Bar-El Inter-regional equilibrium

Inter-regional labor
market equilibrium:
another pattern of
spatial mismatch
objective
► Make
a critical assessment of the “spatial
mismatch hypothesis” and
► Design a general equilibrium model for interregional labor market equilibrium,
 To explain the existence of inter-regional imbalances in
labor demand,
 To explain the existence of differences in types of
regional economic development patterns
 To clarify the role of commuting and migration
 To evaluate the role of the public sector.
The spatial mismatch hypothesis
(Kain 68, Kain 94)
Job dispersal from Center (C) to Suburbs (S).
► Richer (White) migrate from C to S.
► Poorer (Black) stay in C in spite of increasing
labor demand in S, as a result of “housing
segregation” or “income segregation”: gaps
between C and S.
► Public policy solution: compensatory measures
such as subsidies for relocation (Smith and
Zenou, 2003).
►
Some additions and empirical testing
(Martin 2004, Zenou and Brueckner 2000, Rogers 97, Taylor and Ong 95,
Eliasson and Lindgren 2003, Romani, Surinach and Artis 2003, Elhorst 2003,
Renkow 2003)
►
►
Poor (Black) do not migrate, but
alternatively they can commute to jobs in
S. Commuting costs discriminate against
them.
Mostly unskilled, high relative commuting
cost, leading to an increase in the gap
between C and S.
Lack of a model
► No
model that explains a causality
relationship (Arnott, 98)
► Job dispersal should not necessarily lead to
spatial mismatch, even given mobility
constraints.
► Job dispersal or suburb growth cannot be
considered as exogenous.
A general equilibrium model
Following first steps by Arnott (98).
Job dispersal is not exogenous: explained by
agglomeration economies in two regions.
► Existence of a third regional dimension.
► Labor demand behavior depends on types of activities:
non-tradables, tradable manufacturing, tradable services.
► Reject the assumption of “people follow jobs”: migration
also function of housing cost.
► Existence of endogenous growth, through market induced
demand (non-tradables).
► Reject the assumption of dichotomic behavior of migration:
replace segregation with migration cost.
► Public policy as an exogenous variable.
►
►
►2
regions: C (center: the Tel-Aviv district)
and S (Suburbs: the Central district).
► 3 economic sectors: NT, TM, TS.
► Labor demand:
s
s
LDrs  AEr LCr Pr
s
 s ,  s  0,
s  0
Function of agglomeration economies, land cost, and population.
Separate estimation for each region, each sector
Agglomeration economies
AEr  1  e
Increasing return to
agglomeration at
first stage,
decreasing return
later.
 s AG2
A
E
Separate function
for each economic
sector.
A
G
Land cost
 Pr  LDr
LC r  GOV ( LC ) r * 
Ar





Depends on density
of population and of
economic activity
Influenced by
government
Commuting, or inter-regional labor supply
s
s
CCS ,s  ( LDC / LDS ) s * CCCS
Function of relative labor supply in regions, and of
commuting cost, for each specific economic sector
CCCS  DIS * GOV (CC)
Communication costs depend on distance and on
government policy
Migration
M CS ,s  lc * ( LCC / LC S )  ld s * CCCS * ( LDC ,s  LDS ,s )  MC s
Depends on differences in land prices, labor demand
weighted by communication costs, migration cost:
1. Potential for “housing cost driven migration” (as in the
periphery), and especially if cost of commuting is low (as
in suburbs): complementarity (not just trade-off) between
commuting and migration.
2. Migration costs may reflect various levels of “housing
segregation” (Beduins) or of income segregation.
Labor supply
LS r .s  p * Pr * Er
Er  er * GOV ( E) r
Depends on participation
rates, population and
education (skills)
Quality of labor force
influenced by
government policy
Unemployment
UEC , s  LS C , s  LDC , s C SC , s C CS , s
Population
PC  IPC * (1  NGC )  M SC  M CS
Potential inter-regional labor
market patterns
1. Stabilization of
the process at the
Center-Suburbs
level, ignition of a
new process at the
Suburbs-Periphery
level
Table 1: annual growth of labor
demand
Region total 77-90 91-96 97-02
C 1.8% 1.5% 4.3% 0.3%
S 3.9% 2.0% 6.9% 4.9%
P 3.9% 2.5% 8.7% 2.4%
2. Labor demand qualities: focus in C on non-tradables but also
in knowledge intensive services with strong agglomeration
economies, high tech manufacturing to S, non-tradables and
manufacturing to P (increasing location quotients).
Table 2: annual growth of labor supply
3. Slow growth of labor
supply in C, high growth
in S, but stabilization of
growth rate in last
years: effect of shape of
agglomeration
economies curve.
4. P still far from
agglomeration
economies: slowing
growth.
5. Strong migration to S,
gaps between labor
demand and supply, but
no growth in
unemployment.
total
C 1.5%
S 4.0%
P 4.4%
Region
77-90
1.2%
3.2%
3.1%
91-96
2.9%
4.8%
8.5%
97-02
0.9%
4.9%
3.3%
Table 3: average annual migration balance
per thousand population
Region
total 77-90 91-96 97-02
C
-8
-3
-15
-13
S
9
7
7
14
P
0
-5
13
0
6. In spite of decrease share
of C in labor demand,
increasing commuting into C:
not depletion, rather
concentration in specific
economic role.
7. Steady commuting out
from S, in spite of increasing
growth: regional economic
integration.
8. No response yet of P to
changes in labor supply and
to increasing unemployment:
commuting not yet an
appropriate solution.
Table 4: average share of out commuters
from region out of labor supply
Region
Total
77-90 91-96 97-01
C
14%
13%
13%
16%
S
30%
30%
30%
30%
P
12%
12%
12%
12%
Table 5: average share of in commuters
to the region out of labor demand
Region
Total
77-90 91-96 97-01
C
24%
21%
27%
32%
S
18%
17%
19%
21%
P
6%
6%
6%
7%
conclusion
► First
steps for a model explaining the interregional behavior of labor market.
► “people follow jobs”: not necessarily true.
► Commuting as a complement that enables
migration.
► Decreasing returns to agglomeration lead to a
stable equilibrium.
► Public policy is crucial for the solution of
bottlenecks (investments in communication,
infrastructure, education)