Cluster Mapping in
Europe and the United States
Christian H. M. Ketels, PhD Institute for Strategy and Competitiveness
Harvard Business School Trend Chart Workshop Brussels, Belgium 15 November 2005
This presentation has benefited from Professor Michael E. Porter’s articles and books and ongoing research at the Institute for Strategy and Competitiveness as well as joint work with Professor Örjan Sölvell at the Center for Strategy and Competitiveness, Stockholm School of Economics. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means - electronic, mechanical, photocopying, recording, or otherwise - without the permission of the author
Additional information on the Institute for Strategy and Competitiveness is available at www.isc.hbs.edu Additional information on the Center for Strategy and Competitiveness is available at www.sse.edu/csc
2 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
The Role of Cluster Mapping
• Provide a precise language for discussing clusters and their role in regional economies
• Provide data for regional economies to develop competitiveness strategies reflecting their individual cluster portfolios
• Enable regional clusters to systematically compare their size and profile over time and with peers in other locations
• Guide the use of policy instruments tied to the presence of clusters across locations
• Cluster mapping is a key element in moving cluster-based economic
policy to the next level
3 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Cluster Policy Approaches
• Cluster effort often based on national programs
• Strong role of government in initiating cluster efforts
• Lower level of specialization across regional economies
• Business environments tend to be strong on factor input
conditions, often weaker on context for strategy and rivalry
• Cluster effort often based on national programs
• Strong role of government in initiating cluster efforts
• Lower level of specialization across regional economies
• Business environments tend to be strong on factor input
conditions, often weaker on context for strategy and rivalry
• Cluster efforts based on regional initiatives
• Strong role of private sector from the outset of cluster efforts
• Many regional economies highly specialized around strong clusters
• Business environments tend to be very open to cross-regional
competition and have access to strong factor input conditions
• Cluster efforts based on regional initiatives
• Strong role of private sector from the outset of cluster efforts
• Many regional economies highly specialized around strong clusters
• Business environments tend to be very open to cross-regional
competition and have access to strong factor input conditions Europe
Europe United States United States
4 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Use of Cluster Mapping
United States
• Identification of regional clusters
• Assessment of economic performance of regional clusters
• Development of regional strategies to mobilize clusters
• Analysis of the relationship between cluster presence and regional economic performance
Europe
• Intentions as above
• Intention to use cluster definitions to guide public policy programs
5 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Cluster Mapping Approaches
HBS Cluster Mapping Project
• Based on actual co-location of industries; revealed impact of sum of locational factors on company decisions
• Use of U.S. data because the U.S. economy has been exposed to free cross- regional competition among the locations for the longest time
– “a peek into the future of other locations”
Alternatives/complements
• Input – output relationships; supplier relationships
• Cross-company/institution career paths; social networks
• Co-publication/citation data; knowledge spill-overs
6 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
HBS Cluster Mapping Project
• Use of employment data at the 4-digit industry level for regional economies
• Calculation of regional concentration per industry across the U.S.
– No concentration: Local industries
– Significant concentration: Traded clusters and natural-resource driven clusters
• Calculation of correlation patterns among industries in the traded clusters-category
• Based on correlation patterns identification of 41 cluster groups (and
>200 sub-cluster groups) that industries get assigned to
– Narrow cluster definition: Each industry allocated to one cluster
– Broad cluster definition: Industries can be allocated to more than one
cluster
7 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Traded Clusters
Traded Clusters Local Clusters Local Clusters Local Clusters
Natural
Resource-Driven Industries
Natural Natural Resource
Resource- -Driven Driven Industries Industries
30.5%
0.9%
$45,511 129.7%
4.3%
144.1
21.3 590 30.5%
0.9%
$45,511 129.7%
4.3%
144.1
21.3 590
68.8%
2.4%
$29,010 82.7 3.6%
79.3
1.3 241 68.8%
68.8%
2.4% 2.4%
$29,010
$29,010 82.7 82.7 3.6% 3.6%
79.3 79.3
1.3 1.3 241 241
0.7%
-1.2%
$33,066 94.3 1.8%
140.1
7.0 48 0.7% 0.7%
-1.2% - 1.2%
$33,066
$33,066 94.3 94.3 1.8% 1.8%
140.1 140.1
7.0 7.0 48 48
Share of Employment Employment Growth Rate,
1990 to 2002 Average Wage Relative Wage Wage Growth
Relative Productivity
Patents per 10,000 Employees Number of SIC Industries
Note: 2002 data, except relative productivity which uses 1997 data.
Source: Prof. Michael E. Porter, Cluster Mapping Project, Institute for Strategy and Competitiveness, Harvard Business School
Composition of Regional Economies
United States, 2002
8 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Plastics
Oil and Chemical Gas
Products
Pharma- ceuticals
Power Generation Aerospace Vehicles &
Defense
Lightning &
Electrical Equipment Financial
Services
Publishing and Printing
Entertainment
Hospitality and Tourism
Transportation and Logistics
Information Technology
Communi- cations Equipment
Medical Devices
Analytical Instruments Education
and Knowledge
Creation Apparel
Leather and Sporting
Goods
Agricultural Products
Processed Food
Furniture Building
Fixtures, Equipment
and Services
Cluster Overlap in the United States Economy
Common Industries Across Broad Traded Clusters
Note: Clusters with borders or identical colors/shading except gray have at least 20% overlap of industries by number in both directions
Sporting, Recreation and Children’s
Goods
Business Services
Distribution Services Fishing &
Fishing Products
Footwear
Forest Products
Heavy Construction
Services
Jewelry &
Precious Metals
Construction Materials
Prefabricated Enclosures Textiles
Tobacco
Heavy Machinery Aerospace
Engines
Automotive
Production Technology
Motor Driven Products Metal
Manufacturing
9 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
0 1 2 3 4
-50 0 50 100
Specialization of Regional Economies
Atlanta Metro Area
Percentage Share of
National Cluster Employment
in 2000
Percentage Change, 1990–2000
= 0–19,999 = 20,000–49,999 = 50,000–99,999 = 100,000+
Power Generation (1.8, 320.1)
Oil and Gas Agricultural Products Leather Products
Heavy Construction Services
Heavy Machinery Processed Food
Analytical Instruments
Production Technology Metal
Manufacturing
Education and Knowledge Creation Distribution
Services Financial
Services
Transportation and Logistics (4.1, 74.7)
Business Services
Jewelry and Precious Metals Prefabricated
Enclosures
Furniture Lighting and
Electrical Equipment
Apparel
Hospitality and Tourism
Pharmaceuticals and Biotechnology
Atlanta’s Average Share = 1.9%
Note: Uses narrow cluster definitions to avoid overlap
Source: Cluster Mapping Project, Institute for Strategy and Competitiveness, Harvard Business School
Motor Driven Products Aerospace Vehicles
and Defense
Aerospace Engines (0.5, 601.7)
Textiles Building Fixtures,
Equipment and Services
Sporting Products Automotive
IT Communications
Equipment
10 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
$15,000
$25,000
$35,000
$45,000
$55,000
50 100 150 200 250 300
Average Regional Wage, 2001
Share of Traded Employment in Strong Clusters (LQ > .8), Broad Cluster, 2001
y = 96.736x + 16218 R
2= 0.377 New York, NY
Bay Area, CA
Boston, MA
Determinants of Regional Prosperity
Cluster Strength and Wage Levels, U.S. Regions
Source: County Business Patterns; Michael E. Porter, The Economic Performance of Regions”, Regional Studies, Vol. 37, 2003
11 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Specialization and GPP / Capita
y = 336,5x + 35,336 R
2= 0,4358 0
20 40 60 80 100 120 140 160 180
0% 5% 10% 15% 20%
Percent of employment in specialized clusters (SQ>2) G D P / C a p ita p u rc h a s in g pow e r st andar d s of E U av er age
Praha City
Bratislava
Determinants of Regional Prosperity
Cluster Strength and GDP per Capita, EU-10 Regions
Source: Solvell/Ketels/Frederiksson, Regional clusters in the EU-10, 2005
12 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
0.50 0.75 1.00 1.25
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Determinants of Regional Prosperity
Traded Cluster Specialization and Relative Wage Levels: Ohio
Relative Cluster Wage
Relative Employment (LQ) by Traded Cluster Financial Services
Note: Uses narrow cluster definitions to avoid overlap; bubble size proportional to employment bracket Source: Cluster Mapping Project, Institute for Strategy and Competitiveness, Harvard Business School
Automotive Metal Manufacturing
4.29% of U.S. Employment
U.S. average cluster wage Business Services
Production Technology
y = 0.1902Ln(x) + 0.9874
R
2= 0.3403
13 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Determinants of Regional Prosperity
Level versus Mix Effect, U.S. Regions
Cluster Wage Level Effect as % of Wage Gap, 2001
U.S. Economic Areas
Source: County Business Patterns; Michael E. Porter, The Economic Performance of Regions”, Regional Studies, Vol. 37, 2003
Median: 74.2%
14 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Source: County Business Patterns; Michael E. Porter, The Economic Performance of Regions”, Regional Studies, Vol. 37, 2003
Determinants of Regional Prosperity
Change in Cluster Specialization and Wage Growth, U.S. States
2.0%
2.5%
3.0%
3.5%
4.0%
4.5%
5.0%
5.5%
-0.06 -0.04 -0.02 0.00 0.02 0.04 0.06 0.08 0.10
Annual Regional Wage Growth Rate, 1990-2001
Change of Cluster Employment GINI, 1990-2001
y = 8.7905x + 3.6107R2= 0.2626 P-value = .0001
MA NY
Economy becoming less specialized
Economy becoming more specialized
AK
CA CT CO
DE DC
FL GA
HI ID
IL
IN IA
KS
KY
LA
MN
MO NV MT
NJ NC
ND
OK OR
PA RI
SC
TX WA VA
WV WI
WY AL
AR
15 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Explaining Average Regional Wages
Multiple Regression Model
Independent Variable Independent Variable
• Total regional employment
• Patents per capita
• Patentor concentration
• Share of strong clusters in regional employment
• Cluster breadth
• Total regional employment
• Patents per capita
• Patentor concentration
• Share of strong clusters in regional employment
• Cluster breadth
Effect Effect
Positive, significant Positive, significant Negative, significant Positive, significant
Positive, significant Positive, significant Positive, significant Negative, significant Positive, significant
Positive, significant
Dependent variable: Regional Average Wage
Note: Regression uses 2001 data for 172 U.S. economic areas
Source: Michael E. Porter, The Economic Performance of Regions”, Regional Studies, Vol. 37, 2003
Explained Variation (adjusted R
2): 72.8%
16 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Comparative Data on European Clusters
Stockholm Cluster Portfolio
0%
10%
20%
30%
40%
50%
60%
-15% -10% -5% 0% 5%
Change of Share in National Cluster Employment, 1995-2003
Stockholm Share of National Cluster Employment, 2003: 22.9%
Change in Stockholm’s overall share of National Cluster Employment: -0.5%
Note: Bubble size is proportional to employment levels Source: Statistics Sweden (2005), author’s calculations
Biopharmaceuticals
Financial Services
Business Services Communication Equipment
Information Technology
Distribution Services
Education & Knowledge Creation
Heavy Construction Services Tourism Publishing & Printing
Analytical Instruments
Transportation & Logistics Share in National
Cluster Employment,
2003
17 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Region
Region Cluster Cluster Employment Employment
Schleswig-Holstein (DE) Västsverige (SE)
Hamburg (DE) Etelä-Suomi (SF) Stockholm (SE)
Östra Mellansverige (SE)
Mecklenburg-Vorpommern (DE) Warminsko-Mazurskie (PL) Norra Mellansverige (SE) Oslo og Akershus (NO) Småland med öarna (SE) Warminsko-Mazurskie (PL) Norra Mellansverige (SE) Islands (IS)
Agder og Rogaland (NO) Länsi-Suomi (SF)
Schleswig-Holstein (DE) Västsverige (SE)
Hamburg (DE) Etelä-Suomi (SF) Stockholm (SE)
Östra Mellansverige (SE)
Mecklenburg-Vorpommern (DE) Warminsko-Mazurskie (PL) Norra Mellansverige (SE) Oslo og Akershus (NO) Småland med öarna (SE) Warminsko-Mazurskie (PL) Norra Mellansverige (SE) Islands (IS)
Agder og Rogaland (NO) Länsi-Suomi (SF)
Financial Services Automotive
Financial Services Forest Products Business Services Metal Manufacturing Hospitality and Tourism Processed Food
Metal Manufacturing Business Services Metal Manufacturing
Building Fixtures, Equipment and Services Forest Products
Fishing and Fishing Products
Oil and Gas Products and Services Metal Manufacturing
Financial Services Automotive
Financial Services Forest Products Business Services Metal Manufacturing Hospitality and Tourism Processed Food
Metal Manufacturing Business Services Metal Manufacturing
Building Fixtures, Equipment and Services Forest Products
Fishing and Fishing Products
Oil and Gas Products and Services Metal Manufacturing
60,423 43,168 42,420 40,722 38,283 28,706 26,538 21,831 21,240 17,966 16,995 14,431 13,674 11,931 10,752 10,090 60,423 43,168 42,420 40,722 38,283 28,706 26,538 21,831 21,240 17,966 16,995 14,431 13,674 11,931 10,752 10,090
Note: “3 Star” defined as >10.000 employees, > 10% of regional employment, and SQ > 2. Data set does not include Denmark and Russia Source: Institute for Strategy and Competitiveness, author’s calculations
3 STAR-Clusters in the Baltic Sea Region
18 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Estonia
Latvia
Cyprus
Malta Praha Region; CZ
Praha City; CZ
Gdansk; PL Szczecin; PL
Warszawa; PL
Slovenia
Budapest; HU Miskolc; HU
Comparative Data on European Clusters
Transportation and Logistics Clusters in the EU-10
Source: Solvell/Ketels/Frederiksson, Regional clusters in the EU-10, 2005
19 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Conclusion
• Cluster mapping is a tool, not a solution; it is critical for a more fact- driven discussion about cluster-based economic policy
• Cluster mapping efforts should be part of a wider cluster data infrastructure
– Cluster-specific business environment assessments – Impact assessment for cluster-based policy initiatives
• Creating this data infrastructure is a useful task for the European Commission; running cluster-based efforts themselves is not
• The available cluster mapping data suggests that Europe is in the
midst of a relocation process that has already proceeded much
further in the U.S.
20 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Back-Up
21 Copyright 2005 © Dr. Christian H. M. Ketels Cluster-Based Development 03-10-04 CK
Presence of Clusters Across Countries
Selected Countries
“Clusters are common and
deep”
“Clusters are limited and
shallow”
Note: EU members in red (EU-15) and blue (NMS), other countries in green; arrows indicate significant changes since 2002 Source: Global Competitiveness Report 2004-2005, World Economic Forum
Survey Question: “How Common Are Clusters In Your Country?”
Average of all 93 countries
1 2 3 4 5 6 7
Japan Finland
United States Italy Denm
ark Sweden
Ireland
United Kingd om
Canada Germany
Switzerland Norwa
y Austria
France Netherlan
ds Belgium
Portuga l
Australia Spain New Zealand
Lithuania Czech Repub
lic Sloven
ia Poland
Slovak Re pub
lic Latvia
Greece Estonia
Hunga ry
Malta