Linking Defratap and Aglink

Linking DefraTAP and AGLINK

Elisavet Kitoua, Grant Daviesa, George Philippidisb

a Department for Environment, Food and Rural Affairs (Defra) – United Kingdom

b Aragonese Agency for Research and Development (ARAID), Centre for Agro-Food

Research and Technology (CITA) – Spain

A paper presented at the EcoMod2012 Annual Conference

Seville, Spain, July 4 - 6, 2012

(Draft: Please do not quote)

Linking DefraTAP and AGLINK

Abstract

This research is focused on linking two different types of equilibrium models, one general and one partial. An explicit aim of this paper is to exploit the strengths of each modelling approach in a consistent manner which provides an improved evaluation of trade and policy reforms of particular pertinence to EU agriculture.

A General Equilibrium (GE) model was employed to run two Free Trade Area (FTAs) scenarios, whilst the Partial Equilibrium (PE) approach was utilised to generate results for two different sets of Biofuels policies. In each case, the results from each model were fed into the other to better assess the trade (GE) and agricultural sector (PE) impacts of each set of reforms. This research remains a work in progress and we welcome your input.

Introduction

Rational and objective

Our modelling team uses both partial and general equilibrium models to investigate trade and policy reform scenarios, e.g. CAP reform, FTAs, WTO, and their impact on the agricultural sector, with particular emphasis in the UK. Multi-region GE models with their gross bilateral trade flows supported by the GTAP global database, are ideal for evaluating bilateral, regional and multilateral trade reform scenarios and the extent to which the repercussions of such reforms extend to, and beyond, the agricultural sector. Inevitably, however, models of global trade lack the finer sectoral detail which is inherent within their PE counterparts. Consequently, PE models are better suited to a detailed analysis of specific agricultural sector impacts. Consequently, this research seeks to exploit the strengths of each approach to evaluate a set of pertinent policy questions of particular pertinence to EU agriculture.

Modelling approach and methodology

Our GE model, DefraTAP, is built on the GTAP-AGR model with added specificity for EU agricultural policies. Our PE model (AGLINK-Cosimo) is the product of collaboration between the OECD and FAO looking only at agriculture. We want to be able to use both models to answer policy related questions. We are following the approach of a sequential implementation of scenarios, where the outcome from one model serves as the input for the other. For CAP reform related work we are using the PE model, which better represents EU agricultural policy, to get the first round of results to then feed into the GE model and analyse the trade impacts. For FTA and WTO related work we run the GE model and use the results to feed into the PE and investigate the impact on EU agriculture.

GE scenarios

A scenario design of two ongoing free trade area scenarios was implemented into our CGE model, which served as an input into the PE framework.

EU-Canada CETA – description and results

Background

The EU and Canada are traditional trading partners, albeit with a sectoral focus on wines and spirits, education, science and technology etc and a regional dimension linking Canada and the UK and France, Quebec and France etc.

In July 2007 at the annual EU-Canada summit the idea of fostering closer economic relations were first presented. While at the time the Doha Development Round talks looked to be reaching a conclusion, the situation has since changed. This resulted in both the EU and Canada signing or negotiating FTAs with various of their trading partners.

Nine rounds of negotiations have taken place so far between the EU and Canada, with a view to reaching an agreement in 2012 on a Comprehensive Economic and Trade Agreement (CETA).

In our analysis we have used the latest two rounds of offers from both sides (at HS6 digit code) and ran three scenarios in total. Two of them reflect the actual offers and one simulates a complete removal of tariffs. The latter serves as a benchmark for establishing the degree of ambition of the negotiations.

Modelling approach

Our model, DefraTAP is a variant of the standard GTAP model. It incorporates the agricultural improvements made in the GTAP-AGR model and adds further EU agricultural policy specificity. Key features are:

- based on Keeney and Hertel (2005), intermediate and primary factor demands are constant elasticity of substitution (CES) substitutable, as are intermediate feed inputs in livestock sectors. The transfer of labour and capital factors between agricultural/non-agricultural sectors is controlled by an elasticity of transformation (CET) to capture observed differentials in wages and rents in each sub-sector;

- a three-stage weakly separable CET nest capturing land-use heterogeneity across different agricultural activities is employed;

- following the procedure of van Meijl et al. (2006), an econometric nonlinear land supply function is estimated for the 113 regions in version 7 database, complete with an aggregation program for the relevant land supply parameters;

- EU common agricultural policy (CAP) (e.g., set aside, CAP budget, intervention prices, quotas etc.) is modelled, whilst raw milk quotas are modelled in Canada;

- additionally, Canadian and EU tariff-rate quotas (TRQs) are modelled (Elbehri and Pearson 2005), although the broad GTAP sectoral classification restricts TRQ representation to Canadian dairy imports and EU wheat and beef imports;

- to improve the quality of model estimates, closure changes and further secondary data estimates are employed to capture investment accumulation and non tariff barrier liberalisation, respectively.

The regional and sectoral aggregation employed can be found in Table A.I in the Annex and uses version 7 of the GTAP database.

By use of the TASTE software (Horridge and Laborde, 2008), the Doha modalities are modelled. A likely CAP reform scenario is also included, where following Oskam et al. (2004), a digression rate of 2 per cent per annum in nominal euros is applied to the EU single farm payment (SFP) (pillar I) from the 2013 ceiling limits and modulation rates are raised to 20 per cent from the current 5 per cent limit. With complete decoupling, all price support and quantitative constraints are eliminated, whilst the Canadian milk quota is maintained. Finally, a long run closure swap is implemented, allowing an endogenous ‘steady state’ interaction between changes in investment and the capital stock. Thus, it is assumed that investment moves in tandem with fixed savings rates, respecting the long run empirical observation that domestic saving finances domestic investment (Francois et al. 1996).

Results

As the GE model is well suited to treat trade, we used the trade output from the GE to feed to the PE model. Changes in imports and exports from the GE model can be found in Table 1, for the EU and in Table 2 for Canada.

Table 1: EU change in exports and imports vs baseline (%)

Exports / Imports
Scenario 1
vs Baseline / Scenario 2
vs Baseline / Scenario 3 vs Baseline / Scenario 1
vs Baseline / Scenario 2
vs Baseline / Scenario 3 vs Baseline
% change / % change / % change / % change / % change / % change
Rice / -0.401 / -0.155 / -0.163 / 1.446 / 1.315 / 1.249
Wheat / 0.753 / -0.201 / 1.1044 / 24.932 / 0.091 / 23.028
Other grains / -0.156 / -0.078 / 0 / 0.383 / 0.096 / 0.287
Vegetables, fruits, nuts / -0.081 / -0.081 / 0.1616 / 0.094 / 0.000 / 0.000
Oilseeds / 0 / -0.127 / 0.1267 / 0.084 / 0.084 / 0.000
Sugar / -0.329 / -0.117 / -0.141 / 0.223 / 0.334 / 0.223
Plants / -0.444 / -0.177 / -0.177 / 0.205 / 0.103 / 0.103
Other crops / 0 / 0.0815 / 0.2445 / 0.281 / 0.187 / 0.187
Cattle & sheep / -0.118 / -0.118 / 0 / 0.086 / 0.086 / -0.086
Pigs & poultry / -0.09 / -0.09 / 0 / 0.202 / 0.202 / 0.202
Milk / -0.763 / -0.158 / -0.026 / 0.484 / 0.242 / 0.242
Fishing / -0.342 / 0 / -0.085 / 0.589 / 0.295 / 0.295
Red meat / -0.384 / -0.192 / -0.192 / 0.989 / 0.148 / 0.148
White meat / -0.295 / -0.074 / 0 / 3.900 / 0.487 / 0.627
Vegetable oils, fats / -0.094 / 0 / 0.0935 / 0.322 / 0.161 / 0.242
Dairy / 15.948 / -0.244 / -0.244 / 2.765 / 2.572 / 2.637
Processed sugar / -0.223 / -0.223 / -0.223 / 0.188 / 0.094 / 0.126
Other food products / 1.5926 / -0.335 / -0.168 / 0.751 / 0.282 / 0.563
Beverages, tobacco / 0.1925 / 0.2887 / 0.2887 / 0.190 / 0.095 / 0.095

Table 2: Canadian change in exports and imports vs baseline (%)

Exports / Imports
Scenario 1
vs Baseline / Scenario 2
vs Baseline / Scenario 3 vs Baseline / Scenario 1
vs Baseline / Scenario 2
vs Baseline / Scenario 3 vs Baseline
% change / % change / % change / % change / % change / % change
Rice / 874.592 / 881.235 / 847.377 / 4.061 / 3.926 / 4.446
Wheat / 14.503 / -0.409 / 12.354 / 6.255 / 0.774 / 7.315
Other grains / -0.122 / -0.183 / -0.357 / 0.594 / 0.243 / 0.905
Vegetables, fruits, nuts / -1.225 / -0.204 / -1.817 / 0.427 / 0.119 / 0.576
Oilseeds / -1.903 / -0.252 / -2.476 / 0.343 / 0.157 / 0.911
Sugar / -16.793 / 37.692 / 31.433 / 1.905 / 0.481 / 2.867
Plants / -0.985 / -0.417 / -1.382 / 0.468 / 0.214 / 0.560
Other crops / 0.743 / 2.190 / -1.110 / 0.906 / 0.492 / 1.418
Cattle & sheep / -1.373 / -0.448 / -2.177 / 1.299 / 0.308 / 0.962
Pigs & poultry / -0.261 / 0.191 / -0.734 / 1.171 / 0.076 / 0.396
Milk / -1.659 / -1.324 / -3.930 / -0.248 / 1.079 / 2.246
Fishing / 0.132 / 0.324 / 0.040 / 0.989 / 0.717 / 1.183
Red meat / 1.308 / -1.085 / -2.170 / 1.702 / 1.025 / 1.782
White meat / 2.628 / -1.093 / -1.901 / 1.635 / 0.503 / 0.628
Vegetable oils, fats / -0.690 / -0.097 / -1.069 / 0.667 / 0.797 / 1.279
Dairy / 19.360 / 12.511 / 11.784 / 46.903 / 1.013 / 1.468
Processed sugar / 0.582 / -0.756 / -0.396 / -0.060 / 0.560 / 0.871
Other food products / 2.981 / 0.339 / 1.939 / 2.713 / 0.301 / 0.727
Beverages, tobacco / 0.020 / -0.078 / -0.117 / 2.181 / 2.005 / 2.142

EU-Mercosur – description and results

Background

The EU and Mercosur first started FTA negotiations back in 1999. Partly as a result of the ongoing problems in reaching a consensus on the DDA talks, negotiations were formally suspended in 2004. However, renewed interest in the deal re-opened the negotiations in May 2010. In the meantime both regions have gone ahead with other FTAs and trade agreements.

In this analysis we have simulated a complete removal of tariffs that serves as a benchmark for establishing the degree of ambition of the negotiations.

Modelling approach

In this analysis the standard DefraTAP is used. The features of the standard model are:

- a three-stage weakly separable CET nest capturing land-use heterogeneity across different agricultural activities is employed;

- EU common agricultural policy (CAP) (e.g., set aside, CAP budget, intervention prices, quotas etc.) is modelled.

The regional and sectoral aggregation employed in this scenario can be found in Table A.II and in this incarnation version 7.1 of the GTAP database is used.

Results

In this scenario, we again used the trade output from the GE model to feed to the PE one. Table 3 below, show the changes in imports and exports for the EU and Table 4 the changes for the Mercosur countries.

Table 3: EU change in exports and imports vs baseline (%)

Exports / Imports
Scenario
vs Baseline / Scenario
vs Baseline
% change / % change
Rice / 35.705 / 0.381
Wheat / 8.018 / 2.352
Other grains / 4.376 / 7.403
Vegetables, fruits, nuts / 4.191 / -1.174
Oilseeds / 11.479 / -7.845
Sugar / 212.056 / -65.512
Plants / 6.072 / -1.316
Other crops / 13.142 / 2.960
Cattle & sheep / 0.173 / -20.087
Pigs & poultry / 4.242 / -3.013
Milk / 13.462 / -6.625
Red meat / 7.871 / 281.498
White meat / 10.453 / 71.179
Vegetable oils, fats / 5.374 / -4.272
Dairy / 3.414 / 60.030
Processed rice / 2.837 / 12.416
Processed sugar / 60.661 / 52.886
Other food products / 2.311 / -0.351
Beverages, tobacco / 1.183 / 0.539

Table 4: Mercosur change in exports and imports vs baseline (%)

Exports / Imports
Scenario vs Baseline / Scenario vs Baseline
% change / % change
Argentina / Brazil / Paraguay / Uruguay / Argentina / Brazil / Paraguay / Uruguay
Rice / 97.951 / 11.579 / 0.671 / -5.229 / 46.538 / 48.447 / 18.032 / 131.451
Wheat / 3.668 / 6.603 / -0.098 / -39.931 / 27.591 / 22.612 / 57.915 / 31.272
Other grains / 3.850 / -15.212 / 9.692 / 28.420 / 2.757 / 36.909 / 5.611 / 2.452
Vegetables, fruits, nuts / 18.666 / -25.365 / -20.493 / -17.418 / 8.612 / 20.418 / 14.920 / 5.850
Oilseeds / 2.062 / -29.174 / -14.600 / -9.027 / -10.610 / 1.718 / -6.797 / -1.781
Sugar / -66.010 / -82.810 / -79.644 / -66.403 / 14.625 / 128.216 / 562.825 / 13.456
Plants / -1.806 / -40.832 / -19.281 / -20.906 / 13.604 / 18.412 / 8.202 / 4.252
Other crops / -8.009 / 19.870 / -21.010 / -23.206 / 12.602 / 83.946 / 5.292 / -3.118
Cattle & sheep / -9.783 / -58.488 / -45.083 / -32.507 / 11.841 / 204.463 / 72.839 / 31.021
Pigs & poultry / 11.056 / -29.900 / -27.631 / -10.692 / 3.153 / 48.324 / 67.066 / -2.154
Milk / -40.961 / -74.022 / -56.473 / -71.930 / 44.654 / 99.732 / 59.394 / 110.148
Red meat / 9.961 / 520.850 / 403.552 / 19.312 / -10.096 / 58.669 / 21.348 / 21.760
White meat / 165.742 / 2.433 / -45.923 / 111.022 / -15.113 / 97.152 / 17.328 / 15.705
Vegetable oils, fats / -8.788 / -40.033 / -24.752 / 35.658 / 7.239 / 44.442 / 16.191 / 1.162
Dairy / 142.500 / 296.665 / 3038.719 / 79.717 / 11.712 / 53.878 / 19.796 / 41.183
Processed rice / 10.534 / -31.938 / -15.574 / 69.831 / 1.224 / 14.804 / 23.576 / 26.380
Processed sugar / 0.183 / 98.100 / 1190.243 / -21.359 / 1.347 / 60.975 / 37.884 / 0.143
Other food products / 6.135 / -1.294 / -11.735 / -6.277 / 4.166 / 21.519 / 13.819 / 7.497
Beverages, tobacco / 1.403 / 2.192 / -12.544 / -9.396 / 9.543 / 14.333 / 11.119 / 16.040

GE into PE

As a first attempt, we tried employing trade balance results from DefraTAP. Unfortunately, trade balances in AGLINK are an important variable in solving for world commodity prices and are not linked to internal market clearing. Therefore, we opted for the changes in imports and exports in individual regions and countries generated by the DefraTAP model. More specifically, trade changes for both the EU27 and Canada as external shocks to AGLINK.