Spatial Economics

Stephen Redding

Princeton University

1 / 35
 Essential Reading

• Fujita, M., P. Krugman, and A. Venables (1999) The Spatial

Economy: Cities, Regions and International Trade, MIT Press,
Chapters 4, 5 and 14.

• Allen, Treb and Costas Arkolakis (2014) “Trade and the Topography
of the Spatial Economy,” Quarterly Journal of Economics, 129(3),
1085-1140.

• Redding, S. and D. Sturm (2008) “The Costs of Remoteness:

Evidence from German Division and Reunification,” American
Economic Review, 98(5), 1766-1797.

• Redding, S. (2016) “Goods Trade, Factor Mobility and Welfare,”

Journal of International Economics, 101, 148-167, 2016.

2 / 35
 Motivation
• Until the early 1990s, economic geography received relative little
attention in mainstream economic theory

• Despite the fact that production, trade and income are distributed
extremely unevenly across physical space

• Agglomeration of overall economic activity most evident in cities

– In 2016, 54.5 per cent of the world?s population lives in urban areas,
a proportion that is expected to increase to 66 per cent by 2050
– In 2016, 31 “megacities” with a population > 10 Million
– Of these megacities, 24 are located in less developed countries, and
China is home to 6

• Geographical concentration of particular activities

– US manufacturing belt in NE and Eastern Midwest
– Dalton as a carpet manufacturing centre in Georgia
– Silicon Valley and Route 128 in Massachusetts

3 / 35
 Motivation

• What do we mean by economic geography?

– Location of economic activity in space

• First-nature geography
– Physical geography of coasts, mountains and endowments of natural
resources

• Second-nature geography
– The spatial relationship between economic agents

• Our analysis will largely focus on second-nature geography

– How does the spatial relationship between agents determine how
they interact, what they do, and how well off they are?

4 / 35
 Agglomeration Forces
• This lecture introduces Krugman (1991) and Helpman (1998)
– Love of variety, Increasing returns to scale and trade costs

• Marshall (1920) identified three forces for agglomeration

– Market for workers with specialized skills
– Provision of non-traded inputs in greater variety and lower cost
– Technological knowledge spillovers

• Krugman (1991) and Helpman (1998) focus on pecuniary rather

than technological externalities
– Love of variety, Increasing returns to scale and trade costs
(forward & backward linkages)
– Mobile workers (more relevant within than across countries)
– Krugman (1991): immobile agricultural laborers are dispersion force
– Helpman (1998): immobile land is dispersion force

• Krugman and Venables (1995) develop a model of agglomeration

with immobile workers through the introduction of trade in
intermediate inputs
5 / 35
 Helpman (1998)

• Economy consists of N of regions indexed by n

• Each region is endowed with an exogenous quality-adjusted supply of

land (Hi )

• Economy as a whole is endowed with a measure L̄ of workers, where

each worker has one unit of labor that is supplied inelastically with
zero disutility

• Workers are perfectly geographically mobile and hence in equilibrium

real wages are equalized across all populated regions.
• Regions connected by goods trade subject to symmetric iceberg
variable trade costs
– where dni = din > 1 units must be shipped from region i for one unit
to arrive in region n 6= i
– where dnn = 1

6 / 35
 Preferences

• Preferences are defined over goods consumption (Cn ) and residential

land use (HUn )

HUn 1−α
 α  
Cn
Un = , 0 < α < 1.
α 1−α

• Goods consumption index (Cn ) is defined over the endogenous

measures of horizontally-differentiated varieties supplied by each
region (Mi ) with dual price index (Pn ):
" #1 " # 1
Z Mi ρ Z Mi 1− σ
Cn = ∑ ρ
cni (j ) dj , Pn = ∑ pni (j ) 1− σ
dj .
i ∈N 0 i ∈N 0

7 / 35
 Production

• Varieties produced under conditions of monopolistic competition and

increasing returns to scale.
• To produce a variety, a firm must incur a fixed cost of F units of
labor and a constant variable cost in terms of labor that depends on
a location’s productivity Ai .

xi ( j )
li (j ) = F + .
Ai
• Producer of each variety chooses prices to maximize profits subject
to its downward-sloping demand curve
  
xi ( j )
max pi (j )xi (j ) − wi F + .
pi ( j ) Ai

8 / 35
 Profit Maximization and Zero-Profits
• First-order condition for profit maximization implies that prices are a
constant markup over marginal cost
 
σ wi
pi (j ) = pi = ,
σ−1 Ai

pni (j ) = pni = dni pi

• Profit maximization and zero profits implies that equilibrium output
of each variety depends solely on parameters

xi (j ) = x̄i = Ai (σ − 1)F

• Using the production technology, equilibrium employment for each

variety also depends solely on parameters

li (j ) = l¯ = σF .

9 / 35
 Labor Market Clearing

• Labor market clearing requires demand equals the supply for labor

Li = Mi l¯

• Therefore the mass of varieties produced by each location is

proportional to its supply of labor
Li
Mi = .
σF

10 / 35
 Price Indexes

• Using the symmetry of equilibrium pricing, the price index is:

" # 1
1− σ
Pn = ∑ 1− σ
Mi pni
i ∈N

• Using labor marker clearing and the pricing rule, we have:

  1 "  1−σ # 1−1 σ

σ 1 1− σ wi
Pn =
σ−1 σF ∑ Li dni
Ai
.
i ∈N

11 / 35
 Expenditure Shares

• Equilibrium expenditure shares

 1− σ
1− σ
Mi pni Li dni w
Ai
i

πni = 1− σ
=  1− σ .
∑k ∈N Mk pnk

w
∑k ∈N Lk dnk Ak k

• Using the denominator of the expenditure share, the price index can
be re-written as:
  1
σ Ln 1− σ wn
Pn = .
σ−1 σF πnn An

12 / 35
 Land Market Clearing

• Expenditure on land in each location is redistributed lump sum to

the workers residing in that location

• Per capita income in each location (vn ) equals labor income (wn )
plus per capita expenditure on residential land ((1 − α)vn ):

wn Ln
vn Ln = wn Ln + (1 − α)vn Ln = .
α

• Land market clearing implies that the equilibrium land rent (rn ) can
be determined from the equality of land income and expenditure:

(1 − α)vn Ln 1 − α w n Ln
rn = = ,
Hn α Hn

13 / 35
 Population Mobility

• Population mobility implies that workers receive the same real

income in all populated locations:
vn
Vn = = V̄ .
Pn rn1−α
α

• Using the price index, income equals expenditure, and land market
clearing in the population mobility condition, we obtain:
σ (1−α)−1
−α/(σ−1) −
Aα H 1−α πnn Ln σ − 1
V̄ = n n
σ

α 1 1− σ 1− α  1− α
  α 
α σ− 1 σF α

14 / 35
 Gains from Trade and Market Access

• A region’s welfare gains from trade depend on the change in its

domestic trade and share and the change in its population
σ (1−α)−1
VnT  − α  LT − σ−1
T σ −1 n
= πnn .
VnA LA
n

• Rearranging the population moblity condition to obtain an expression

for Ln , and dividing by total labor supply L̄ = ∑n∈N Ln , population
shares depend on productivity, land supply and market access
h i σ −1
A α H 1−α π −α/(σ −1) σ(1−α)−1
Ln n n nn
λn = = i σ −1 ,
L̄ h
α H 1−α π −α/(σ −1) σ(1−α)−1
∑ k ∈N k k
A kk

• where market access summarized by the domestic trade share (πnn )

15 / 35
 General Equilibrium

• General equilibrium : two systems of equations across locations

– Gravity of trade flows
– Population mobility

16 / 35
 Population Mobility
vn
Pnα = .
V̄ rn1−α
• Using vn = wn /α and land market clearing (rn = 1−
α
α wn Ln
Hn )

 1− α "  #1
1 − α 1− α
  α
wn Hn α
Pn = , W̄ ≡ α V̄ .
W̄ Ln α
• Recall
  1 "  1−σ # 1−1 σ
σ 1 1− σ wi
Pn =
σ−1 σF ∑ Li dni
Ai
.
i ∈N

• Obtain a first wage equation from population mobility

  (1− σ ) 1− α
wi1−σ H
  1− σ i α
1 σ Li
W̄ 1−σ = 1− σ .
σ−1

σF w
∑n∈N Ln din Ann

17 / 35
 Gravity I

• Gravity and income equals expenditure implies:

 1− σ
Li σ wi
σ −1 dni Ai
∑
σF
wi Li = w n Ln .
n ∈N Pn1−σ

• Recall that the price index can be expressed as

  1− σ
Ln σ wn
σF σ − 1 An
Pn1−σ = .
πnn
• Obtain a second wage equation from gravity

wiσ A1i −σ = ∑ 1− σ σ 1− σ
πnn dni wn An .
n ∈N

18 / 35
 Gravity II
• Recall price index from previous slide
  1− σ
Ln σ wn
σF σ − 1 An
Pn1−σ = .
πnn
• Recall price index from population mobility

 1− α "  #1
1 − α 1− α
  α
wn Hn α
Pn = , W̄ ≡ α V̄ .
W̄ Ln α

• Equating these two expressions, we obtain the following solution for

the domestic trade share
  1− σ
1 σ 1−(σ−1) 1−α α (σ−1) 1−α α σ−1
πnn = W̄ 1−σ Ln Hn An .
σF σ − 1

19 / 35
 Gravity III

• Recall our earlier wage equation from gravity

wiσ A1i −σ = ∑ 1− σ σ 1− σ
πnn dni wn An .
n ∈N

• Using our expression for the domestic trade share on previous, slide
this wage equation from gravity becomes
 1− σ
wiσ A1i −σ

1− σ 1 σ
W̄ = .
σF σ−1 1− σ 1−(σ−1) 1−α α (σ−1) 1−α α
∑n∈N dni Ln Hn wnσ

20 / 35
 General Equilibrium I

• Two systems of equations for wages and population

  (1− σ ) 1− α
wi1−σ H
  1− σ i α
1 σ Li
W̄ 1−σ = 1− σ .
σ−1

σF w
∑n∈N Ln din Ann
 1− σ
wiσ A1i −σ

1− σ 1 σ
W̄ = .
σF σ−1 1− σ 1−(σ−1) 1−α α (σ−1) 1−α α
∑n∈N dni Ln Hn wnσ
• Under our assumption of symmetric trade costs (dni = din ), this
system of equations has the following closed-form solution
(σ−1) 1−α α −(σ−1) 1−α α
wn1−2σ Anσ−1 Ln Hn = φ.

21 / 35
 General Equilibrium II

• Using this closed-form solution, we obtain a single system of

equations that determines equilibrium population

σ̃γ −σ̃(σ−1) −σ̃σ 1−α α

Ln 1 An Hn =
  1− σ   γ2
1 σ σ̃ (σ −1) 1− α
W̄ 1−σ ∑
σ̃γ γ
dni Li 1 1 Aiσ̃σ Hi α
,
i ∈N
σF σ−1

σ−1
σ̃ ≡
,
2σ − 1
1−α
γ1 ≡ σ ,
α
σ 1−α
γ2 ≡ 1 + − ( σ − 1) .
σ−1 α

22 / 35
 Existence and Uniqueness
Proposition
Assume σ (1 − α) > 1. Given the land area, productivity and amenity
parameters {Hn , An , Bn } and symmetric bilateral trade frictions {dni } for
all locations n, i ∈ N, there exist unique equilibrium populations (Ln∗ )
that solve this system of equations.

Proof.
The proof follows Allen and Arkolakis (2014). Assume σ (1 − α) > 1.
Given the land area and productivity for each location {Hn , An } and
bilateral trade frictions {dni }, there exists a unique fixed point in this
system because γ2 /γ1 < 1 (Fujimoto and Krause 1985).

• Intuition: Unique equilibrium requires that agglomeration forces are

sufficiently weak relative to dispersion forces
– Higher (1 − α) implies that land accounts for a larger share
consumer expenditure
– Higher elasticity of substitution (σ) implies that varieties are closer
substitutes for one another
23 / 35
 Market Access

• Model provides micro-foundations for a theory-consistent measure of

market access
– Ad hoc measures of market potential following Harris (1954)

Lit
MPnt = ∑ dist ni
i ∈N

– Theory-based measure highlights the role of price indexes

(connection with Anderson and Van Wincoop 2003)

• We now examine the predictions of the model for the equilibrium

relationship between wages, population and market access

• Market access is itself an endogenous variable

24 / 35
 Wages and Market Access
• From profit maximization
 
σ wi
pi (j ) = pi = ,
σ−1 Ai

• From profit maximization and zero profits

xi (j ) = x̄i = Ai (σ − 1)F

• From CES demand and market clearing, we have:

x̄i = piσ ∑ 1− σ
dni (wn Ln ) (Pn )σ−1 ,
n ∈N

25 / 35
 Wages and Market Access
• Combining profit maximization, zero profits, CES demand and
market clearing, we obtain the following wage equation:

σ wi σ
 
1
= FMAi
σ − 1 Ai x̄i
 σ −1
σ−1
 σ −1
σ
− σ1 1
wi = Ai σ
(l¯) (FMAi ) σ .
σ
• where firm market access is defined as

FMAi ≡ ∑ 1− σ
dni (wn Ln ) (Pn )σ−1 ,
n ∈N

• Wages increase in productivity Ai and firm market access (FMAi )

• Reductions in transport costs (dni ) increase firm market access and
wages (wi )
• For empirical evidence, see Dekle and Eaton (1999), Hanson (2005),
Donaldson and Hornbeck (2016)
26 / 35
 Price Indexes and Market Access

• Market access also affects the price index, which depends on

consumers’ access to tradeable varieties

• We summarize this access to tradeable varieties using consumer

market access (CMAn ):
1
Pn = (CMAn ) 1−σ ,

CMAn ≡ ∑ Mi (dni pi )1−σ ,

i ∈N

• where we use symmetric trade costs

27 / 35
 Estimating Market Access
• Use international trade data to “reveal” market access
• CES demand function

Xni = Mi pi xni = Mi pi1−σ dni

1− σ
Xn Pnσ−1

• Can be re-written in terms of market and supply capacity

1− σ
Xni = si dni mn ,

si ≡ Mi pi1−σ , mn ≡ Xn Pnσ−1 ,
• Firm and consumer market access can be written:

FMAi = ∑ 1− σ
dni mn , CMAn = ∑ si dni1−σ
n ∈N i ∈N

• where we again use symmetric trade costs

• Redding and Venables (2004) estimate these market access
measures using fixed effects gravity equation estimation
28 / 35
 Population and Market Access
• Population mobility implies:
vn
Vn = = V̄ .
Pn rn1−α
α

• Using income equals expenditure, the price index and land market
clearing, together with the definitions of firm and consumer market
access, this population mobility condition can be written as:

( α σ −1 ) α α
Ln = χAn 1−α σ Hn (FMAn ) (1−α)σ (CMAn ) (1−α)(σ−1) ,

• Equilibrium population is increasing in productivity, housing supply,

and firm and consumer market access
• For evidence, see Redding and Sturm (2008)
• In our earlier GE system of equations, we solved for equilibrium
population as a function of the exogenous variables

29 / 35
 Firm and Consumer Market Access
• Firm and consumer market access are closely related
• From the definitions of firm and consumer market access:

FMAi ≡ ∑ 1− σ
dni (wn Ln ) CMAn−1 .
n ∈N

• From the definition of consumer market access and using profit

maximization and labor market clearing, we have:
  1− σ
Li σ wi
CMAn ≡ ∑ σF σ − 1 Ai
dni .
i ∈N

• Note that CES demand implies the following gravity equation for
bilateral exports from location i to location n:
  1− σ
Li σ wi 1− σ
Xni = dni (wn Ln ) CMAn−1 ,
σF σ − 1 Ai
• where we have used Xn = αvn Ln = wn Ln .
30 / 35
 Firm and Consumer Market Access
• Summing across destinations and using goods market clearing
(Xi = ∑n∈N Xni = wi Li ), we obtain:

σ wi 1 − σ
 
Li
wi Li =
σF σ − 1 Ai ∑ dni1−σ (wn Ln ) CMAn−1 .
n ∈N

• Using the definition of firm market access, this relationship implies:

  1− σ
Li σ wi w i Li
= .
σF σ − 1 Ai FMAi
• Using this result in the expression for consumer market access, we
obtain:
CMAn ≡ ∑ dni1−σ (wi Li ) FMAi−1 ,
i ∈N
• which reversing the notation becomes:

CMAi ≡ ∑ 1− σ
din (wn Ln ) FMAn−1 .
n ∈N

31 / 35
 Firm and Consumer Market Access

• Assuming symmetric trade costs (dni = din ), any solution to these

two systems of equations must satisfy:

MAi ≡ FMAi = ψCMAi ,

• as can be confirmed by using this relationship under symmetric trade

costs in these equations to obtain the recursive solutions:

FMAi ≡ ∑ 1− σ
dni (wn Ln ) ψ−1 FMAn−1 ,
n ∈N

FMAi ≡ ∑ 1− σ
dni (wn Ln ) ψ−1 FMAn−1 ,
n ∈N
• where we determined wn and Ln from our GE system above
• Therefore with symmetric trade costs firm and consumer market
access can be reduced to a single market access measure

32 / 35
 References
• Allen, Treb and Costas Arkolakis (2014) “Trade and the Topography
of the Spatial Economy,” Quarterly Journal of Economics, 129(3),
1085-1140.
• Anderson, James and Eric van Wincoop (2003) “Gravity with
Gravitas: A Solution to the Border Puzzle,” American Economic
Review, 93(1), 170-192.
• Dekle, Robert and Jonathan Eaton (1999) “Agglomeration and Land
Rents: Evidence from the Prefectures,” Journal of Urban
Economics, 46(2), 200-214.
• Donaldson, D., and R. Hornbeck (2016) “Railroads and American
Economic Growth: A Market Access Approach,” Quarterly Journal
of Economics, 131(2), 799-858.
• Fujimoto, T. and U. Krause (1985) “Strong Ergodicity for Strictly
Increasing Nonlinear Operators,” Linear Algebra and its
Applications, 71, 101-112.

33 / 35
 References

• Harris, Chauncy D. (1954) “The Market as a Factor in the

Localization of Industry in the United States,” Annals of the
Association of American Geographers, 44(4), 315-348.
• Hanson, Gordon H. (2005) “Market Potential, Increasing Returns,
and Geographic Concentration,” Journal of International Economics,
67(1), 1-24.
• Helpman, E. (1998) “The Size of Regions,” in Topics in Public
Economics: Theoretical and Applied Analysis, (ed.) by D. Pines, E.
Sadka, and I. Zilcha. Cambridge: Cambridge University Press.
• Krugman, P. (1991) “Increasing Returns and Economic Geography,”
Journal of Political Economy, 99(3), 483-99.
• Krugman, P. (1991) Geography and Trade, Cambridge: MIT Press.
• Krugman, P. and A. J. Venables (1995) “Globalisation and
Inequality of Nations,” Quarterly Journal of Economics, 60, 857-80.

34 / 35
 References

• Marshall, A. (1920) Principles of Economics, London: Macmillan.

• Redding, S. J., and D. M. Sturm (2008) “The Costs of Remoteness:
Evidence from German Division and Reunification,” American
Economic Review, 98(5), 1766-1797.
• Redding, S. J. and A. J. Venables (2004) “Economic Geography and
International Inequality,” Journal of International Economics, 62(1),
53-82.

35 / 35