Modelling socio-economic and energy aspects of urban systems

G Model

ARTICLE IN PRESS

SCS 136 1–15

Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

Contents lists available at ScienceDirect

Sustainable Cities and Society journal homepage: www.elsevier.com/locate/scs

Modelling socio-economic and energy aspects of urban systems

1

2

Q1

Athanasios Dagoumas a,b,∗ a

3

b

4

Electricity Market Operator S.A., 72 Kastoros str., 185 45 Piraeus, Greece Department of European & International Studies, University of Piraeus, 80 Karaoli & Dimitriou str., 185 34 Piraeus, Greece

5

a r t i c l e

6 14

i n f o

a b s t r a c t

7

Keywords: Cities London Energy poverty Climate change Integrated assessment

8 9 10 11 12 13

Q3

15

16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

There is an urgent need to limit greenhouse gas emissions from cities if ambitious mitigation targets are to be met. On the other hand the economic crisis and the ambiguous relationship of inequality with economic growth have raised the issue of energy poverty. The need to connect economic activity with employment, energy poverty, climate change is becoming increasingly recognised. This paper describes the socioeconomic–energy–environmental component of an urban integrated assessment facility developed by the Tyndall Centre for Climate Change Research, which simulates socio-economic change, energy demand, climate impacts and greenhouse gas emissions over the course of the twenty first century at the city scale. The research is focussed upon London, UK, a city that has taken a lead role in the UK and globally with respect to energy poverty and climate protection. The paper demonstrates, through the implementation of several scenarios, quantifiable synergies and conflicts between economic development, employment and energy poverty in order to improve decision making in achieving sustainable and equality outcomes for cities. © 2013 Published by Elsevier B.V.

1. Introduction Urbanisation is not merely a modern phenomenon, but a rapid and historic transformation of human social roots on a global scale. However, over the last century it has been considerably increased in the developed countries and is expected to be further increased in developing countries by 2050 (UN, 2012). As urbanisation is closely linked to modernisation and to industrialisation, the rapid technological developments are expected to accelerate the concentration into cities. Urban areas occupy less than 2% of the Earth’s land surface (Balk, Pozzi, Yetman, Deichmann, & Nelson, 2005), but house just over 50% of the world’s population, a figure that was only 14% in 1900 (Douglas, 1994) and one which is expected to increase to 60% by 2030 (UN, 2012). Urban activities acquire significant amounts of energy and release greenhouse gases (GHGs) that drive global climate change directly (e.g. fossil fuel-based transport) and indirectly (e.g. electricity use and consumption of industrial and agricultural products). Considering that over the last decade, two challenges – namely economic crisis and climate change – have been prioritised in the political and scientific agenda, the role of cities in tackling those

∗ Corresponding author at: Department of European & International Studies, Q2 University of Piraeus, 80 Karaoli & Dimitriou str., 185 34 Piraeus, Greece. Tel.: +30 2104142394/2109466810; fax: +30 2104142328/2109466901. E-mail addresses: [email protected], [email protected], [email protected]

challenges is becoming crucial. The economic crisis and the inequalities among different economic quintiles have raised the issue of energy poverty, namely the lack of access to modern energy services. Modern energy services are crucial to human well-being and to a country’s economic development (IEA, 2010) and yet globally over 1.3 billion people are without access to electricity and 2.6 billion people are without clean cooking facilities (IEA, 2010). The lack of access to modern energy services is a serious hindrance to economic and social development, and must be overcome if the UN Millennium Development Goals are to be achieved. Moreover cities are potential hot spots of vulnerability to climate change impacts by virtue of their high concentration of people and assets. Responding to climate change by mitigating greenhouse gas emissions and adapting to the impacts of climate change is placing new and complex demands upon urban decision makers. The climate drivers that adaptation is responding to will amplify over a period of decades but manifest themselves most vividly in the form of intermittent extreme climate shocks of windstorms, floods, droughts or heat waves. There is increasing understanding of the synergies and conflicts in the objectives of mitigation and adaptation (McEvoy et al., 2006). These interactions are no more vivid than Q4 in urban areas, where they play out through land use, infrastructure systems and the built environment. Without sensible planning, climate change can induce energy intensive adaptations such as air conditioning or desalination, driving higher emissions. Urban decision makers need to understand the implications of these interactions and the potential influences of future global changes. The processes of influence and interaction within urban areas are so

2210-6707/$ – see front matter © 2013 Published by Elsevier B.V. http://dx.doi.org/10.1016/j.scs.2013.11.003

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62

G Model SCS 136 1–15 xxx.e2 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107

108

109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125

ARTICLE IN PRESS A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

complex (Hall et al., 2010), that aspiration for integrated responses to the challenges facing urban areas is widely articulated (Walsh et al., 2011). There is therefore a need for methods and tools that can help to facilitate and inform integrated assessment. Considering that the first step in integrated assessment modelling is the derivation of assumptions about population and economic activity (GDP growth), which will feed then the different more technical models, it is crucial to provide a robust analysis of the socio-economic aspects of urban development. Therefore, a detailed quantitative representation of the socio-economic and energy aspects of urban systems is of high importance, as providing credible input to the different sub-models that consequently face more technical aspects of the urban systems, enables the exploration of feasible solutions of the climate-change and energy poverty problems. The economics for avoiding dangerous climate change requires analysis from many disciplines. The multi-disciplinary risk analysis carried out by the Stern Review team (Stern, 2006) and the IPCC 4th Assessment Report (IPCC AR4, 2007) has revealed critical weaknesses of the traditional, neoclassical approach, especially as regards the treatment of uncertainty and risks such that equilibrium-based models by themselves are not considered appropriate for providing an adequate understanding of the climate change problem (Beinhocker, 2006; DeCanio, 2003). For these reasons, an approach has been adopted that emphasises the dynamics of the energy–environment–economy system in its historical setting, using a UK model that incorporates many aspects of the “new economics” described in Barker (2008). This paper aims to describe in more detail the socioeconomic–energy–environmental branch of an integrated assessment model (Hall et al., 2010; Walsh et al., 2011) of the Greater London Area (GLA), together with a scenario analysis of different socio-economic futures of the GLA, including London, South-East and East England, over the period 2000–2100. A useful and important contribution of the paper, is that due to high uncertainty of the economic structure in the future, especially when examining 2100, the analysis is made by classifying the 41 economic sectors of the UK economy on 8 aggregate sectors, based on their technological characteristics and on the likely effects of three pervasive technologies (information technology, biotechnology and nanotechnology). This analysis is proving to be of value in stimulating more integrated thinking about cities and informing complex decision making problems. Therefore, a decision maker can take signals from this analysis, on which economic sectors can provide higher economic output, how employment issues can be tackled at macro level and how energy policies can tackle energy poverty issues.

2. Modelling framework For the needs of this paper, the MDM-E3 model (Barker & Peterson, 1987) of the UK economy has been used and developed. MDM-E3 is a very detailed, integrated energy–environment–economy (E3) model (Hall et al., 2009) of the UK economy. The model has been developed by the University of Cambridge (4CMR) and Cambridge Econometrics (CamEcon), a leading economic consultancy. It is one of a suite of E3 models:MDM-E3: Multisectoral Dynamic Model of the UK Economy, including energy–environment–economy (E3) interactions, E3ME: E3 Model of Europe and E3MG: E3 Model at a Global level. All three follow the same overall principles in their economics, construction and operation and are appropriate in exploring longterm policies of the UK economic–energy–environment system (Dagoumas & Barker, 2010). The model is a demand-driven model and as it is based on Post Keynesian macro-economics, it assumes that equilibrium is not a useful concept for market analysis, unlike the General Equilibrium approach. It is a dynamic simulation

model, putting an emphasis on ‘history’, as it is based on time series and cross-section data, using input–output data from Office of National Statistics (ONS). It uses cointegration techniques to identify long-run trends in 22 sets of equations. It is a structural and hybrid model, where the disaggregation of the variables is further extended in the submodels, focusing for a more detailed representation of the UK economy and especially for the energy system as there exist several relevant submodels for (1) Energy demand and fuel-shares, (2) the electricity sector, including an Energy Technology Model (ETM), (3) CHP, (4) Household energy use and (5) Transport energy use and emissions. The model contains 41 production sectors, which enables a more accurate representation of the effects of policies than is common in most macroeconomic modelling approaches. The model addresses the issues of energy security and climate stabilisation both in the medium and long terms, with particular emphasis on dynamics, uncertainty and the design and use of economic instruments, such as emission allowance trading schemes. MDM-E3 is a non-equilibrium model with an open structure such that labour, foreign exchange and public financial markets are not necessarily closed. It is disaggregated, with 12 energy carriers, 19 energy users, 28 energy technologies, 14 atmospheric emissions and 41 production sectors, with comparable detail for the rest of the economy. The model represents a novel long-term economic modelling approach in the treatment of technological change, since it is based on cross-section and time-series data analysis using formal econometric techniques, and thus provides a different perspective on stabilisation costs. The author, within the Cambridge Centre for Climate Change Mitigation Research (4CMR), has extended the MDM-E3 model up to 2100 (Hall et al., 2009, 2010) with an overall objective to provide output tables of economic activity with regional and industrial disaggregation (measured in terms of economic value added at constant prices), of employment with regional and industrial disaggregation (measured in terms of full-time-equivalent – FTE – employees) and of energy demand (in terms of thousands tonnes of oil equivalent (toe) consumed by different fuel type) at national level with industrial disaggregation as input to the land use and population distribution model, the transport emissions accounting model and the energy use emissions accounting model.

2.1. Description of the treatment of aggregate energy demand in MDM-E3 For the energy demand, a 2-level hierarchy is being adopted. A set of aggregate demand equations on annual data covering 19 fuel users/sectors and 12 UK regions (where GLA is one of them) is estimated and is then shared out among main fuel types (coal, heavy fuel oil, natural gas and electricity) assuming a hierarchy in fuel choice by users: electricity first for “premium” use (e.g. lighting, motive power), non-electric energy demand shared out between coal, oil products and gas. The energy demand for the rest of the 12 energy carriers is estimated based on historical relations with the main 4 energy carriers. An autoregressive distributed lag (ARDL) model is developed for modelling energy consumption, considering also that some users’ aggregate demands are affected by upward movements in relative prices only (ratchet or asymmetrical price effects). Historical data for the last 40 years are used, while an error correction model (ECM) distinguishes between long-term and adjustment parameters. A long-term behavioural relationship is identified from the data and embedded into a dynamic relationship allowing for short-term responses and gradual adjustment (with estimated lags) to the long-term outcome. The equations and identities are solved iteratively for each year, assuming adaptive expectations, until a consistent solution is obtained.

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165

166 167

168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188

G Model SCS 136 1–15

ARTICLE IN PRESS A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

216

The aggregate energy demand is affected by industrial output of user industry, household spending in total, relative prices, temperature, and technical progress indicator. Energy efficiency improvements are modelled through relevant investment, augmented by time trends and/or accumulated. The economy aggregates, such as GDP, are found by summation. This enables representation of the wider macroeconomic impacts of policies focused on particular sectors, including rebound effects (Barker, Dagoumas, & Rubin, 2009). These long-run energy demand equations incorporate explanatory variables that represent an indicator of activity, relative prices (relative to GDP deflators for energy) and a technological progress indicator (TPI). TPI is measured by accumulating past gross investment enhanced by R&D expenditures with declining weights for older investment. The original energy demand equations are based on work by Barker, Ekins, and Johnstone (1995) and Hunt and Manning (1989). Since there are substitutable inputs between fuels, the total energy demand in relation to the output of the fuel-using industries is likely to be more stable than the individual components. This total energy demand is also subject to considerable variation, which reflects both technical progress in conservation, and changes in the cost of energy relative to other inputs. Aggregate and disaggregate energy-demand equations’ specifications follow similar lines including economic activity, technology, relative price effects, spending and R&D investment and are in the process of being respecified so as to also capture the temperature effect. As an activity measure, gross output is chosen for most sectors, but household energy demand is a function of total consumers’ expenditure.

217

2.2. Description of the treatment of the labour market in MDM-E3

189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215

218 219 220 221 222 223 224 225 226 227 228 229 230 231 232

233 234 235 236 237 238 239 240 241 242 243 244 245

246 247 248 249 250

The energy–environment–economy (E3) model of the UK, MDM-E3, includes a detailed treatment of the labour market with stochastic equations for employment (as a head count) (Lee, Pesaran, & Smith, 1990) and average wages (Lee & Pesaran, 1993), both by industrial sector and region, for hours worked (Wilson & Bosworth, 1990; Wilson, Homenidou, & Dickerson, 2006) by industrial sector and for labour market participation by gender and regions. This treatment plays an important role in analysis involving tax switches or inflation, particularly in cases where tax revenues are recycled through reductions in taxes on labour. Unemployment is calculated as the difference between employment and the active labour force and is a key explanatory variable in the equations for wages, labour market participation and trade. Unlike equilibrium models, MDM-E3 does not assume full employment, even in the long run. 2.2.1. The employment equations Employment is modelled as a total headcount number for each industry and region and is explained as a function of industry output, wages and hours worked. The theory underlying the equations is described in (Lee et al., 1990), although we now interpret the relationship in terms of institutional behaviour, rather than in a neoclassical factor demand framework. Industry output is assumed to have a positive effect on employment, while the effect of higher wages and longer working hours is assumed to be negative. Depending on available data, it is possible to estimate different equations for male and female employment, but this is not done at present, and the allocation of total employments across gender and age groups is done on the basis of projecting historical trends. 2.2.2. The hours worked equations Hours worked is a simple equation, where average hoursworked by industry is a function of “normal hours-worked” (hours worked in other industries). The resulting estimate of hours worked is an explanatory variable in the employment equation (see above).

xxx.e3

Hours worked is defined as an average across all workers in an industry, so incorporates the effects of higher levels of part-time employment in certain industries. 2.2.3. The wage equations The wage equations are based on the theory of real wage bargaining (Lee & Pesaran, 1993). The wage rates for an industry in a region depend on productivity effects and prices and wage rates in the wider economy. Other important factors include unemployment, tax rates and cyclical effects. Generally it is assumed that higher prices and productivity will push up wage rates, but rising unemployment will reduce wages. A single average wage is estimated for each region and sector. The estimates of average wages are a key input to both the employment equations and the price equations in MDM-E3.In the absence of growing output, rising wages will increase overall unit costs and industry prices. These prices may get passed on to other industries (through the input–output relationships), building up inflationary pressure. 2.2.4. The labour market participation equations Labour market participation is estimated as a rate between 0 and 1 for male, female and total working-age population. At present there is no disaggregation by age groups. Labour market participation is a function of output, wages, unemployment and benefit rates. Participation is assumed to be higher when output and wages are growing, but falls when unemployment is high, or benefits create a disincentive to work. In addition, there is a measure of economic structure and the relative size of the service sector of the economy; this has been found to be important in determining female participation rates. The participation rates determine the stock of employment available (by multiplying by workingage population, which is exogenous). This is an important factor in determining unemployment, which in turn feeds into wages and back to labour market participation. 3. Scenarios Scenarios represent alternative storylines of the future rather than predictions or forecasts and are used to assist in the understanding of the behaviour and long term changes to complex systems to support policy making. Scenarios provide an internally consistent and reproducible set of assumptions about the key relationships and driving forces of change in order to integrate qualitative narratives of future global change and quantitative estimates of future emissions, economic growth and population change scenarios. The model has been extended to examine the evolution of industrial economic output and employment for all 12 MDM-E3 regions focusing mainly on three regions in Greater London area: London, South East England (SE) and East England (EE). Three main scenarios were designed, representing base, low and high growth scenarios. A sensitivity analysis was carried out on the number of average working hours per week, disaggregated at industrial but not at regional level. As mentioned above on the description of the labour market in MDM-E3 model, the average working hours per week variable is used as an explanatory variable in the employment equation. An increase of part-time jobs leads to a reduction in the average working hours per week, leading then to an increase of employment, measured in full-time equivalent jobs. Data from national or international sources (ILO; Maddison, 2005; ONS) shows that during the last 50 years average working hours have decreased from the levels of 44–39 h per week in the period 1950–2000 in most developed countries. But during the period 1985–2005, average working hours remained at the levels of 39 h per week, where for developed countries some sectors showed a slight decrease e.g. services while other sectors (e.g. heavy

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

251 252 253

254 255 256 257 258 259 260 261 262 263 264 265 266 267

268 269 270 271 272 273 274 275 276 277 278 279 280 281 282

283

284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312

G Model SCS 136 1–15 xxx.e4 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334

335 336 337 338 339 340 341 342 343 344 345

346 347 348 349 350 351

352

353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373

ARTICLE IN PRESS A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

industries) showed slight increase for some years, due mainly to a loss of competitiveness compared to developing countries’ relevant industries. So, the average working hours depend on the future structure of each economy. A recent study (Poncet, 2006) projects that working hours per week will have a differentiation even between developing countries, depending on the evolution of each sector. In our analysis we assume that UK will continue to be competitive in the services sectors, while the traditional industrial sectors will steadily decrease, if not be expunged, as UK will become a consuming country for these sectors’ products. Based on this consideration and on the fact that during the last years the average working hours in UK remained almost stable, we assume firstly that working hours will remain stable in period up to 2020 and secondly a steady decrease in the period 2021–2100. We assume that the trend in working hours will be the same for all sectors, leading to a decrease of 25% in year 2100 compared to year 2000. This assumption considers the results of a relevant work (Poncet, 2006). We carry out a sensitivity analysis on the evolution of the average working hours, applicable to all sectors and regions, in order to examine the influence of this variable on the employment of each sector and region. Summarising, the scenarios examined are I. Baseline scenario: GDP growth rate is projected to slow steadily from an average of 2.5% each year over 2010–2020 to 1.5% for UK by the year 2100. In the different regions, London growth rate is slightly lower to national growth rate, while South East and East England have higher growth rates. II. Low growth scenario: GDP growth rate at national and regional level is 0.3% less than in the baseline scenario, steadily slowing to the rate of 1.2% for UK by the year 2100. III. High growth scenario: GDP growth rate at national and regional level is 0.3% higher than in the baseline scenario, steadily slowing to the level of 1.8% for UK by the year 2100. In all 3 scenarios, average working hours per week are assumed to be decreased by 25% in the period 2000–2100. A number of sub-scenarios are examined in order to explore the creation of part-time jobs, influenced by the average working hours per week. Fig. 1 shows the evolution of this variable. All scenarios and their assumptions are summarised in Table 1. 4. Discussion of results Long-term forecasting of sectoral growth is sensitive to technological diffusion. The socio-economic system seems to be characterized by on-going fundamental change. Therefore, it is necessary to consider the dynamic processes of socio-economic development. MDM-E3 incorporates endogenous technological change and so it can examine the influence of technological development in the different sectors. But because of the long-term forecasting period (100 years), any adequate forecast of future economic structure requires us to assess the risk related to the maturity of the new technologies and also to the timing of their entering the market compared to competitive technologies. Although MDME3 belongs to the family of the evolving “new economics” models enabling a ‘positive feedback’ which means that an earlier market availability of a technology is a crucial factor concerning its market penetration compared to competitive technologies, it is not certain which technologies will be developed first and to what extent. Based on the technology evolution, some sectors may to shrink and even die, while new firms/industries will emerge rapidly, as it happened the last 30 years with the computer services and the telecommunications. For this reason it is safer to aggregate the sectors into broad technological categories, considering the likely

effects of technological change within those sectors. The MDM 41 sectors are classified on 8 aggregate sectors, listed in Table 2, based on their technological characteristics and on the likely effects of three pervasive technologies (information technology, biotechnology and nanotechnology) on their input–output structure (Dewich, Green, & Miozzo, 2004; Dewick, Green, Fleetwood, & Miozzo, 2006). These 8 categories are most likely to remain relevant in the future even if many of the 41 industries that currently define these categories do not. The incorporation of this technological taxonomy of sectors in the MDM-E3 model has been done by aggregating the relevant sectors into the final 8 sectors for all equations of the model. The econometric equations have been re-estimated in order to consider the new structure of the model. Therefore the model has been practically transformed from a 42 sectors-model to an 8 sectors-model. Figs. 2–11 show the evolution of the employment and of the gross value added for each sector for London region over the period 2000–2100. For practical reasons, it has been selected to represent results only for London and not for the other two regions (East England, South-East England), as was done for the needs of the Tyndall research project. Moreover, from the whole range of scenarios examined, again for practical reasons few of them have been selected to be presented. The first important conclusion derives from the comparison of the growth rate in sectoral Gross Value Added (GVA) for the three examined regions. London is predicted to have a growth rate at the level of 2.5–3% up to 2060, which slows steadily to the rate of 1.4% in year 2100. This growth rate is similar to national projected growth rate. On the other hand, the other two regions (South East and East England) are projected to have higher growth rates, which are at the level of 3% up to 2060, steadily slowing to the rate of 1.8% in 2100. The growth rate of East England is slightly higher, by 0.1%, than that of South East England almost for the whole examined period. The effect of different regional GVA growth rates on the regional employment is obvious, especially in case of London where its total employment is projected to be decreased in the baseline, low and high growth scenarios. For the other two regions, total regional employment is projected to be increased or stabilised for the same scenarios. The evolution of ‘worked hours per week’ variable affects total employment, as the total regional employment is increased/decreased by a further 2–2.5% when this variable is projected to be decreased by 35%/15% compared to 25% of the baseline scenario by 2100 respectively for the three regions. The same conclusion comes when examining the effect of this variable on the low and high growth scenarios, irrespective of the region. This variable, which is related to the number of part-time jobs, have a considerable effect on total employment but its influence is relative smaller than that of GVA growth rates. From a sectoral perspective, Scale Intensive Information Networks category, including banking/finance, communications, professional, business and other services, are projected to have very high growth rates, becoming by far the most dominant category for all three regional economies concerning output. Very high growth rates are projected also for the Science Base Service Suppliers category, which becomes the second biggest category beyond 2060 in GVA. Both these categories are project to require high productivity personnel, which is attributed by the lower increase growth rate in employment. On the other hand Supplier Dominated General and Scale Intensive General categories which include traditional heavy industries show low or negative growth rates, as UK and its regions become consumers rather producers for these sectors. Furthermore scale intensive physical networks and Supplier Dominated Services categories, including transportation, education, hotels and public administration show a considerable increase in their output. Within the Supplier Dominated Services category there is projected a different evolution curve in employment, as hotels, public

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439

G Model

ARTICLE IN PRESS

SCS 136 1–15

A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

xxx.e5

Table 1 Assumptions on scenarios examined with the MDM-E3 model. Scenario name

Growth rate % for London

Growth rate % for South Eastern England

Growth rate % for Eastern England

BASa

2.95% in 2020, reduced at 2.42% in 2060 and at 1.37% in 2100 As above As above As above As above

3.09% in 2020, reduced at 2.83% in 2060 and at 1.74% in 2100

2.98% in 2020, reduced at 2.89% in 2060 and at 1.87% in 2100 As above As above As above As above

2.65% in 2020, reduced at 2.11% in 2060 and at 1.07% in 2100 As above As above As above As above

2.78% in 2020, reduced at 2.53% in 2060 and at 1.43% in 2100

3.26% in 2020, reduced at 2.73% in 2060 and at 1.67% in 2100 As above As above As above As above

3.4% in 2020, reduced at 3.14% in 2060 and at 2.04% in 2100

BAS1 BAS2 BAS3 BAS4 LOW

LOW1 LOW2 LOW3 LOW4 HIGH

HIGH1 HIGH2 HIGH3 HIGH4

As above As above As above As above

As above As above As above As above

As above As above As above As above

2.67% in 2020, reduced at 2.59% in 2060 and at 1.57% in 2100 As above As above As above As above 3.29% in 2020, reduced at 3.2% in 2060 and at 2.18% in 2100 As above As above As above As above

Decrease % in average working hours per week in period 2000–2100b 25%

15% 20% 30% 35% 25%

15% 20% 30% 35% 25%

15% 20% 30% 35%

a

UK GDP growth rate is similar to London growth rate (at the level of 1.5% in 2100). 25% means working hours per week are reduced by 25% up to 2100. This reduction is applied to all industries/sectors. Reduction in the variable that represents average working hours per week, means that more part-time jobs are available. This has an effect on the variable that represents full-time-equivalent (FTE) employment. b

Table 2 Taxonomy of MDM 42 sectors into 8 aggregate sectors based on their technological characteristics. MDM-E3 sectors

Technological taxonomy of sectors

1 Agriculture, etc. 2 Coal 3 Oil and Gas, etc. 4 Other Mining 5 Food, Drink and Tobacco 6 Textiles, Clothing and Leather 7 Wood and Paper 8 Printing and Publishing 9 Manufacturing Fuels 10 Pharmaceuticals 11 Chemicals 12 Rubber and Plastics 13 Non-Metallic Mineral Products 14 Basic Metals 15 Metal Goods 16 Mechanical Engineering 17 Electronics 18 Electrical Engineering and Instruments 19 Motor Vehicles 20 Other Transport Equipment 21 Manufacturing 22 Electricity 23 Gas Supply 24 Water Supply 25 Construction 26 Distribution 27 Retailing 28 Hotels and Catering 29 Land Transport, etc. 30 Water Transport 31 Air Transport 32 Communications 33 Banking and Finance 34 Insurance 35 Computing Services 36 Professional Services 37 Other Business Services 38 Public Administration and Defence 39 Education 40 Health and Social Work 41 Miscellaneous Services

Supplier Dominated General Supplier Dominated General Supplier Dominated General Supplier Dominated General Scale Intensive General Supplier Dominated General Scale Intensive General Supplier Dominated General Scale Intensive General Science Based General Scale Intensive General Scale Intensive General Scale Intensive General Scale Intensive General Scale Intensive General Specialised Suppliers General Science Based General Scale Intensive General Scale Intensive General Specialised Suppliers General Supplier Dominated General Supplier Dominated General Supplier Dominated General Supplier Dominated General Supplier Dominated General Scale Intensive Physical Networks Scale Intensive Physical Networks Supplier Dominated Services Scale Intensive Physical Networks Scale Intensive Physical Networks Scale Intensive Physical Networks Scale Intensive Information Networks Scale Intensive Information Networks Scale Intensive Information Networks Science Based Service Suppliers Scale Intensive Information Networks Scale Intensive Information Networks Supplier Dominated Services Supplier Dominated Services Supplier Dominated Services Scale Intensive Information Networks

SDG SDG SDG SDG SIG SDG SIG SDG SIG SBG SIG SIG SIG SIG SIG SSG SBG SIG SIG SSG SDG SDG SDG SDG SDG SIPN SIPN SDS SIPN SIPN SIPN SIIN SIIN SIIN SBS SIIN SIIN SDS SDS SDS SIIN

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

G Model SCS 136 1–15 xxx.e6

ARTICLE IN PRESS A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

Fig. 1. Mean working hours per week over the period 2000–2100 (assuming different decrease rate, from 15 to 35% over the period 2000–2100).

Fig. 2. Employment in London by sector over the period 2000–2100 (in thousands) for the BAS scenario.

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

G Model SCS 136 1–15

ARTICLE IN PRESS A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

xxx.e7

Fig. 3. Employment in London by sector over the period 2000–2100 (in thousands) for the LOW scenario.

440 441 442 443 444

administration and defence sectors have a significant decrease in projected jobs while health and by a lesser extent education are projected to create new jobs. Finally Scale Intensive General, Science Based General and Specialised Suppliers General categories are projected to have more moderate increase growth rates, which

are enough to allow them cover a small but almost constant percent of the total output and employment for the whole examined period 2000–2100. The influence of the growth rate in GVA, reflected by the establishment of base, high and low growth scenarios, and the influence

Fig. 4. Employment in London by sector over the period 2000–2100 (in thousands) for the HIGH scenario.

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

445 446 447 448 449

G Model SCS 136 1–15 xxx.e8

ARTICLE IN PRESS A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

Fig. 5. Employment in London by sector over the period 2000–2100 (in thousands) for the BAS1 scenario.

450 451 452 453 454

of the creation of part-time jobs, reflected by the establishment of different decrease rates in the ‘worked hours per week’ variable, on the sectoral employment for the different sectors is similar to its influence on the total employment discussed above. This means that the different scenarios (baseline, low and high growth)

together with their variants have been established so as to illustrate some feasible ranges of the regional and sectoral output and employment. Figs. 12 and 13 provide the final energy demand for each sector and for each energy type over the period 2000–2100. Fig. 14 focuses

Fig. 6. Employment in London by sector over the period 2000–2100 (in thousands) for the BAS4 scenario.

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

455 456 457 458 459

G Model SCS 136 1–15

ARTICLE IN PRESS A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

xxx.e9

Fig. 7. Gross value added (GVA) in London by sector over the period 2000–2100 (in thousands £) for the BAS scenario.

Fig. 8. Gross value added (GVA) in London by sector over the period 2000–2100 (in thousands £) for the LOW scenario.

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

G Model SCS 136 1–15 xxx.e10

ARTICLE IN PRESS A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

Fig. 9. Gross value added (GVA) in London by sector over the period 2000–2100 (in thousands £) for the HIGH scenario.

460 461 462 463 464

on the transportation sector, where it is obvious that the road transportation is fully electrified beyond 2050. This comes from the model assumptions enabling endogenous positive feedback which accelerates the rate with which a new alternative technology or energy type penetrates into the system. Moreover, it has to be

considered, that compared to other energy studies, hydrogen is not considered to take a share into the market, considering the current R&D status and the current deficiencies of the hydrogen cars compared to the electric cars. The results concerning energy demand are not provided for London but for the UK, as the energy

Fig. 10. Gross value added (GVA) in London by sector over the period 2000–2100 (in thousands £) for the BAS1 scenario.

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

465 466 467 468 469

G Model SCS 136 1–15

ARTICLE IN PRESS A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

xxx.e11

Fig. 11. Gross value added (GVA) in London by sector over the period 2000–2100 (in thousands £) for the BAS4 scenario.

Fig. 12. Final energy demand for each sector over the period 2000–2100 for the UK (in ktoe) for the BAS scenario.

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

G Model SCS 136 1–15 xxx.e12

ARTICLE IN PRESS A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

Fig. 13. Final energy demand for each energy type over the period 2000–2100 for the UK (in ktoe) for the BAS scenario.

Fig. 14. Final energy demand of transportation for each energy type over the period 2000–2100 for the UK (in ktoe) for the BAS scenario.

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

G Model SCS 136 1–15

ARTICLE IN PRESS A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508

509 510 511 512 513 514

515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533

sub-model of the MDM-E3 model does not incorporate regional detail. It would be expected that the load profile of the London does not change significantly; therefore Figs. 12–14 can be considered representative of energy demand trend for the London area. Therefore, from Figs. 12–14, clear results concerning urban energy demand can be taken out. Fig. 12 shows that the “Residential” and the “Other final use” (which incorporates the Services) sectors have the highest growth in energy demand, dominating the mix in final energy demand by 2100. Fig. 13 shows that electricity and natural gas show a considerable increase, and together with combustible waste are almost the only fuels left by 2100. Combining Figs. 12–14, the following assumptions can be taken out: a full electricification of urban transportation is expected, natural gas will dominate the urban energy demand for heat and other uses (hot water, cooking, etc.), and energy demand in the cities is expected to be increased higher compared to the other sectors. Although, due to the different level of data availability concerning economic activity and energy consumption, the direct linkage of all economic sectors with the final energy demand is not possible, the MDM-E3 can provide critical insights on which sectors or category of sectors is more energy intensive. Therefore, decision makers can have insights on how economic sectors affect employment and energy consumption at macro level. The paper for practical reasons does not provide further graphs on the evolution of energy demand over the different scenarios, described above. This arrives from the fact that already Figs. 2–11 show the effect of factors affecting the economic activity and the total employment. Considering that, within the MDM-E3 model, energy demand equations are linked with economic activity it derives that the effect of the GDP rate increase will proportionally (the estimated economic elasticities are in the range of 0.7, diversifying per sector and energy type) affect the rate of the energy demand increase. Similarly, considering that, within the MDM-E3 model, the employment equations are linked with economic equations, the change in average working hours per week will affect the full-time employment (similarly all over sectors) and consequently will affect the energy demand. The IEA uses an energy development index in measuring energy poverty (IEA, 2010), combining three indicators: • the share of households using cleaner, more efficient cooking and heating fuels (liquefied petroleum gas, kerosene, electricity and biogas), • the share of households with access to electricity, • electricity consumption per capita. This indicator is used to capture the level of overall energy development. The analysis provided through the MDM-E3 model enables the examining the evolution of only the third indicator. Fig. 15 provides the evolution of the electricity consumption per capita for the UK over the period 2000–2100 for the Baseline, Low growth and High growth scenarios. The figures depict that the electricity consumption is expected to be increased much faster than the population, depicting the evolution of electricity use in the whole economy. The considerable increase in this indicator is highly related to the fact that, as Figs. 13 and 14 show, the transportation is expected to be fully electrified. Moreover, Fig. 16 provides an new indicator of energy poverty, namely the evolution of the electricity consumption per Full-Time-Equivalent (FTE) employee for the UK over the period 2000–2100 for the baseline, low growth and high growth scenarios. This figure again depicts the expected growth in electricity consumption compared to the evolution of employment and parallel, by comparing Figs. 15 and 16, it provides the linkages of economic and employment growth. Therefore, the MDM-E3 model can be used as a tool of analysing the level of overall energy development at macro level, and if used in applying portfolio of energy

xxx.e13

polices (Dagoumas & Barker, 2010), economic or employment policies it can depict how those policies affect energy poverty at macro level. Moreover, considering that the nature of the MDM-E3 enables the examination of the complementarity of policies, this would enable the proper selection of policies that tackle energy poverty issues at the macro level. However, the drawback of a top-down analysis, as such that is undertaken for this paper, is that it does not enable to have a detailed picture on the different economic quintiles. This would enable a more detailed analysis of energy poverty issues. However, the analysis provided in this paper can be used as a useful tool for decision making for energy poverty issues, as it provides the trends and the linkages between economic activity, employment and energy demand. Moreover, the paper faces the high degree of uncertainty, which is resulting from the fact that the analysis period covers 100 years, by aggregating the economic sectors into 8 sectors and by providing sensitivity analysis on the crucial factors such as the GDP growth rate and the average working hours which affects employment. Therefore, a decision maker can take signals from this analysis, on which economic sectors can provide higher economic output, how employment issues can be tackled at macro level and how energy policies can tackle energy poverty issues, e.g. by enhancing the support of certain energy types that are not vulnerable on energy crises. The results of this analysis can be further useful when seen within the framework of an integrated system analysis. The MDME3 model is a vital component of this integrated assessment framework, as this socio-economic component together with a global climate component provide the boundary conditions for the city scale analysis, in the case of London. These boundary conditions drive scenarios of regional economy and land use change, ensuring that whilst they are influenced by local policy, these scenarios are also globally consistent. It is at the level of land use modelling that the analysis becomes spatially explicit. Scenarios of land use and city-scale climate and socio-economic change inform the emissions accounting and climate impacts modules. The final component of the framework is the integrated assessment tool that provides the interface between the modelling components, the results and the end-user. A foremost challenge facing urban planners and decision makers today is the need to deliver plans, designs and strategies to support economic sectors with high gross value added, to support the creation of new jobs, to tackle energy and climate problems such as energy poverty and meeting of GHG mitigation targets. It is important to recognise that their implementation varies and a more informed understanding of their synergies, conflicts and trade-offs is essential in developing a more integrated climate policy and integrated portfolio of effective management options for our urban environments. The complexity of the urban system requires the development of integrated assessment models, such as the Tyndall Centre’s Urban Integrated Assessment Facility (UIAF). This tool brings together long-term projections of demography, economy, land use, climate impacts and GHG emissions within a coherent assessment framework. Indeed, whereas the UIAF has already demonstrated the capability to test adaptation and mitigation policies, this paper explores in more detail crucial components of the integrated system, aiming at providing insights at socio-economic and energy issues. Considering that the UIAF is now being used to help inform decision making surrounding the new London Plan, those further insights can enable development of integrated portfolios of options in the transition to decarbonised and energy equal cities. This paper aims to describe in more detail a crucial component of the UIAF, which examines socio-economic, energy and environmental aspects of London. It therefore provides an overview of the MDM-E3 model of the UK economy, which has been further

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599

G Model SCS 136 1–15 xxx.e14

ARTICLE IN PRESS A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

Fig. 15. Final electricity demand per capita over the period 2000–2100 for the UK (in MWh/capita) for the BAS, LOW and HIGH scenarios.

Fig. 16. Final electricity demand per full-time-equivalent (FTE) employee over the period 2000–2100 for the UK (in MWh/employee) for the BAS, LOW and HIGH scenarios.

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

G Model SCS 136 1–15

ARTICLE IN PRESS A. Dagoumas / Sustainable Cities and Society xxx (2013) xxx.e1–xxx.e15

600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636

Q5 637 638

639

developed in order to provide forecast up to 2100. A number of scenarios are applied aiming at providing the trends and insights on the linkages between economic activity, employment and energy demand, and therefore on energy poverty issues, although the topdown “nature” of the model does not enable to have a detailed picture on the different economic quintiles. Moreover, the paper faces the high degree of uncertainty, which is resulting from the fact that the analysis period covers 100 years, by aggregating the economic sectors into 8 sectors and by providing sensitivity analysis on the crucial factors such as the GDP growth rate and the average working hours that affects employment. Therefore, a decision maker can take signals from this analysis, on which economic sectors can provide higher economic output, how employment issues can be tackled at macro level and how energy policies can tackle energy poverty issues, e.g. by enhancing the support of certain energy types that are not vulnerable on energy crises. The paper demonstrates, through the implementation of several scenarios, how the socioeconomic–energy–environmental component of the urban integrated assessment facility quantifies synergies and conflicts between economic development, employment and energy poverty in order to improve decision making and to facilitate the development of portfolios of planning policies that together have a realistic prospect of achieving sustainable and equality outcomes for cities. Considering the long forecast period, the high degree of uncertainty in forecasts needs to be considered. Therefore, the MDM-E3 model is extended in order to incorporate treatment of uncertainty in crucial variables, related to energy pricing, economic growth and labour productivity. This will enable a more probabilistic analysis of the socio-economic and energy aspects of the future urban systems. Another important limitation that will be considered in on-going and future developments of the model is to incorporate more detail in critical parameters related to energy poverty issues, such as economic quintiles, tariffs and subsidies policies for energy commodities. Those developments would enable a more robust analysis of certain energy poverty policies, rather than simply providing insights and trends. Uncited reference Government (2004). Acknowledgements

643

This paper has been prepared as a contribution to the research of the Tyndall Centre for Climate Change Research. The author is grateful for the support of the Centre and of the funding from the UK Research Councils.

644

References

640 641 642

645

xxx.e15

Balk, D., Pozzi, F., Yetman, G., Deichmann, U., & Nelson, A. (2005). The distribution of people and the dimension of place: Methodologies to improve the global estimation of urban extents. In Proceedings of the conference on Urban Remote Sensing, International Society for Photogrammetry and Remote Sensing Tempe, AZ, Barker, T. (2008). The economics of avoiding dangerous climate change. Climatic Change, 89, 173–194. Barker, T., & Peterson, W. (1987). The Cambridge multisectoral dynamic model of the Q6 British economy. Cambridge University Press. Barker, T. S., Ekins, P., & Johnstone, N. (1995). Global warming and energy demand. London: Routledge. Barker, T., Dagoumas, A., & Rubin, J. (2009). The macroeconomic rebound effect and the world economy. Energy Efficiency, 2, 411–427. Beinhocker, E. (2006). The origin of wealth: Evolution, complexity and the radical remaking of economics. Random House Business Books. CamEcon, Cambridge Econometrics limited. Retrieved from http://www.camecon.com/ Dagoumas, A., & Barker, T. (2010). Pathways to a low-carbon economy for the UK with the macro-econometric E3MG model. Energy Policy, 38, 3067–3077. DeCanio, S. J. (2003). Economic models of climate change: A critique. New York: Palgrave MacMillan. Dewich, P., Green, K., & Miozzo, M. (2004). Technological change, industry structure and the environment. Futures, 36, 267–293. Dewick, P., Green, K., Fleetwood, T., & Miozzo, M. (2006). Modelling creative destruction: Technological diffusion and industrial structure change to 2050. Technological Forecasting and Social Change, 73, 1084–1106. Douglas, I. (1994). Human settlements. In W. B. Meyer, & B. L. Turner (Eds.), Changes in land use and land cover: A global perspective (pp. 149–169). Cambridge: Cambridge University Press. Actuary’s Department GAD (2004). Retrieved from Government http://www.gad.gov.uk/Population/index.asp?v=Table &pic=2004|uk|totpop. Hall, J. W., Dawson, R. J., Walsh, C. L., Barker, T., Barr, S. L., Batty, M., et al. (2009). Engineering Cities: How can cities grow whilst reducing emissions and vulnerability? UK: The Tyndall Centre for Climate Change Research. Hall, J., Dawson, R., Barr, S., Batty, M., Bristow, A., Carney, S., et al. (2010). City-scale integrated assessment of climate change: Impacts, mitigation and adaptation. In R. K. Bose (Ed.), Energy efficient cities: Assessment tools and benchmark practices. USA: The World Bank. Hunt, L., & Manning, N. (1989). Energy price- and income elasticities of demand: Some estimates for the UK using the cointegration procedure. Scottish Journal of Political Economy, 36, 183–193. IEA (International Energy Agency). (2010). Energy poverty: How to make modern energy access universal. Special early excerpt of the World Energy Outlook Edition ILO, International Labour Organizations. Retrieved from http://www.ilo.org IPCC. AR4. http://www.ipcc.ch/ Lee, K. C., & Pesaran, M. H. R. (1993). The role of sectoral interactions in wage determination in the UK economy. Economic Journal, 103, 21–55. Lee, K. C., Pesaran, M., & Smith, H. R. (1990). Aggregation bias and labour demand equations for the UK economy. In T. S. Barker, & M. H. Pesaran (Eds.), DisaggreQ7 gation in economic modelling. Routledge. Maddison, A. (2005). Evidence submitted to the Select Committee on Economic Affairs, House of Lords. London: For the inquiry into Aspects of the Economics of Climate Change. ONS, Office for National Statistics. Retrieved from http://www.ons.gov.uk Poncet, S. (2006). The long term growth prospects of the world economy: Horizon 2050. CEPII. No 2006-16 Stern Review on the Economics of Climate Change. (2006). HM treasury. UN (United Nations). (2012). World urbanisation prospects: The 2011 revision. New York: United Nations Publications. Walsh, C. L., Dawson, R. J., Hall, J. W., Barr, S. L., Batty, M., Bristow, A. L., et al. (2011). Assessment of climate change mitigation & adaptation in cities. ICE – Urban Design and Planning, 164(2), 75–84. Wilson, R., & Bosworth, D. L. (1990). Hours of work. Book Chapter. In M. Gregory, & A. Thompson (Eds.), A portrait of pay 1970–82: Analysis of the new earnings survey. Oxford: Oxford University Press. Wilson, R., Homenidou, K., & Dickerson, A. (2006 February). Working futures 2004–2014. National Report, SSDA.

4CMR, Cambridge Centre for Climate Change Mitigation Research, Dept. of Land Economy, University of Cambridge. Retrieved from http://4cmr.org

Please cite this article in press as: Dagoumas, A. Modelling socio-economic and energy aspects of urban systems. Sustainable Cities and Society (2013), http://dx.doi.org/10.1016/j.scs.2013.11.003

646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712