Appalachian technology clusters

1
Appalachian technology clusters economic
development outcomes

• An assessment of cluster growth and change
• Edward Feser
• University of Illinois at Urbana-Champaign
• October 2005

2
Overview

• The cluster phenomenon
• Question
• Antecedents objectives
• 2002 ARC study as baseline
• Initial findings

3
The cluster phenomenon

• Interdependence Each member firms competitive
  position depends on one, some, or all other
  members of the group. Business is the core.

Trading sectors
Related sectors
Supporting institutions
Intermediate suppliers Capital good
suppliers Producer services Consultants Contract
RD
Similar technologies Share pool of labor Similar
strategies
Education (univ, colleges) Training (ccs) RD
(univ, fed labs) Development agencies Regulatory
agencies
4
Question

• What is the relationship between firm and
  industry performance and clustering?
• Do clustered firms and industries out-perform
  non-clustered firms and industries?
• Are there measurable benefits to clustering?
• What is the relationship between other cluster
  members performance and clustering?
• Academic institutions
• Contract RD houses, federal labs

5
Question

• Performance should be some function of cluster
  intensity

Performance
Clustering
6
Antecedents

• Revealed location preference
• Agglomeration economies
• Employment growth
• Productivity
• Case study analysis

7
Cluster cycles

• Net benefits of clustering are not necessarily
  positive in perpetuity
• Lock in
• Flammang
• Mancur Olsen
• We would expect to see wide variation in
  relationship between cluster levels and
  performance over given period t to t1, depending
  on cluster vintage

Advantage
Cluster benefits
Liability
time
Cluster size
time
8
Objectives

• Begin to investigate link between performance of
  cluster members and overall cluster levels
  change
• Utilize results of extensive 2002 study of
  technology-based clusters in Appalachian Regional
  Commission area

9
2002 ARC study

• Identified geographic concentrations of science
  and technology assets in the Appalachian region
• ST assets RD, innovation, human capital
  creation and training, and tech-intensive
  industry
• In what technology specializations?
• How competitive are the strengths?
• To what degree is there a match between RD
  assets and the tech-intensive industries, by
  sector and spatially?
• Where are matches located?
• Described the overall geographic pattern of
  clustering in the region as of late 1990s

10
ARC area

• 406 counties of ARC
• Study included adjacent border region
• 62 metro areas (shaded)

11
2002 ARC study approach
University RD
Private research houses
Geography
Geography
Federal RD labs
Knowledge creation, innovation, education
Industry patents
Goods services production
Tech Clusters
Technology value-chains
SBIR grants
Science technical workers
Teaching universities
Specializations
Specializations
Community colleges
Tech clusters Joint spatial concentrations of
related high-tech industry and innovation/knowledg
e activity. Places where a moderately to highly
sophisticated knowledge infrastructure is joined
with a substantial related industrial base.
12
Methodological challenges

• Assembly of multiple, overlapping sources of
  evidence

Measures, indicators
Industry, tech area concordances
Geographic units of analysis
Data sources
13
Measurement of tech industrial base
14
(No Transcript)
15
(No Transcript)
16
Industrial clustering Chemicals
Map Legend
ARC boundary
Analysis buffer
significant Gi, counties, zips
significant Gi, county growth
location quotient gt 1.1
17
Occupation clustering Scientists engineers
Rochester Chem, Matrl, Indust
Albany Enviro
Erie Chem, Matrl
Youngstown Matrl
Binghamton Elect
Cleveland Chem, Matrl, Indust
Newburgh Chem
Newark Chem, Enviro
Akron Chem
Allentown Chem
Reading Chem, Matrl
Map Legend
ARC boundary
State College Elect
Pittsburgh Math, Matrl
Analysis buffer
State boundary
Metro areas with employment location quotients gt
1.25 in given technology area indicated. Source
BLS Occupational Employment Survey, 1999. See
Table 5 for abbreviated codes.
Metro area
18
Measurement of innovative activity
19
U.S. Rankings, RD funding 1999
20
University RD strengths1st and 2nd tier
strengths based on U.S. rank
21
Innovation clustering Patents 90-99
Map Legend
ARC boundary
Analysis buffer
significant Gi, counties
location quotient gt 1.25
22
Innovation clustering SBIR/STTR/ATP
Ithaca
Binghamton
Pittsburgh
State College
Blacksburg
Birmingham
Knoxville/Oak Ridge
Huntsville
Dots scaled by number of awardees per zip code.
23
Skill base clusters Degree completions
Total degree completions by county 2-year
colleges institutes, 1997/8, All
technology-related fields
24
Innovation clustering Laboratories
25
Tech clusters Chemicals plastics
Buffalo, Rochester
Binghamton
Pittsburgh
Erie, Cleveland, Akron
Albany-Schenectady
Cornell
Newburgh
Wheeling
Penn State
Reading, Allentown
Parkersburg
State College
Charleston
Concentrations of both production (value-chain or
ST occ. employment) and innovation (patents,
ranked universities, or federal innovation
grants) are labeled. All indicators are specific
to technology area.
Cincinnati
Washington, DC
Map Legend
ARC boundary
Analysis buffer
Huntsville, Decatur
Johnson City
1-3 SBIR/STTR/ATP awards (scaled)
Research universities, 1st Tier
Asheville
Research universities, 2nd Tier
GA Tech
Metro area, ST workers LQ gt1.25
Greenville, Spartanburg
Auburn
Sig. emp. Gi or LQgt 1.1, counties
Chattanooga
Sig. patent Gi or LQ gt 1.25, counties
Atlanta
Two-year college completions, 1997/98
gt 750
250 - 749
100 - 249
26
Tech clusters Industrial machinery
Albany-Schenectady
Buffalo, Rochester
Erie
Binghamton
Cornell
Pittsburgh
Penn State
State College
Cleveland, Canton, Akron, Youngstown
Reading, Allentown
Mansfield
Harrisburg
Altoona
Concentrations of both production (value-chain or
ST occ. employment) and innovation (patents,
ranked universities, or federal innovation
grants) are labeled. All indicators are specific
to technology area.
Cincinnati, Middleton
Johnstown
Lynchburg
Map Legend
Nashville-Davidson
U. of Tennessee, Knoxville
Greensboro
ARC boundary
Analysis buffer
Statesville
1-2 SBIR/STTR/ATP awards (scaled)
Starkville, Columbus
Research universities, 1st Tier
Charlotte
Research universities, 2nd Tier
Asheville
U. of Alabama, Huntsville
Metro area, ST workers LQ gt1.25
GA Tech
Greenville-Spartanburg
Sig. emp. Gi or LQgt 1.1, counties
Sig. patent Gi or LQ gt 1.25, counties
Atlanta
Huntsville
Two-year college completions, 1997/98
gt 750
250 - 749
100 - 249
27
Tech clusters Info tech instruments
Binghamton
Rochester
Cornell
State College
Poughkeepsie
Penn State
Carnegie Mellon
Columbus
Concentrations of both production (value-chain or
ST occ. employment) and innovation (patents,
ranked universities, or federal innovation
grants) are labeled. All indicators are specific
to technology area.
Washington, DC
Virginia Polytechnic
Lynchburg
Map Legend
Huntsville
ARC boundary
Analysis buffer
1-28 SBIR/STTR/ATP awards (scaled)
Research universities, 1st Tier
Atlanta
Research universities, 2nd Tier
Metro area, ST workers LQ gt1.25
GA Tech
Sig. emp. Gi or LQgt 1.1, counties
Sig. patent Gi or LQ gt 1.25, counties
Two-year college completions, 1997/98
gt 750
250 - 749
100 - 249
28
(No Transcript)
29
(No Transcript)
30
2002 study Overall findings

• Identified 100 tech clusters of varying size and
  strength
• Uneven spatial distribution
• 45 northern, 19 central, 29 southern
• In central, Cincinnati and Washington account for
  9 of 19
• In south, Atlanta, Greenville-Spartanburg, and
  Huntsville account for 16 of 19
• Tech areas vary in geography
• Chemicals and IT more even than industrial
  machinery, which is nearly exclusively northern
  and southern
• Over half of the tech clusters are peripheral to
  the ARC region proper
• Role of (positive) spillover effects?

31
Next step

• ARC study represented picture of regional
  clustering as of 1998/99
• Investigate relationship between clustering and
  subsequent tech sector growth
• 1998-2001 (national expansion period)
• 2001-2002 (national recession period)
• 2002-forward (recovery and growth)
• Investigate relationship between clustering and
  academic laboratory performance

32
Three base sector clustering measures

• Employment, cluster r, 1998
• Employment change, cluster r, 1989-1998
• Utility patent grants, cluster r, 1990-99

33
Focus on sectoral growth

• Problem of SIC to NAICS concordance
• Use of Hecker (1999) technology sector scheme

34
(No Transcript)
35
(No Transcript)
36
(No Transcript)
37
(No Transcript)