Cities and Growth: What Do We Know?

1

• Cities and Growth What Do We Know?
• By Peter Gordon and Bumsoo Lee
• University of Southern California
• University of Illinois, Champaign-Urbana
• September 17, 2008

2
What Do We Know?

• Why are there cities?
• People have found ways to organize themselves so
  that agglomeration benefits dominate associated
  congestion costs.
• Cities are the engines of growth.
• Economic growth springs from entrepreneurial
  activity (e.g., discovery, J. Schumpeter).
• Optimal urban scale discussions are not
  helpful they suggest a static analysis.

3

• 5. Prefacing a recent NBER symposium on
  agglomeration economies, (http//www.nber.org/book
  s/glae08-1/),
• Ed Glaser writes
• a central paradox of our time is that in
  cities, industrial agglomerations remain
  remarkably vital despite ever easier movement of
  goods and knowledge across space.

4
Questions

• How can cities be congenial hosts to
  entrepreneurship and discovery?
• Can we understand the role of urban structure?
• Which spatial arrangements internalize positive
  externalities while avoiding negative ones, in
  light of all other trade-offs?
• Gordon and Moore (E and P(A), 1989) discussed
  land use optimizations that suggest an answer.
  Are there plausible tests?

5
Traffic and Transportation

1. Modern lifestyles generate massive volumes of
   non-work travel during peak and off-peak hours.
2. Combining this with the absence of pricing, the
   continued growth of cities (metro areas) is
   remarkable.
3. Impending traffic doomsday is forever
   impending.

6
Distribution of daily person-trips by trip
purpose and period of the week, 2001
In billions
7
Distribution of daily person-trips by trip
purpose and period of the week, 2001
In percent
8
Growth of average daily person-trips per person
by trip purpose and period of the week, 1990 to
2001
Work Non-Work
All
Family / School / Social /
Personal Church
Recreation
9
Industry Churning and Growth

• Cities succeed and maintain their status in the
  urban ranking by churning industries (Duranton,
  2006) as circumstances change.
• Spatial form must change to make all of this
  possible. Test more than metro-area average
  densities.
• Y year
• Zsector
• eemployment
• cMSA

10
Employment shares by location type
Results from GWR procedure.
11
Mean commute time by workplace location type vs.
metro population size (drive alone mode)
12
Urban Spatial Structure and Metro Growth
Empirical Model 1

• Glaesers (2003) supply-side urban growth model
• where Nt and Nt-1 denote population (employment)
  size in 2000 and 1990, respectively
• X is a vector of metropolitan attributes
• F is vector of spatial structure variables
  dispersion and polycentricity

13
Estimation Results
14
Growth effects of spatial structure at different
metro sizes
6.05
15
Urban Structure and GrowthModel 2 Specification

• Locally Weighted Regression (LOESS)
• Allows the coefficients of spatial structure
  variables to vary without restrictions
• At each data point, a supply-side economic growth
  model is fit to a subsample of observations that
  are similar in Pop/Emp size (window size 41
  obs.)
• Similar size observations get more weight

16
LOESS Results

1. LOESS results corroborate interaction term
   significance of OLS results.
2. Estimate of coefficient for dispersion is about
   zero near sample mean size pop (log pop 14, 1.25
   million), positive for larger metros, negative
   for smaller metros.
3. Estimate of coefficient for polycentricity always
   near zero.

17
Model 2 Population Growth Model, Varying
Coefficients of Spatial Structure
18
Model 2 Employment Growth Model, Varying
Coefficients of Spatial Structure
19
Summary of Results

1. The growth effects of spatial structure
   (dispersion) were found to be contingent on
   metropolitan size.
2. When small, a metropolitan area with more
   clustered spatial form grows faster, realizing
   agglomeration economies in this way. But as a
   metro area becomes larger, more dispersion
   accommodates greater growth.
3. Just as a metro takes on higher-order economic
   functions to move up within an hierarchical urban
   system, it also (concurrently) restructures its
   spatial form in ways to mitigate congestion or
   other diseconomies of size for continued growth
   to be possible.

20
Conclusions

• Growth can be accommodated.
• 2. Continued productivity, growth and prosperity
  require adaptable (open-ended) urban forms.
  Suggests importance of flexible land, labor and
  capital markets. Openness to spontaneous forms,
  learning feedback and endogenous clustering (J.
  Jacobs) -- rather than approaches like
  pro-cluster policies.
• 3. None of the involved markets are (or can be)
  completely free or unhampered, but evidence of
  performance and spatial change is impressive.