85
North Africa
210.97
35.98
-152.36
EU-25
136.94
270.13
347.64
USA
1639.47
1310.39
927.59
Cairns
1502.17
1204.23
966.11
Japan
859.69
824.22
146.75
ROW
702.50
-93.03
-120.04
Impacts on the value of GDP
Scenario I
Scenario II
Scenario III
SSA
0.29
0.26
0.26
North Africa
0.75
0.48
0.16
EU-25
-0.04
-0.03
0.02
USA
0.07
0.06
0.04
Cairns
0.23
0.19
0.18
Japan
-0.13
-0.08
0.00
ROW
0.03
0.07
0.06
Trade balance (change $ mln)
Scenario I
Scenario II
Scenario III
SSA
106.54
82.69
68.44
North Africa
-105.89
-58.45
-12.27
EU-25
-46.36
-393.44
-295.1
USA
573.36
326.08
182.79
Cairns
623.65
461.43
370.68
Japan
-923.11
-870.51
-341.84
ROW
-228.18
452.19
27.29
Value Added (% change)
Scenario I
Scenario II
Scenario III
SSA
NAF
SSA
NAF
SSA
NAF
Cereals
1.98
1.79
1.88
2.04
2.05
1.95
Veget.
1.77
2.49
0.96
1.59
-0.01
0.37
Sugar
-0.31
2.05
-0.25
1.13
-0.16
0.12
Cotton
6.37
2.3
6.07
2.40
4.94
2.50
Cattle
-0.25
0.59
-0.21
0.26
-0.15
-0.05
Milk
-0.16
1.61
-0.13
0.94
-0.09
0.25
Fishing
0.05
0.03
0.07
0.01
0.03
0.01
VegOil
-1.12
245
-0.78
139
-0.22
26.96
Price index of global merchandise exports
Veg_Oil
Scenario III
Fishing
Scenario II
Scenario I
Milk
Cattle
Cotton
Sugar
Vegetables
Cereals
0
0.5
1
1.5
2
% change in price
2.5
3
3.5
4
Points on economic impacts
 Scenario I with no sensitive products and with deep
cuts is the most ambitious.
 Scenario III shows the actual impact of both the
sensitive products and lower ambition in the
market access.
 Scenario III indicates the significance of trade-off
between two important pillars.
Cost of sensitive products and lower
ambition
Welfare implications of lower ambition
200.00
Percent change in welfare from scenario I
150.00
100.00
Impact of sensitive products
Impact of lower ambition
50.00
0.00
SSA
-50.00
-100.00
-150.00
-200.00
NAF
EU25
USA
Cairns
Japan
ROW
Should market access be traded for
domestic support?
Approximate cost of lowering ambition in addition to sensitive products
80.00
EU25
60.00
40.00
20.00
0.00
-20.00
ROW
SSA
Cairns
USA
-40.00
-60.00
-80.00
-100.00
Japan
NAF
What it could mean for Africa
 Ambitious coefficients in agriculture remain the
best result for Africa.
 Africa loses through sensitive products what it
is supposed to gain from the ambitious tariffs
cut.
 A trade-off by the advanced developing
countries on market access that would reduce
ambition may not be in Africa’s interest.
 www.uneca.org/atpc