Post on 22-Mar-2016
description
India-EU trade reforms: Comments on modelling and policy analysis
R. Jongeneel (LEI-WU)
Agricultural Trade Project
25 April 2008 LEI Seminar
Agricultural Trade Policies and Development
Contents
Introduction Some experiences & lessons Reasons for scrutiny post-model
analyses Specific issues India application Concluding remarks
Introduction
Science/models: Looking for an appropriate map
Why use models ? Need for ‘systematics’ in the analysis Need for quantification Need for explicitizing assumptions Need for framework to discuss
disagreements Need to clarify costs and benefits
(welfare impacts) from policy changes Need to explore policy options …
Why use models ?
Check for alternative policy scenario’s Check for all kind of impacts (income,
welfare, markets, budget, environment, …
Do sensitivity analysis about uncertainties and show impacts
Qualifications (i) Models simplify reality Models often hide uncertainties Models use a lot of basic assumptions and
supplementary assumptions Models are weak in accounting for changes
in behavior Models are often too restrictive wrt market
structure (e.g. deviations from full competition such as monopolistic comp., etc)
Qualifications Model closure (and non-considered
feedback links) are important Models don’t prescribe policy, but can
be easily abused for this Institutional issues are often
downplayed or presumed It is difficult to include the full real
world dynamics (expectations, non-linearities, comparative static)
Some experiences and lessons
Some examples
Use models for their strengths not their weaknesses
Power of GTAP and its ‘family’-members lies in world-wide impact analysis of trade policy changes
See overview partial (Harbinson) (slide 1) and full trade liberalisation analysis (slide 2) (source J-C. Bureau)
Model use and WTO /TrLib (i)
Models and WTO/TrLib (ii)Even with
same model and same scenario different
researchers come up
with different results
Lessons from modelling WTO/TrLib (i)
Only a few model families are used Caution needed for artifical
consensus More liberalisation leads to larger
gains: already by assumption Gains are actually quite small (0.x%
of GDP) In particular developing countries
gain relatively little (although they did in older studies).
Lessons from modelling WTO/TrLib (ii)
More recent model version show smaller impacts due to better data (applied tariffs, TRQ-treatment, other NTBs)
Better desaggregation of DC (still weak on impacts of binding overhang, tariff data precision, somtimes simplistic assumptions wrt consumer surplus changes offsetting producer surplus changes)
Few models still take the relationship between intervention prices, tariffs and export subsidies properly into account (EU)
Post-modeling analysis
Model use matters
Model have limited direct impact, but large and increasing indirect impact
Models are (more) used in trade panels (with country focus!)
Need for good quality management & accountability about performance and limitations
When focus on specific country-market-impacts post-model analysis is required
Post-modeling analysis Models are calibrated: lack empirical testing Specification errors (aggregation,
heterogeneity, down-scaling, macro-micro, lacking detail in policy implementation)
No market power Instantaneous adjustments (signal
transmission, responsiveness) Dynamics & structural change not well-
captured Balance of trade-closure
Also plea for pre-modeling analysis
Specific issues wrt India
Comments/questions on analysis
Modeling assumptions
BoP closure rule => dX = dM Factor mobility assumption Meat import fixation
Comments/questions on analysis
Context: quantitative assessment of India-EU FTA
Q1 : PTA or FTA? Q2 : ‘external’ tariff assumption? Q3: TC and TD (how to explain) Q4: TRQ treatment Scenario design
Comments/questions on analysis Post-modeling analysis (fed/state level desagr?) Check for main affected products– Padi rice– Processed rice– Sugar cane, sugar beet?– Textiles and leather?– Wool, silkworm cocoons?– (Manufactures)
Analyse in detail– Policy representation– Price transmission
Concluding remarks
Some conclusions GTAP contains lot and still increasing
amount of expertise on modeling, trade volume, price and policy data
CGE is encompassing but captures not everything properly and with proper detail
Plea for (pre-) and post-modeling analysis– (scenario design)– (scenario implementation: policy transl.)– interpretation and modification of results– derived impact analysis