Transcript SIMULATION MODELS - Universidad Autonoma de Madrid

Macro-economic Impact of
Migrations in Madrid Region
INTRODUCTION
Centre of Migration Research,
Warsaw, November 2007
Dr. Mahía & Dr. de Arce
Professors Univ. Autónoma de Madrid
WHY DO POLICY-MAKERS NEED MODELS?
Reference paper: Piermartini, R. and Teh, R. (2005): “Demystifying Modelling Methods for
Trade Policy”. WTO Discussion Papers, num. 10. September, 2005.
• Economic models provide a theoretically
consistent, rigorous and quantitative way of
evaluating different policy analysis
• A simulation of the model can confirm that
judgment and provide estimate of the likely gains
• Models can alert about effects in inter-related
issues
• But, what’s first?
EXPONENTIAL GROWTH
 Advances in theoretical analysis
 Largeness and wideness Statistical collection
 Increase of Computational power
STRENGTHS
 Reflects Inter-dependency of economic variables
 Computer based models allow us to track all of
these interactions
 Simulation of several scenarios is available
 Simulations are performed in a transparent
hypothesis frame
 They discipline thinking about how economies
actually work
Reference paper: Piermartini, R. and Teh, R. (2005): “Demystifying Modelling Methods for Trade
Policy”. WTO Discussion Papers, num. 10. September, 2005.
LIMITATIONS
Aggregations can obscure important underlying
relationships
Data are not always “high quality”
Responsiveness of supply and demand to price
changes are not necessarily accurate
Choices among scenarios and model
specification can imply very different results
Technical comprehension is only available for
experts
Reference paper: Piermartini, R. and Teh, R. (2005): “Demystifying Modelling Methods for Trade
Policy”. WTO Discussion Papers, num. 10. September, 2005.
CHOICE THE MODEL
 COMPARATIVE STATIC AND DYNAMIC ANALYSIS
• Static: changes in policies and direct transmission to
endogenous variables of model
• Dynamic: final and intermediate process of change
 PARTIAL OR GENERAL EQUILIBRIUM ANALYSES
• “Ceteris Paribus” criteria
• Whole economy linkages
KIND OF THECNICAL STRATEGIES
FREQUENTLY USED IN INTERNATIONAL
SIMULATIONS
• Traditional Econometrical Multiequational
Models
• Optimization Models: Izquierdo, M., Jimeno, JF. and Rojas,
JA (2007): “On the Aggregate Effects of Immigration in Spain”. Banco
de España Working Papers.
• Gravitational Models
• Panel Data Models
• Leontief Based Models (I-O Models): Vicens et
al. (2005): “Impacto Macroeconómico de la Inmigración en la
Comunidad de Madrid”.
• Social Account Matrix Models
Inmigration Model “Comunidad de
Madrid 2005”
STARTING POINT
• More than 12% of Employed population are
immigrants in Comunidad de Madrid
• Majority of immigrants have come to CM in last
five years
• Estimation of economic aspects of
immigration in Spanish Comunidad de Madrid
in last five years
Immigration CM Model Schema
FORMAL ECONOMY
TOTAL
IMMIGRANTS
ESTIMATION
EMPLOYMENT
& WAGES
CONSUMPTION
BASKET
INFORMAL ECONOMY
PADRÓN
REGULARIZ. LAW
DELPHI
BRANCHES DESEGREGATION
PUBLIC
SOURCES
ASSOCIATIONS
EMBASSY
C
O
E
F.
E
M
P
L
O
Y
M
E
N
T
ACADEMICALS
▲ INPUT
PRODUCTION
& DEMAND
TIO
Direct Effect
Indirect Effect
▲ OUTPUT
HEALTH
EDUCATIONAL
OTHER PUBLIC
EXPENSES
REMITTANCES
BRANCHES
DISTRIBUTION OF
CONSUMPTION
YIELD
NEW DOMESTIC
EMPLOYMENT
GENERATION
RATIO CONSUMPTION/INVESTMENT
IMMIGRANTS CONSUMPTION
STRUCTURE
OTHER
Main Outputs in CM Model
1)
2)
3)
4)
5)
6)
7)
8)
Estimation of the Total Number of Immigrants in “Comunidad de
Madrid”
Estimation of wages in immigrant population (sector distribution
and formal and informal differences)
Simulation Model: An unrealistic Economy without Immigrants –
Gosh I-O Model
Productivity aspects in the whole economy
Firms Margins derived from immigrants activities
Indirect Effects in domestic employment derived from immigrants
activities (production and consumption)
Public sector issues about Immigration: revenues and
expenditures. Vital Cycle considerations in Fiscal Balance
Remittances of immigrants estimation: domestic financial sector
implications and remittances destination considerations
Ghosh Supply Model (I)
Distribution Coefficients: delivers of
sector “i” to sector “j” over total Outputs of
sector “i”. Sells structure of each sector
(interdependence between sectors)
dij 
xij
Tot _ Outputi
Ghosh Supply Model (and II)
Followed from previous coefficients and in
the same terms that Leontief Model, we
can build one expression to impact from
Value Added to Total Production:
x' = VA(I-D)-1 +[(Tr+M+Taxes)*D](I-D)-1
Value Added and Employment Generation from
Ghosh Production: Static vs. Dynamic
coefficients of VA and Employment
Coef .VA j 
VA jañoTIO
Pr od j1añoTIO
VA jt  prodjt * Coef .VA j
Coef .Empjt 
Empljt
V . A. jt
Empljt  V .A. jt * Coef .Empljt
I-O Ghosh Model Implementation: Production Effect
CHANGES IN VALUE ADDED VECTOR
ESTIMATION OF EFFECT ON TRANSFERS, IMPORTS & TAXES
DIRECT
PRODUCTION
EFECT
DISTRIBUTION BY BRANCHES
(Matrix of Distribution Coefficients)
TOTAL
PRODUCTON
EFECT
Δ INTERMEDIATE
PRODUCTION
(Ghosh Inverse (I-D)-1)
Δ PRODUCTION
Δ ADDED VALUE
COEFFICIENTS
TRANSLATION
Δ EMPLOYMENT
Distribution Matrix D
x' = VA(I-D)-1 +[(Tr+M+Taxes)*D] (I-D)-1
VA jt  prodjt * Coef .VA j
Empljt  V .A. jt * Coef .Empljt
Leontief I-O Model Implementation: Demand Efect
Δ TOTAL PRODUCTION
GROSS SALARIES
TAXES
YIELD
SAVE
Δ CONSUMPTION YIELD
DIRECT DEMAND
EFECT
INDIRECT
DEMAND EFECT
Δ EMPLOYMENT
DISTRIBUTION BY
BRANCHES
Δ INTERMEDIATE
DEMAND
(Leontief Inverse)
TOTAL DEMAND EFECT:
DIRECT + INDIRECTS
Δ PRODUCTION / A.V
Δ PRODUCTION / A.V.
Δ EMPLOYMENT
Δ PRODUCTION / A.V.
Δ EMPLOYMENT
MAIN RESULTS
Salaries & Wages
Margins & Benefits
Production Taxes
VA (Prod. Effect)
VA (Induced
Demand Effect)
Total VA CM
Total
Industry
Building
Market
Services
4.863.137
3,27%
7.334.185
4,93%
100.207
0,07%
12.297.529
8,27%
65.645
0,04%
44.717
0,03%
240
0,00%
110.602
0,07%
1.854.759
1,25%
987.676
0,66%
43.398
0,03%
2.885.833
1,94%
2.883.087
1,94%
6.297.398
4,24%
56.543
0,04%
9.237.027
6,21%
Non
Market
Services
59.647
0,04%
4.394
0,00%
26
0,00%
64.067
0,04%
2.085.248
1,40%
9,68%
351.908
0,24%
0,31%
122.063
0,08%
2,02%
1.609.920
1,08%
7,30%
1.119
0,00%
0,04%