Dynamics, growth and geography

1
Chapter 10

• Dynamics, growth and geography

2

• Long term equilibrium adjustment
• d?i/?i ?(wi ?)
• Value of ? sets the speed of adjustment but in
  general does not have an effect on the outcome
• Exceptions
• overshooting
• unstable equilibrium
• not the nearest equilibrium is reached

3
Figure 10.1 Regular adjustment dynamics
4
Figure 10.1 Regular adjustment dynamics
5
Figure 10.1 Regular adjustment dynamics
6
Figure 10.2 Special adjustment dynamics (along
vertical axis, ? horizontal axis, number of
reallocations)
7
Figure 10.2 Special adjustment dynamics (along
vertical axis, ? horizontal axis, number of
reallocations)
8
Figure 10.2 Special adjustment dynamics (along
vertical axis, ? horizontal axis, number of
reallocations)
9
Can growth periods be simulated?

• Convergence/divergence
• Different for countries/regions
• Convergence/divergence speed different in
  different periods

10
Figure 10.3 Histogram of per capita income,
selected years
11
Figure 10.3 Histogram of per capita income,
selected years
12
Figure 10.4 Regional convergence in the EU, speed
of convergence estimates
13
Figure 10.5 Regional income inequality in the EU
Lorenz curves
14
EU 1995-2001
15
EU 1995-2001
16
EU 1995-2001
17
EU 1995-2001
18

• Between-country inequality
• Assuming equal gdp/cap inside each country
• Within-country inequality
• Assuming equal national gdp/cap worldwide

19
Figure 10.6 Regional income inequality in the
EU Theil index and Gini coefficient
20
Figure 10.7 Leaders and laggards in the world
economy, 1-2003

income per capita ( of world average)
Switzerland
USA
Australia
Netherlands
UK
Italy
Italy
Iran
Iraq
New Zealand
China
Many
India
Many
W Offshoots
Australia
oAfrica
year
21
To be explained

1. For almost all countries even increasing level of
   income
2. Differences between countries may persist for a
   long time
3. Long periods of stagnation can be followed by
   long periodes of growth
4. Frequent changes in economic ranking
   (leap-frogging)

22
Theories

• Endogenous growth Y A f (K,L)
• Total factor productivity A as a function of K or
  L can explain (1)
• In a closed model A can be structurally different
  per country can explain (2)
• (3) and (4) cannot be explained by endogenous
  growth theory
• Need for geographical economics

23
Recall Krugman Venables (1995)
Fig 4.10 The bell-shaped curve
1
?1
0,5
0
T
Unstable equilibria
Stable equilibria
VL model with lowering T from dispersion to
agglomeration to dispersion Very simple
explanation of (3)
24
Simulations of (3) and (4)

• Using the 24 region racetrack model with
  congestion, unchanged e5, d0.6, t0.05
• Simulating a change in transport costs over time
• Some random initial distribution (history)
• Find long term equilibirum with T3, then
  decreasing
• Herfindahl index HS?i2

25
Figure 10.8 Distribution of manufacturing and
Herfindahl index
26
Figure 10.9 Evolution of agglomeration, the
Herfindahl index
27
Figure 10.10 Several phases of the reallocation
process
28
Figure 10.10 Several phases of the reallocation
process
29
Figure 10.10 Several phases of the reallocation
process
H does not tell anything about spikes
30
Figure 10.11 Dynamics of regional size regions
3, 6, and 9
31
Combine agglomeration and growth

• Baldwin Forslid (2000)
• Extend the CP model with capital K produced by
  sector In (investment sector)
• With global knowledge spill-overs location of In
  does not matter
• With local knowledge spill-overs location of In
  does matter
• High policy relevance many governments stimulate
  knowledge flows to periphery with
  universities/high-tech industrial parks etc.

32
Baldwin Martin (2004)

• Cost function M sector R Wßxi
• K is produced under perfect competition with only
  variable labor aI under knowledge spill-overs aI
  falls with rising output
• QkLI / aI
• aI 1 / K-1 ? K-1
• With
• Qk flow of new capital
• LI employment in investment sector
• K stock of knowledge ( other region)
• ? parameter degree of spillovers
• (capital depreciates in one period)

33
Intertemporal utility

• U St (1/1?)t ln (Ft1-dMtd)
• See box 10.1
• Mobility related to difference in present value
  of real wages in each region

34
Main results

• For two region model only spreading and complete
  agglomeration into one region are stable
  long-term equilibria
• -gt same as in CP model
• with increasing ? more stable equilibria possible

35
Figure 10.12 Stability in the Baldwin-Forslid
economic growth model
36
"deep determinants of growth"

• growth different because A is localized, but
• why is A localized?
• institutions (table 10.4) relevant
• again the discussion on first nature returns
• climate, land-locked
• tropical diseases (Sachs)
• missing second nature the role of geography
  relative to other geographies
• spatial autocorrelation (first block of the
  course by Paul Elhorst)
• spatial autocorrelation of institutions?

37
Figure 10.13 Scatterplot of own and neighboring
institutions
38
Conclusions (p451)

• Integrated endogenous growth/geographical
  economic models can deal with (1)-(4) but
• do not pay attention to deep determinants of
  differentiated growth
• models that do take account of deep determinants
  ignore second nature/spatial interdependence
• long term history and path-dependence (box 10.2)
• country borders change over time
• cities do not research more promising
• needed integration of deep determinants, history
  and second nature geography