This document proposes a new methodology called the PB-TO DFM model to assess the performance and efficiency of 38 global cities. It uses data envelopment analysis (DEA) with two viewpoints - human environment and socioeconomic activity - to evaluate the cities. For less efficient cities, the PB-TO DFM model provides step-by-step strategies to improve performance by adjusting certain inputs and outputs. Case studies on Amsterdam and Stockholm demonstrate how the model can determine efficiency improvement pathways for specific cities. The methodology aims to help decision-making and planning to strengthen global cities as "urban empires" in the new urban world.
Urban Empires – Cities as Global Rulers in the New Urban World
1. In Pursuit of High-Performance Global Cities
–An Extended DEA Benchmark Model for Assessing Urban
Socio-economic Environmental Welfare Indicators
Soushi Suzukia, Karima Kourtitb,d and Peter Nijkampb,c,d
aHokkai-Gakuen University, Sapporo, Japan
b KTH Royal Institute of Technology, Stockholm, Sweden
c Tinbergen Institute, Amsterdam, The Netherlands
dAdam Mickiewicz University, Poznan, Poland
Advanced Brainstorm Carrefour
Urban Empires -Cities as Global Rulers in the New Urban World
Augustus 15, 2016, Adam Mickiewicz University, Poznan
Urban Empires –
Cities as Global
Rulers in the New
Urban World
2. Space in Transition
• Peter Gould (1963):
- Man against Nature
- Locational patterns
• Lucassen and Willems (2011):
- Challenge and Response
- Adaptation and rising urbanisation
• Kourtit and Nijkamp (2012):
- Globalisation and migration
- Agglomeration advantages and creative cities
THE NEW URBAN WORLD
3. The New Urban World
2 Trends:
• Persistent Urbanization
• Fast Urbanization
4. Megatrends – The New Urban World
• Rising urbanization everywhere (not every city)
• Cities as ‘the home of man’
• Urban areas as centres of development and of concerns
• Pluriformity in urban appearance and socio-economic
development
• Dominance of sustainability conditions (XXQ, Nijkamp,
2010)
• No natural or economic limit to city size
• The law of Van Loon (1932)
• Smart specialisation
• Need for effective long-range policy responses
• Challenges for Regional Science
5. Figure 1.
Percentage of
population in
city areas in
Japan
• We live nowadays in the ‘urban century’.
• The role of urban systems is becoming more and more important.
The megatrend of population concentration in city areas does not
come to a standstill, even not in a depopulating and ageing
society like Japan (Figure 1).
Urban Century
6. • Global cities play a role as global ‘rulers’ in the
‘New Urban World’.
• In the globalization and environmentalization age, large
urban areas act as:
- international communication stations, with a high
human intelligence ability and a powerful technological
and socio-economic activity (Socioeconomic-cognitive
activity),
- environmental coexistence stations, with a high-
quality residential profile and an ecologically-friendly
human environment structure
(Human environment profile).
• There is a rising interest in ranking and rating systems
for cities on the basis of systematically designed
comparative benchmark principles.
7. • A novel multidimensional analysis based on a Data
Envelopment Analysis (DEA) which can evaluate an
efficiency of Decision Making Units (DMUs) will be
adopted in our study.
• Our study ties seeks to offer an advanced
methodological contribution to the identification of
high-performance cities (HPCs) on the basis of an
extensive multivariate database on a set of 38 global
cities.
• We also employ a ‘smart’ improvement strategy for less
efficient cities in our sample, based on a newly
developed efficiency improvement projection model in
DEA.
8. Global cities in the GPCI database
Source: Global Power City Index (GPCI) (2015), p.7
10. Outline of DEA
uv,
max
,0mv 0su
(FPo)
s.t.
: an efficiency score
xmj : the volume of input m in DMU j
ysj : the volume of output s in DMU j
vm and us : the weights given to input m and output s
I1(x1)
I2(x2)
O
A C
B
C’
( =OC’/OC)
DMU
,1
m
mjm
s
sjs
xv
yu
m
mom
s
sos
xv
yu
DEA was developed to analyze the relative efficiency of
Decision Making Unit (DMU), and projecting the
performance of each DMU onto the efficient frontier.
11. • The efficiency improvement projection:
The original DEA models have only focused on a uniform
input reduction in the improvement projections.
The solution of an efficient improvement problem is not
only just one point.
I1(x1)
I2(x2)
O
A C
B
C’
( =OC’/OC)
DMU
12. • Suzuki and Nijkamp et al.(2010) proposed a DFM model that can
compute more effectiveness solutions than the original projection.
AOriginal
Original Projection
A
ADFM
DFM-Projection
Weighted
Input 2
(v2
*x2)
Weighted Input 1 (v1
*x1)
• DFM does not need to incorporate subjective value judgments of a
decision maker.
• Nevertheless, the strategies to improve a city’s performance are
also based on political targets and preferences of city stakeholders.
• Therefore, in many decision-making situations, a balance between
input and output targets has to be found. It seems more plausible
that this balance is to be co-determined by a DMU’s preference
pattern.
Outline of Distance Friction Minimization (DFM) Approach
13. • The target values in a Preference Based model, which are allocated
between input efforts and output efforts based on Output
Augmentation Parameter (OAP)(Examples OAP=0.7).
• This model is able to calculate both input reduction value and
output increase value so as to reach an efficiency score of 1.0,
despite the fact that in reality this might be difficult to achieve for
low-efficiency DMUs.
Target Value
(OAP=0.7)
Input score
Target value
(DFM model)
70%
30%
(Input)(Output)
Fair allocation
target
Output score
Preference Based(PB) Model in DFM
14. Input 1
Input2
O
A
F
D
C
B
E
F’
Normal DFM projection (TES0 = 1.000)
Non-Attainment DFM projection
(θ*<TES0 <1.000)
CCR(original)-Projection
Target Oriented (TO) Model in DFM
takes for granted a given prior
target-efficiency score (TES).
• This approach is able to calculate an efficient input
reduction value and an efficient output increase value
in order to attain this TES.
15. A Proposal for a PB-TO DFM Model
Distance Friction Minimization (DFM) model
Preference-Based Approach
(PB)
Target-Oriented Approach
(TO)
PB-TO-DFM model
16. Performance assessment of global cities
No. DMU No. DMU No. DMU No. DMU
1 Amsterdam 11 Fukuoka 21 Mumbai 31 Sydney
2 Barcelona 12 Hong Kong 22 New York 32 Taipei
3 Beijing 13 Istanbul 23 Osaka 33 Tokyo
4 Berlin 14 Kuala Lumpur 24 Paris 34 Toronto
5 Boston 15 London 25 San Francisco 35 Vancouver
6 Brussels 16 Los Angeles 26 Sao Paulo 36 Vienna
7 Cairo 17 Madrid 27 Seoul 37 Washington, D.C.
8 Chicago 18 Mexico City 28 Shanghai 38 Zurich
9 Copenhagen 19 Milan 29 Singapore
10 Frankfurt 20 Moscow 30 Stockholm
We refer to the “score by indicator” datasets in the GPCI-2015 report.
These indicator data are converted into a standardized indicator value,
falling in between 100 and 0, so that the data can be evaluated
according to a uniform standard. The highest performance of an
indicator receives a score equal to 100, and the poorest a score of 0.
17. Viewpoint 1: Human environment
(human well-being, labour market and environment)
We consider 1 Input and 4 Outputs :
(I1) Total Employees
(O1) CO2 Emissions
(O2) Nominal GDP,
(O3) Level of Satisfaction of Employees with their Lives,
(O4) Percentage of Renewable Energy Used
38 Global Cities
Level of Satisfaction
Employees
CO2 Emissions
GDP
Renewable
Energy Used
18. Viewpoint 2: Socioeconomic-cognitive activity
(human resources, communication, and cognitive performance)
We consider 3 Inputs and 2 Outputs :
(I1)Interaction Opportunities between Researchers
(I2)Research and Development (R&D) Expenditures
(I3)Number of Employees
(O1)Nominal GDP
(O2)Number of Registered Intellectual Industrial Property Rights (Patents)
38 Global Cities
Employees
GDP Number of Patents
Interaction
Opportunities
R&D Expenditures
19. Efficiency Evaluation Based on Super-Efficiency Model
No ‘double crown winner’ global city
we may need an
efficiency
improvement
projection for
inefficient cities.
20. Illustration of Efficiency Improvement Projection,
Original Model vs DFM (Stockholm)
• CCR; reduction in Total employees by 23.1%, together
with an increase in Satisfaction of Employees of 66.2%
and a reduction in CO2 emission of 74.2%.
• DFM: reduction in Total employees by 13.0%, together
with an increase in Nominal GDP of 18.5%.
• It appears that the empirical ratios of change in the DFM
are smaller than in the CCR (more effective solution).
21. Illustration of Efficiency Improvement Projection,
Original Model vs DFM (Amsterdam)
• These models are able to compute target input and
output values to reach an efficiency score of 1.0; in reality
this may hard to achieve.
• Reduction of R&D expenditures by 86.3% in a CCR model
and by 80.8% in a DFM model… less feasible.
22. Efficiency-Improvement Projection of the PB-TO DFM model
• The previous findings have demonstrated that the pathway to
an efficient outcome may require rather extreme ‘draconic’
measures and strategies.
• We will resort to PB-TO DFM model to explore whether an
intermediate or mitigating strategy is possible in order to arrive
at an entirely efficient city or to a pre-specified target level.
• Amsterdam - socioeconomic-cognitive activity;
OAP is carried out in successive steps from 0.0 to 1.0 with
intervals of 0.1, while the TES is set on 0.600 (note: the present
efficiency score equals 0.426).
• Stockholm - human environment;
OAP uses the same OAP, while the TES is set equal to 0.850
(note: the present efficiency score equals 0.769)
23. • OAP amounting to 0.7 (i.e., 70 percent of the total efficiency gap is
allocated for output, and 30 percent of the total efficiency gap is allocated
for input)
• a reduction in Number of Researchers of 5.9 percent, and an increase in
Nominal GDP of 32.6 per cent are required to raise the efficiency score to
0.600.
Amsterdam: efficiency improvement of socioeconomic-cognitive activity
24. Stockholm: efficiency Improvement of Human Environment
• OAP is equal to 0.3 (i.e. 30 percent of the total efficiency gap is allocated
for output, and 70 percent of the total efficiency gap is allocated for input)
• a reduction in Total employees of 6.3 percent, and an increase in Nominal
GDP of 5.0 percent would be needed to raise the efficiency score to 0.850.
• If such a plan would have an OAP of 1.0 (i.e. 100 per cent of the total
efficiency gap is allocated for output), then even an increase in Nominal
GDP of 14.9% would be required to raise the efficiency score to 0.850.
25. Conclusions
• We have assessed 38high-performance global cities based
on DEA.
• We also have presented a new methodology, the PB-TO
DFM model. This model is able to provide operational and
helpful step-by-step policy information on governance
strategies of global cities.
• The results of this new methodology may provide a
meaningful quantitative contribution to decision making
and planning on the improvement of the performance for
each global city, as illustrated by our case studies, and
hence may reinforce the position of ‘urban empires’ in the
‘New Urban World’.