economic models & consulting for public decision making: remi training & workshop july, 2003...
TRANSCRIPT
Economic Models & Consulting for
Public Decision Making:REMI Training & Workshop
July, 2003
Regional Economic Models, Inc.
Amherst, Massachusetts
The Planner’s Monday Lament(or how to lose millions painfully)
“Right back where we started from,
Do-dah, do-dah,
All the livelong day.”
Central Importance of the Economy
Economy
Geography
Environment
Land Use
Demography
Energy
Transportation
Epidemiology
Are You Involved In:
• Formulating policy?
• Making recommendations?
• Collecting and analyzing data?
• Implementing models?
• Consulting, using consultants,
other?
Competition & Advantage
• Competition is any area in the world with quality goods and services to sell, and access to markets and transportation.
• Advantage goes to relatively productive total resources, not to merely cheap labor.
• Competition and advantage are regional.
Assess Long-range Effects• Workforce training and public health planning• School, elderly, ethnic, and total populations• Employment, budgets, taxes, land use planning• Transportation alternatives (ports, corridors, …)• Cluster policy examples:
– Communications– Chemicals (e.g., polymers, plastics, …)– Motor Vehicles (e.g., multi-state regional industry)– Furniture Industry and Lumber Industry products– Misc Professional (e.g., geo-spatial NASA spinoffs)
Where Does REMI Come In?
REMI provides:
•A software tool & knowledgeable client support.
•Consulting & economic analysis.
•Model integration.
Where Does REMI Come In?
REMI provides:
•A software tool & knowledgeable client support.
•Consulting & economic analysis.
•Model integration.
REMI Policy Insight Features
Applies economic theory, inter-industry links: Includes New Economic Geography Theory
Is calibrated & estimated using regional data
Is dynamic and predicts when results will occur
Cause & effect model structure explains results
Alternative structures allow sensitivity tests
Is the leading Policy Analysis model in the U.S.
REMI Model Applications
• Economic development
• Energy
• Transportation
• Environment/Regulation
• Taxation
• National-level applications
Unexpected EventsModeling and Analysis
• Understanding the Problem
• REMI’s Modeling Framework
• Application Examples
Unexpected Event Applications
• Earthquakes for the U.S. Insurance Industry• Flooding of the Mississippi River• Hurricane Effects on Florida• Port Disruption in the Northwest U.S. and Canada• 9-11 Attacks; Economic Effect on New York• Potential Economic Effects of a Bio-terror Attack
in a Major U.S. City
Energy Model
Land Use ModelPolicy Insight® Economic and Demographic
Model
Transportation Model
Financial and Exchange Rate
Model
Key Security Constraints
Energy Prices Economic Activity
Housing Demand
Land Prices
Economic Activity
Flow Capital Funds
Key Structural Breaks
Economic Effects
Travel Time Changes
Economic Flows
Examples of Model Integration
• Energy 2020• E-GAS• GIS• UrbanSim• HERS-ST• NFIB
How Does REMI Do It?
Alternative Simulation Forecast
Control Forecast
REMI Policy Variable Examples
•Output (by firm or industry)
•Spending (by demand type)
•Taxes or Subsidies
•Employment (firm, industry)
•Productivity (i.e., output per
worker per year by industry)
•Migration (international,
economic, retired)
•Non-pecuniary (amenity)
•Labor Force Participation
Rates (by cohort)
•Birth and Survival Rates (by cohort)
•Occupational Supply (by
occupation)
•Production Costs (by industry)
•Fuel, Labor, Capital Costs (by
industry)
•Prices (by consumer commodity)
•Exports to Non-local Markets (by
industry)
•Imports from Non-local Markets (by
industry; e.g., import substitution)
REMI Policy Insight Results: Sample Results by Industry
•Sector Employment (to 3 digits)
•Intermediate Output Demand Empl
•Local Consumption Employment
•Investment Activity Employment
•Government Demand Employment
•Export Employment (Gross)
•Relative Production Cost
•Relative Composite Labor Cost
•Relative Fuel Cost
•Relative Capital Cost
•Relative Composite Input Cost
•Relative Factor Productivity
•Regional Purchase Coefficient
•Labor Intensity
•Average Annual Wage Rate
•Industrial Mix Index
•Demand, Imports, Self Supply
•Exports: Region, US, World
•Sector Output (to 3 digits)
•Value Added
•Wage & Salary Disbursements
REMI Policy Insight Results: (Continued)
• Personal Income
• Transfer Payments
• Taxes
• Disposable Personal Income
• Price (PCE) Index
• Real Disposable Income
• GRP (by 22 final demand
categories, & by sector)
• Empl (by demand, & sector)
• Population
• Migrants
• Natality Rate
• Survival Rate
• Labor Force
• Participation
Rate
Aggregate Variables: Age/Gender/Cohort Variables:
REMI Model Structure (2002 - )WITH ECONOMIC GEOGRAPHY LINKAGES
Output
Market SharesLabor & Capital Demand
Population & Labor Supply
Wages, Costs & Prices
REMI Model Linkages (Excluding Economic Geography Linkages)
State and Local Government Spending
Investment
(1) Output(1) Output
Output
Exports
Consumption
Real Disposable Income
(4) Wages, Prices, and Production Costs(4) Wages, Prices, and Production Costs
Employment Opportunity
Housing Price
Wage Rate Composite Wage Rate Production Costs
Composite PricesReal Wage RateConsumer Price
Deflator
(3) Demographic(3) Demographic (2) Labor & Capital (2) Labor & Capital DemandDemand
Optimal Capital Stock
Employment
Labor/Output Ratio
(5) Market Shares(5) Market Shares
Domestic Market Share
International Market Share
Population
ParticipationRate
Migration
Labor Force
Economic Geography Linkages
Intermediate Input Productivity
Intermediate Inputs (1) Output Block
(4) Wages, Prices, and Production Costs
Composite Wage
Production Costs
Composite Prices
(3) Demographic and Labor Supply
(2) Labor & Capital Demand
Labor Access Index
Employment
Labor Productivity
(5) Market Shares
Domestic Market Share
International Market Share
EconomicMigrants
Output
Commodity Access Index
Integrated Modeling Approach REMI integrates Economic Geography (labor
& product agglomeration, accounting for distance costs & non-homogenous factor & product markets) with key modeling tools:
Model Tool Aspect
I-O Inter-industry processing from & to sectors (input-output relationships)
CGE Computable general equilibrium tendencies in factor and material input markets
Econometric Estimates derived from panel data
Types of ModelsCustomized Models Using Region Specific Data
SingleRegionModel
Single Region Model
Multi-Regional Model (linkages among regions)
Gross Commuter FlowsMulti-region exports and imports
Region1
Region2
Region3
Trade and Commuter Flow Linkages
Local Demand
Output Local Demand
Output Local Demand
Output
Flows based on estimated trade flows
Disposable Income
Local Earnings by non-Residents
Local Earnings by non-Residents
Disposable Income
Local Earnings by non-Residents
Commuter linkages based on historic commuting data
Disposable Income
Steps in Model Building
• Data preparation
• Model building programs
• Quality assurance
• Ongoing research & development
The New Economic Geography• Monopolistic Competition Base• Labor Productivity Concepts• Commodity Price Concepts• Revised Equations• New Policy Variables• “Nullify” Variables• Bulletin on Demand Variables• Group Policy Variables• Market Share/Trade Flow, & Price Elasticities• Transportation Costs Matrix• Economic Geography Conclusion
New Economic Geography(Based on Monopolistic Competition)
• Increasing economies of scale for firms (decreasing marginal costs).
• Access to product varieties from different regions.• Different delivered prices in the same industry due
to transportation or other distance deterrence costs.• Cross hauling, cluster agglomeration, and labor
productivity are explained by microeconomic gains from having choice and variety.
New EG FeaturesLabor Productivity Concepts
• Labor Access Index– Occupation & industry effects of selecting best workers
• Composite Labor Cost– Relative wage rate adjusted for labor access index
– Includes commuter distance cost blended into wage rates
• Occupation Codes– Standard occupation codes (SOCs) are implemented
– Better groupings with sectors for analysis & display
New EG FeaturesCommodity Price Concepts
• Delivered Price (wtd avg by sector at place of use)– Source production cost + transport cost to place of use– Production & transport costs reflect every source location
• Commodity Access Index (by sector at place of use)– Producer & consumer effects of better commodity access– Reflects market share changes due to composite price changes
• Composite Price (by sector at place of use)– Delivered price adjusted for commodity access index
• Production Cost (by sector at place of use)– Includes intermediate inputs from all sectors & locations– Local production cost captures intermediate input composite
prices; composite labor cost; and capital and fuel costs
New EG FeaturesRevised Equations
• The Housing Price Equation– Population & Real Disposable Income Changes
• The Wage Equation– Employment Opportunity & Occupational Demand
• The Stock Adjustment Process– New Parameters
• The Economic Migration Equation– Empl Opportunity, RWR, Consumption Access Index
• The Market Share Equations– Economic Geography & New Price Elasticities
– More sensitive to production cost changes
New EG FeaturesNew Policy Variables
• Industry Sales & Industry Employment-nonlinear– Ignores firms being displaced in a region by new firms
who are competing locally or in nearby markets.
– Treated as “international exports” for “demand source.”
– Applies input value less the local economy’s indirect effects on international exports due to production costs.
• Firm Sales and Firm Employment-nonlinear– Accounts for firms being displaced in a region by new
firms who are competing locally or in nearby markets.
– Displacement effect correlates strongly to the degree to which current firms already serve competitive markets.
New EG Features“Nullify” Variables
• Industry Sales and Employment: nullify applies.• Firm Sales and Employment: nullify not defined.• Nullify Workaround (Firm Sales & Employment)
– Defined for nullify investment only.
– Apply to non-residential capital stock variable only, in same amounts, sign, & year as the inputs for detailed construction (any non-residential) and for investment (non-residential).
– Apply to residential capital stock variable only, in same amounts, sign, & year as the inputs for detailed construction (any residential), and for investment (residential).
– Apportion as appropriate between non-residential capital stock and residential capital stock for the input to construction (non-detailed).
New EG FeaturesBulletin on Demand Variables
• Demand Variables They now cause an output response in every region which supplies the region with the changed demand. The regional trade flow shares drive the response.
• Applicability– Exogenous demand– Intermediate demand– Investment (structures) demand and PDE demand– Consumer demand and government demand– Indirect demand from bridge tables such as I-O
table, demand vector tables, and translator tables
New Policy Insight FeaturesGroup Policy Variables
• Users can define several policy variables as one.• Users can include input values or not as desired.• Useful for applying the same variables to many
simulations, or applying then editing as needed.• Example: define your own visitor days variables as
group variables, with or without values, e.g.:– Consumer demand variables (include transportation, food
& beverage, clothing & shoes, gasoline, other services, etc)– Hotel sector– Amusement & Recreation sector (or detailed output sales)– Motion Picture sector (or detailed output sales)
New EG FeaturesMarket Share/Trade Flow,
& Demand Price Elasticities
• Market Share/Trade Flow Displays– Share of local demand met by regions in model.
• Demand Price Elasticities– Alternative elasticities option for new regional
controls.– For any given simulation, test your model’s
sensitivity.
1994
PRODUCTION COSTS - CREDIT & FINANCE
C AL IF O R NIA
N E W Y O RK
TE X A S
FL O RID A
P E N NS Y L V AN IA
ILL INO ISO H IO
M ICH IG AN
N O RT H C AR O LIN A
G E O R G IA
V IR G IN IA
M A
IN
M IS S O UR I
W IS CO NS IN
TE N N E S S E E
W AS H IN G TO N
M AR Y LA ND
M INN E S O TA
LO UIS IA NA
A L
K E NT UC KY
A RIZO NA S C
C O LO R AD O
O K LA HO M A
O R E G O N
IO W A
M S
K AN S AS
A RK AN S AS
U TA HN E BR AS K A
N E W M E X IC O
M AIN E
N E V AD A
N HID AH O
M O NT AN A
S O U TH DA KO T A
N D
W Y O M ING
M exico
B ah amas
B erm ud a
0 2 0 0 4 0 0
M i le s
M a p La ye rsC o u n tr y (L o w R e s )S ta te (L o w R e s )C o u n ty (L o w R e s )
P ACT610 .0 0 0 0 to 1 .0 0 4 0 (1 2 2 )1 .0 0 4 0 to 1 .1 0 8 3 (1 2 2 )1 .1 0 8 3 to 1 .1 5 2 2 (1 2 2 )1 .1 5 2 2 to 1 .1 9 1 8 (1 2 2 )1 .1 9 1 8 to 1 .2 1 6 7 (1 2 2 )1 .2 1 6 7 to 1 .2 3 5 4 (1 2 2 )1 .2 3 5 4 to 1 .2 4 7 1 (1 2 3 )1 .2 4 7 1 to 1 .2 5 7 0 (1 2 2 )1 .2 5 7 0 to 1 .2 6 4 7 (1 2 2 )1 .2 6 4 7 to 1 .2 7 1 1 (1 2 2 )1 .2 7 1 1 to 1 .2 7 7 5 (1 2 2 )1 .2 7 7 5 to 1 .2 8 4 0 (1 2 2 )1 .2 8 4 0 to 1 .2 8 8 2 (1 2 3 )1 .2 8 8 2 to 1 .2 9 2 2 (1 2 2 )1 .2 9 2 2 to 1 .2 9 6 4 (1 2 2 )1 .2 9 6 4 to 1 .3 0 0 7 (1 2 2 )1 .3 0 0 7 to 1 .3 0 6 3 (1 2 2 )1 .3 0 6 3 to 1 .3 1 3 5 (1 2 2 )1 .3 1 3 5 to 1 .3 2 0 3 (1 2 3 )1 .3 2 0 3 to 1 .3 2 7 9 (1 2 2 )1 .3 2 7 9 to 1 .3 3 6 8 (1 2 2 )1 .3 3 6 8 to 1 .3 5 0 5 (1 2 2 )1 .3 5 0 5 to 1 .3 6 9 0 (1 2 2 )1 .3 6 9 0 to 1 .3 9 4 0 (1 2 2 )1 .3 9 4 0 to 1 0 .0 0 0 0 (1 2 3 )O th e r (1 7 9 )
Delivered Price – Credit & Finance
1994
COMPOSITE PRICE - CREDIT & FINANCE
1994
PRODUCTION COSTS - RETAIL TRADE
1994
COMPOSITE PRICE - RETAIL TRADE
1994
PRODUCTION COSTS - MOTOR VEHICLES
C AL IF O R NIA
N E W Y O RK
TE X A S
FL O RID A
P E N NS Y L V AN IA
ILL INO ISO H IO
M ICH IG AN
N O RT H C AR O LIN A
G E O R G IA
V IR G IN IA
M A
IN
M IS S O UR I
W IS CO NS IN
TE N N E S S E E
W AS H IN G TO N
M AR Y LA ND
M INN E S O TA
LO UIS IA NA
A L
K E NT UC KY
A RIZO NA S C
C O LO R AD O
O K LA HO M A
O R E G O N
IO W A
M S
K AN S AS
A RK AN S AS
U TA HN E BR AS K A
N E W M E X IC O
M AIN E
N E V AD A
N HID AH O
M O NT AN A
S O U TH DA KO T A
N D
W Y O M ING
M exico
B ah amas
B erm ud a
0 2 0 0 4 0 0
M i le s
Map LayersC o u n tr y (L o w R e s )S ta te (L o w R e s )C o u n ty (L o w R e s )
PACT3710 .0 0 0 0 to 0 .9 2 2 5 (1 2 2 )0 .9 2 2 5 to 0 .9 5 2 6 (1 2 2 )0 .9 5 2 6 to 0 .9 6 9 3 (1 2 2 )0 .9 6 9 3 to 0 .9 8 0 3 (1 2 2 )0 .9 8 0 3 to 0 .9 8 9 1 (1 2 2 )0 .9 8 9 1 to 0 .9 9 6 5 (1 2 2 )0 .9 9 6 5 to 1 .0 0 3 3 (1 2 3 )1 .0 0 3 3 to 1 .0 1 1 4 (1 2 2 )1 .0 1 1 4 to 1 .0 1 8 2 (1 2 2 )1 .0 1 8 2 to 1 .0 2 4 9 (1 2 2 )1 .0 2 4 9 to 1 .0 3 1 4 (1 2 2 )1 .0 3 1 4 to 1 .0 3 6 9 (1 2 2 )1 .0 3 6 9 to 1 .0 4 3 9 (1 2 3 )1 .0 4 3 9 to 1 .0 5 1 0 (1 2 2 )1 .0 5 1 0 to 1 .0 5 7 9 (1 2 2 )1 .0 5 7 9 to 1 .0 6 5 0 (1 2 2 )1 .0 6 5 0 to 1 .0 7 1 8 (1 2 2 )1 .0 7 1 8 to 1 .0 8 0 7 (1 2 2 )1 .0 8 0 7 to 1 .0 9 1 0 (1 2 3 )1 .0 9 1 0 to 1 .1 0 1 7 (1 2 2 )1 .1 0 1 7 to 1 .1 1 1 4 (1 2 2 )1 .1 1 1 4 to 1 .1 2 4 0 (1 2 2 )1 .1 2 4 0 to 1 .1 4 0 3 (1 2 2 )1 .1 4 0 3 to 1 .1 5 8 8 (1 2 2 )1 .1 5 8 8 to 1 0 .0 0 0 0 (1 2 3 )O th e r (1 7 9 )
Delivered Price – Motor Vehicles
1994
COMPOSITE PRICE - MOTOR VEHICLES
1994
PRODUCTION COSTS - TEXTILES
O K LA HO M A
1994
COMPOSITE PRICE - TEXTILES
1994
PRODUCTION COSTS - CHEMICALS
1994
COMPOSITE PRICE - CHEMICALS
New EG FeaturesTransportation Costs Matrix
The transportation cost matrix is a new component of the economic geography model of REMI Policy Insight. There are three components of the matrix:
– Commuting costs
– Accessibility costs
– Transportation costsThe three components respond to changes in the effective distance between regions.
Effective DistanceEffective Distance is Derived as Follows
• Capture all trade flows among all counties.• Back calculate the distance (“effective distance”)
which yields the observed trade flows, given the gravity model’s distance decay parameter (β).
• The β parameter by economic sector is the factor which best explains a change in regional output, given a change in demand. Thus, the β parameter drives the regional purchase coefficient (RPC), which is the ratio of self-supply to demand for each economic sector.
Commuting Costs
Commuting costs affect the effective distance in the occupational labor access productivity equation (by adjusting labor cost relative to the average wage rate), which feeds into industry labor access productivity and from there into composite wage rates, ultimately affecting both the composite cost of production by industry and the relative wage rate in the economic migration equation.
Accessibility Costs
Accessibility costs affect the effective distance in the intermediate input access index, which reflects a price elasticity of demand (price is sensitive to distance) and which feeds into the moving average commodity access index. The commodity access index feeds the composite input cost equation and the consumer access equation, ultimately affecting both the composite cost of production by industry and the consumption access index in the economic migration equation.
Transportation Costs
Transportation costs affect the effective distance in the delivered price equation for each sector (relative to time and to wage rates, or only to time if wage rates are unchanged), which feeds into the composite input cost equation to producers by sector and also into the consumption equation to consumers by commodity, ultimately affecting both the composite cost of production by industry and the consumption access index in the economic migration equation.
Matrix Input Internal UsageEffective Distance Equations
• Intermediate Input Access Index (Eq. 1-4)
• Labor Productivity by Occupation Due to Relative Labor Access by Occupation, and
Relative Labor Productivity Due to Industry
Concentration (Eq. 2-1a, 2-1b: Commuting)
• Delivered Average Price (Eq. 4-3: Transport)
• Share of Domestic Demand Supplied (Eq. 5-1)
Multi-Regional Price and Wage Linkages
Commodity Access
Commodity Access
Commodity Access
At market impedance estimated costs based on
dynamically estimated price elasticity
Occupational Labor Access Productivity
Composite Labor Costs
Industry Labor Access Productivity
Occupational Labor Access Productivity
Composite Labor Costs
Industry Labor Access Productivity
Occupational Labor Access Productivity
Composite Labor Costs
Industry Labor Access Productivity
Based on commuting in hours per day at one-half of the daily wage
Delivered Prices and Costs
Delivered Prices and Costs
Delivered Prices and Costs
Based on estimated transportation costsComposite and
Production Costs
Composite and Production Costs
Composite and Production Costs
Matrix Input ProcessingStandard Policy Variables vs Matrix Inputs
• Standard Policy VariablesIn most simulations, inputs are made by sector or demand source based on how each direct shock will affect each part of the economy as an element within the overall shock being assessed.
• Matrix InputsEconomic Geography matrix inputs for Commuting, Accessibility, and Transportation work their way through all sectors and regions in a set-up process before participating in the simulation itself.
Matrix Response SensitivityAssuming 1% Change in Effective Distance
Elasticity is inverse to the degree to which:• Region or state is compact in size and shape.• Supply chain and labor sources are remote from the
region and from nearby compact regions or states.• Shock is focused (i.e., degree to which shock applies
a matrix change to one directional effective distance concept rather than simultaneously to intra- and inter-distances).
Economic Geography Conclusion
A Forecasting and Policy Analysis Model that
includes product differentiation and distance
deterrence costs will enable the user to incorporate
endogenous agglomeration effects on productivity and
on consumer utility in the user’s forecasts and policy
analyses. This should open the way to more realistic
and insightful simulations for use in regional
economic policy analysis. – George Treyz
REMI’s Objective
Economic theory
Economic theory
Policy application
Policy application