USITC Analysis of Trade Policy Issues

1

• USITC Analysis of Trade Policy Issues
• Presentation to Center For Strategic Research,
  Moscow Russia, April 6, 2006

Bob Koopman, Chief Economist Director, Office of
Economics United States International Trade
Commission These remarks are my own and do not
necessarily reflect the views of the USITC or any
of its individual Commissioners.
2
Overview of seminar

• Provide an overview of
• some of the questions we address and the economic
  models used at the USITC to analyze trade policy
  issues.
• how the ITCs work is used in U.S. trade policy
  formulation and deliberation.
• ITC is an independent agency not part of
  Presidential Administration. (See HSE
  presentation for more on this.)
• ITC operates under a number of specific statutes
  or laws
• Sections 131, 332, 2104,etc. of various trade
  acts 1974, 2002
• Office has approximately 48 staff, some visiting
  scholars. Very few support staff research
  assistants. Mostly PhDs, but some masters and
  ABDs.
• ITC staff level is 350, Office of Industries
  (partner office on most studies) has
  approximately 100 staff.
• USITCs independence is important and one of its
  most valued assets along with its employees.

3
Where is the ITC in the policy process?
4
Some analytical tools we use

• The USITC uses partial equilibrium and general
  equilibrium simulation models, and econometric
  models to analyze trade measures. Will provide
  examples as we go.
• PE models are largely commodity specific,
  Armington specification, applied at 6 or 8 digit
  tariff line. Some multi sector PE models are
  used also.
• GE models include
• large scale, 500 sector model of the U.S.,
  United States Applied General Equilibrium Model
  (USAGE).
• Large scale world model, Global Trade Analysis
  Project (GTAP) model (based at Purdue
  University.)
• Econometric models applied recently include
  gravity models, reduced form supply and demand
  models, logit and probit specifications.
• Choice of econometric specification depends very
  much on the question we are asked to answer.

5
Broad context for ITC analysis

• USITC analysis is used at many stages in the
  trade policy formulation and deliberation
  process.
• Our economic analysis, and that of others, NEVER
  determines the policy choice. It is used as one
  input to the policy process.
• ITC analysis is used before negotiations begin on
  FTAs (largely commodity specific PE analysis.)
  wont provide examples but can discuss if you are
  interested.
• ITC analysis is used after negotiations are
  completed, but before Congressional vote on FTA
  (usually GE analysis for economywide effects and
  broad sectoral impacts.)
• ITC analysis is used to inform debate on many
  other topics
• Significant U.S. import restraints and steel
  special safeguards (examples follow large
  single country CGE model)
• Chinas accession to WTO and FTAs (examples
  follow global CGE model)
• The benefits of improved trade efficiency
  logistics services and express delivery services.
  (example follows - econometrics)
• Non-Tariff Measures ( example follows
  econometrics)
• WTO cases

6
Example of independence

• Trade liberalization and growth
• Policymakers (USTR) wanted to assert that trade
  liberalization causes growth
• Asked ITC to prepare report that examined the
  theoretical and empirical research, summarized
  current findings, and recommended areas for
  further research.

7
Results were controversial

• Some Commissioners were skeptical that economic
  theory and empirical work could answer the
  question correctly.

8
Our answer was not what they wanted to hear
9
Recent summary from OECD Trade Committee summary
- 2006
10
But we still get askedsome recent questions

• Market for Medical Devices Japan
• Trade facilitation and security economic
  effects
• HK EBA
• Asia trade patterns and FDI
• Peru FTA, Colombia FTA, Ecuador FTA, etc
• South Korea and Malaysia probable economic
  effects advice
• We usually get 6 to 12 months to complete
  studies, but FTA studies usually done in 3 to 4
  months

11
Economic Models used at the ITC

• The USAGE model developed jointly with Peter
  Dixon, COPS, Monash U.
• A dynamic, regional applied general equilibrium
  (AGE) model of the U.S. economy with capability
  to analyze effects for distinct labor and
  household categories
• 513 sectors, dynamic
• Four closures
• Historical, decomposition, forecast, and policy
• Calibrates the model more tightly to observed
  data over time
• Examines changes to parameters on tastes and
  technology
• Helps identify areas in the data or model that
  need further attention
• Add-ons
• Take AGE results as an input and produce results
  for variables not included in AGE model
• Offer computational advantages
• State-level effects all 50 states - completed
• Distinct labor categories with different economic
  characteristics flexible, but up to 753
  occupations, self employment, and farming in
  development
• Distinct household categories with different
  primary factor endowments and demands (BLS data
  from Current Household and Payroll Surveys)
  again flexible, but probably no more than 10
  data gathering.
• Refinements
• Sweetener markets U.S. sugar policy explicit
  modeling of US quotas using complementarities.
• Jones act, upcoming

12
USAGE

• Ill elaborate a bit on this model will play
  important role in our future work by addressing a
  number of concerns customers have raised about
  economic models in general.
• Model validity, dynamics, state level results,
  labor detail, and household breakouts
• Data source 1992 U.S. Input Output table
• Updated annually annual BLS 182 sector macro
  data and trade data - currently to 2004.
• Major new challenge BEA changed from SIC to
  NAICs classifications, most up-to-date IO table
  is in NAICs 2002. We must redo all our data.
• Then projected to 2015 using outside forecasts
  for key macro variables and labor.

13
USAGE closures to ensure model fits historical
data and official government forecasts

• We follow Australian Monash approach and use
  closures to validate the model on recent data
  and fit projections to official government
  forecasts CBO, BLS, etc.
• Historical closure
• Estimating changes in tastes/technology,
  1992-2001
• Decomposition closure
• Decomposing effects of economic developments,
  1992-2001
• What were the effects of changing tastes and
  technologies versus observed changes in policy
  variables.
• Forecasting (projection) closure
• Projecting, 2001-15, annually forms baseline
  consistent with government macro forecasts
• Policy closure
• Policy analysis, 2001-15, annually examines
  potential effects of policy changes as deviations
  from baseline.

14
Historical simulation

• An attempt to rectify problem of infrequent and
  outdated Input-Output (I-O) tables
• an attractive alternative to mechanical update
  methods such as RAS
• Historical simulation approach
• Flexible approach regarding data requirements
• Based on USAGE-ITC theory
• By-product
• Detailed and interpretable estimates of changes
  in household preferences and industry
  technologies
• Central input to decomposition and forecast
  simulations
• By-product overshadows up-to-date I-O tables as
  the major output of historical simulation
• Done by swapping endogenous and exogenous
  variables in closure.

15
Historical simulation, 1992-2001
Naturally exogenous data, 2001 e.g., CIF import
prices, foreign currency Tariff rates
Changes in tastes/technology, 1992-01 e.g.,
Shifts in import/domestic preferences Shifts in
foreign demand Primary-factor-saving technical
change and capital/labor bias in technical change
Naturally endogenous data, 2001 e.g., Imports,
exports by commodity Employment and capital
inputs by industry
Historical closure, long run
16
Historical and decomposition simulations,
1992-2001
Naturally exogenous data, 2001
State level decomposition
Naturally endogenous data, 2001
Changes in tastes technology, 1992-01
State-level add-on
Decomposition of effects of trade policy and
technological change, 1992-2001
Historical closure
Decomposition closure
17
Annual forecast simulations, 2001-15
Annual forecasts naturally exogenous
endogenous variables Macro Trade policy
Exports, Imports
Changes in tastes technology industry,
commodity specific
Forecast paths
Shifts in functions (e.g., foreign demands,
export supplies) and macro coefficients, e.g., APC
Forecast closure, short run
18
Annual forecast simulations

• Contributes to generating believable AGE analysis
• what if answers depend significantly on the
  basecase forecast
• Incorporates detailed information from several
  expert groups (in our case official govt agency
  forecasts or projections.)
• USAGE-ITC forecasts/projections are of interest
  by themselves
• Establishes a foreign trade baseline in some
  detail. Not a part of original projections.

19
Annual forecast and policy simulations, 2001-15
Forecasts naturally exog. endogenous Macro
Trade policy Exports, Imports
Forecast paths
Shocks modified forecasts for naturally
exogenous variables
Policy effects as deviations from forecast path
Shifts in functions
? tastes technology
Policy paths
Forecast closure
Policy closure
20
Model validation in USAGE-ITC framework

• USAGE-ITC has most of requirements for
  statistically tested AGE forecasting model
• Dynamic specification to allow forecasting to a
  year outside data
• Fitting of historical events (historical
  simulations)
• Incorporates detailed trend estimates of
  technology and preferences
• But no statistical measure of validity
  estibration, a combination of calibration and
  highly restricted estimation of select parameters
  with no statistical basis.