85 aiutsa, B., L. Puangchit, R. Kjelgren, and W. Arunpraparut. 2008. “Urban Green Space, Street Tree and Heritage Large Tree Assessment in Bangkok, ailand.” Urban Forestry & Urban Greening 7 (3): 219–29.UNEP/GRID (United Nations Environment Programme/Global Resource Information Database). 2005. “National Carbon Dioxide (CO2) Emissions per Capita.” http://maps.grida.no/go/graphic/national_carbon_dioxide_co2_emissions_per_capita.UNEP (United Nations Environment Programme) and UN-HABITAT (United Nations Human Settlements Programme). 2005. “Climate Change: e Role of Cities.” http://www.unhabitat.org/downloads/docs/2226_alt.pdf.UNESCAP (United Nations Economic and Social Commission for Asia and the Paci c). 2009. “Key Statistics of Population and Households of Bangkok.” http://www.unescap.org/esid/psis/population/database/thailanddata/central/bangkok.htm.UNFCCC (United Nations Framework Convention on Climate Change). 2003. “Italy: Greenhouse Gas Emissions per Capita.” Data for 1990–2003 submitted to the United Nations Framework Convention on Climate Change, UNFCCC greenhouse gas inventory. http://globalis.gvu.unu.edu/indicator_detail.cfm?IndicatorID=199&Country=IT.Weatherbase [database]. http://www.weatherbase.com.World Bank. 2007. “Strategic Urban Transport Policy Directions for Bangkok.” Urban Transport Development Partnership report, June. http://siteresources.worldbank.org/INTTHAILAND/Resources/333200–1177475763598/2007june_bkk-urban-transport-directions.pdf.Yusuf, S., and K. Nabeshima. 2006. Postindustrial East Asian Cities: Innovations for Growth. Washington, DC, and Stanford, CA: World Bank and Stanford University Press.
87GHG Emissions, Urban Mobility, and Morphology: A HypothesisAlain Bertaud, Benoit Lefèvre, and Belinda YuenIntroduction is chapter explores the link between greenhouse gas (GHG) emissions, transport mode, and city shape. Urban productivity is dependent on people’s mobility within a metropolitan area. GHG emissions, however, are only weakly linked to the number of kilometers traveled per person because of large varia-tions between the emissions per passenger kilometer of di erent transport modes and di erences in the carbon content of the various energy sources used for transport. us, to reduce urban GHG emissions due to transport, it is important to look at all the parameters that contribute to emissions. In this chapter, three concurrent strategies that could contribute to reducing GHG emissions due to urban transport are reviewed: technological improvements within mode, mode shi , and land-use strategy allowing spatial concentration of jobs. In particular, the chapter explores options for improving travel in urban areas by investigating the links between GHG emissions and transport modes, with consideration of associated travel costs and city shape. However, it is our contention that none of these strategies are likely to succeed if not supported by an energy pricing policy directly linking energy price to carbon content. e central hypothesis is that carbon-based energy pricing could trigger a demand shi toward transit in dominantly monocentric cities, providing adequate zoning changes were made. More speci cally, this chapter seeks to develop and determine the following:4
CITIES AND CLIMATE CHANGE Hypothesis 1: Price signals, including energy prices and carbon market–based incentives, road tolls, and transit fares, are the main drivers of techno-logical change, transport modal shi , and land-use regulatory changes. Hypothesis 2: Price signals could shi transport mode from individual cars to public transit for trips from the periphery to the central business district (CBD) only in cities that are densely populated (more than 50 people/hectare(ha) in built-up areas) and already dominantly monocentric.GHG Emissions and Urban TransportUrban GHG emissions per person in large cities are a fraction of the national average ( gure 4.1). is di erence appears as a paradox because cities have a higher gross domestic product (GDP) per person than the national average, and it is usually assumed that higher GDP means higher GHG emissions. In fact, modern cities with a large proportion of service jobs consume less energy per capita than smaller towns and rural areas. However, because GHGs are emitted in urban areas by a very large number of small sources—cars, appli-Figure 4.1 CO2 Emissions in Cities Compared with CountriesSource: EIU 2008; World Resources Institute 2009.Note: CO2e = carbon dioxide equivalent.United StatesUnited KingdomNew York City London Tokyo Stockholm RomeJapanSwedenItalyCO2e emissions per capita 25,00020,00015,00010,000Kilograms CO2e per capita per year5,000
GHG EMISSIONS, URBAN MOBILITY, AND MORPHOLOGY
89ances, individual buildings—as opposed to concentrated sources such as power plants or factories, it is di cult to develop an emission reduction strategy that would work for all emitters.Reliable data on emissions in cities are di cult to collect because of ambi-guity in determining which sources to include as urban. Should urban GHG emissions be limited to sources located within metropolitan boundaries? Or should emissions be counted on the basis of urban residents’ consumption in urban areas? e data for cities shown in gure 4.1 correspond to the rst de -nition, although emissions from electricity are accounted for on the basis of consumption and not on emissions at the location of the power plant.Some analyses solve the problem posed by emission location versus location of consumption by including life-cycle emissions (Button 1993; McKinsey and Co. 2007; Schipper, Unander, and Marie-Lilliu 1999). For instance, the emissions of a car are not limited to the fuel consumed but include also the energy used to manufacture it, to maintain it, and to scrap it a er its useful life. Although this type of de nition is reasonable, the resulting numbers are di cult to calculate, and the method implies a number of assumptions, in particular, concerning the number of years and the number of kilometers traveled during the useful life of a vehicle. It is important to be aware of the limitations of the data set avail-able when comparing cities’ performance in GHG emissions. Some apparent inconsistencies in the data presented below can be attributed to slightly di erent assumptions in the data collected about emissions attributions. e sample of ve large cities1 in high-income countries shown in gure 4.1 gives a range of emissions from 4 to 7 tons per person per year in 2005 (EIU 2008). It is likely that GHG emissions in cities in low- and middle- income countries, for which no reliable data are available, are even higher than the Organisation for Economic Co-operation and Development (OECD) cities shown in gure 4.1. e use of older cars and buses, and the prevalence of two-stroke engines for motorcycles and three-wheelers, might contribute to higher GHG emissions per capita. e three main sources of GHG emissions in cities are buildings, transport, and industries. In the sample of ve high-income cities included in gure 4.1, the proportion of GHG emissions due to transport var-ies from 25 percent of total emissions in New York City to 38 percent in Rome ( gure 4.2). is chapter will be limited to identifying the best strategies to reduce GHG emissions due to transport in a context of increasing urban productivity. e con-clusions of this study would be particularly relevant to cities that have more than 1million inhabitants. According to United Nations data and projections, cities with populations above 1 million accounted for about 1.2 billion, or 18percent of the world population, in 2005. By 2025, it is expected that this will increase to1.85billion and will then represent 23 percent of the world population.
CITIES AND CLIMATE CHANGETransport is a key driver of the economy and is highly dependent (98 percent) on fossil oil. Although already a signi cant sector of GHG emissions, it is also the fastest growing sector globally. Between 1990 and 2003, emissions from the trans-port sector grew 1,412 million metric tons (31 percent) worldwide. e sector’s share of carbon dioxide (CO2) emissions is also increasing. In 2005, the transport sector contributed 23 percent of CO2 emissions from fossil fuel combustion. It is also the sector where the least progress has been made in addressing cost-e ective GHG reductions (Sperling and Cannon 2006). As mentioned earlier, the frag-mentation of emissions sources and the complexity of demand and supply issues in urban transport explain the lack of progress. Making transport activity more sustainable must be a top priority policy if climate change is to be addressed.In most cities, numerous urban problems are transport related, such as con-gestion on urban roads, poor air quality, fragmented labor markets, and social fractioning due to poor access to economic and social activity and the like (Ng and Schipper 2005; World Bank 2009). Road transport accounts for, by far, the largest proportion of CO2 emissions from the transport sector, principally from automobile transport. Against the projected increase in car ownership world-wide (expected to triple between 2000 and 2050), road transport will continue to account for a signi cant share of CO2 emissions in the coming decades. Within cities, modal share and measures facilitating less GHG-intensive modes Figure 4.2 CO2 Emissions in Five High-Income CitiesSource: EIU 2008.1009080706050403020100New York City London Tokyo Stockholm RomeIndustryBuildingsTransportPercent emission of CO2 per person per year (2005) per sector 25%28%31% 31%38%
GHG EMISSIONS, URBAN MOBILITY, AND MORPHOLOGY
91such as public transport require closer examination. Modal shi policies are generally inadequately assessed in CO2 policy (OECD 2007). Because GHG emissions caused by urban transport have to be reduced while urban produc-tivity has to increase, it is important to establish the links between urban trans-port, labor mobility, and city productivity.Mobility and Cities’ EconomiesEconomic literature, both theoretical and empirical, linking the wealth of cities to spatial concentration is quite abundant and no longer controversial in academic circles (Annez and Buckley 2009; Brueckner 2001; Brueckner, isse, and Zenou 1999). e World Bank’s World Development Report (2009), “Reshaping Economic Geography,” and the Commission on Growth and Development report “Urbanization and Growth” (Annez and Buckley 2009) exhaustively summarize and document the theoretical and empirical arguments justifying the economic advantage provided by the spatial concen-tration of economic activities in large cities. e necessity to manage urban growth rather than to try to slow it down is eventually reaching mayors, city managers, and urban planners. e size of cities is not critical; what matters is the connectivity insured by urban transport networks2 between workers and rms and between providers of goods and services and consumers, whether these consumers are other rms or individuals. is connectivity is di cult to achieve in large cities. It requires coordination between land uses and invest-ments in transport networks; di cult pricing decisions for road use, parking, and transit fares; and nally, local taxes and user fees that makes the main-tenance and development of the transport network nancially sustainable (Staley and Moore 2008).Tra c congestion in slowing down mobility represents a management fail-ure on the part of city managers. Congestion has a double negative e ect: It acts as a tax on productivity by tying down people and goods, and it o en increases GHG emissions even for vehicles that would otherwise be performing satisfac-torily. It is conceivable that mismanaged large cities may reach a level of con-gestion that negatively o sets the economic advantage of spatial concentration. In this case, these cities would stop growing. However, the positive economic e ect of agglomeration must be very powerful to o set the chronic congestion of cities such as Bangkok and Jakarta that are still the economic engine of their region in spite of their chronic congestion.Poor migrants moving to large cities o en have di culties in participating in the urban economy, either because their housing is located too far from the urban transport networks or because they cannot a ord the cost of transit or motorized transport. It has been observed that some slums appear to be self-
CITIES AND CLIMATE CHANGEsu cient and that many slum dwellers are able just to walk to work. Some have argued that slum dwellers’ lack of motorized mobility and inclination toward walking would constitute an advantage in terms of GHG emissions and should be emulated by higher-income groups. is argument is a cruel joke on the poor because their lack of mobility condemns them to live in large cities with all its costs but none of its bene ts. e lack of mobility in many slums and in some badly located government housing projects constitutes a poverty trap rather than an advantage to be emulated in the future (Gauteng in South Africa being a case in point).Although walking and cycling do constitute an indispensable transport mode in large cities, people using these modes should do it by choice, not because they are forced to do so by lack of access or a ordability of other means of transport. Because mobility is a necessity for economic survival in large cit-ies, a reduction of GHGs should not be made by reducing mobility and cer-tainly not by preventing an increase in mobility for the poor. e reduction of the number of passenger kilometers traveled (PKmT) should not be targeted for reduction to reduce GHG emissions. To the contrary, because of the lack of mobility of a large number of poor people living in large cities, PKmT should increase in the future. Various alternative solutions to decrease GHG emissions while increasing PKmT are discussed.Identifying Key Parameters in Urban Transport GHG Emission SourcesGHG emissions from transport are produced by trips that can be divided into three broad categories:1. Commuting trips2. Noncommuting trips3. FreightCommuting trips are the trips taken to go from residence to work and back. In most low-income cities, commuting trips constitute the majority of trips using a motorized vehicle (with exceptions in some East Asian cities where nonmo-torized trips still constitute a large number of commuting trips). Noncommut-ing trips are trips whose purpose is other than going to work, for instance, trips to schools, to shops, or to visit family or other personal reasons.In high-income countries, commuting trips constitute only a fraction of total trips. For instance, in the United States, commuting trips represented 40 percent of all motorized trips in 1956; in 2005, they represented only slightly less than 20 percent of all motorized trips (Pisarski 2006). In low-income cities, most of
GHG EMISSIONS, URBAN MOBILITY, AND MORPHOLOGY
93these trips involve short distances and are using nonmotorized transport. When noncommuting trips become more numerous and longer they tend to be made via individual cars or motorcycles because destinations are not spatially concentrated and transit networks cannot easily accommodate them. For instance, in New York City in 2005, transit was used for 30.8 percent of all commuting trips but only for 9.6 percent of all commuting and noncommuting trips (O’Toole 2008).Freight trips, including public vehicle travel and urban goods and services travel, constitute a sizable portion of all trips but vary signi cantly between cities. Because freight trips within urban areas are always done by individual vehicle and cannot use transit, these trips are adversely a ected by road conges-tion, which results in signi cant costs to the economy of cities.Will the trends observed in the United States anticipate what will happen in other parts of the world when these cities reach a level of income comparable to that of the United States today? is appears unlikely because of di erences in city density between the United States and other parts of the world. Most cities outside the United States have a density far higher than U.S. cities, o en by two orders of magnitude. Although densities of large cities tend to decrease over time, the decrease is slow and is unlikely to ever reach the low density of U.S. cities. It is probable that in high-density cities noncommuting trips will largely use nonmotorized transportation, taxis, or transit, as is the case in high-density Manhattan today.Analysis in this research will therefore concentrate on emissions from commuting trips because these trips are the most common type in low- and middle-income cities. In addition, commuting trips require the most capi-tal investment because of the transport capacity required during peak hours. Commuting trips o en de ne a transport network whereas the other types of trips, including freight, piggy-back onto the transport investments made ini-tially for commuting trips.In East Asia, commuting trips, using walking or bicycles, constituted the majority of commuting trips in the 1980s and 1990s. During the past 20years, because of the physical expansion of cities and increase in oor space con-sumption due to rising incomes, the share of nonmotorized transport has unfortunately been shrinking. In 2006, for instance, the share of nonmotorized commuting trips has been reduced to about 20 percent in Shanghai from about 75 percent in the early 1980s.Disaggregating Commuting Trips by ModeCommuting trips can be disaggregated into three modes: nonmotorized mode (walking and cycling, and increasingly included in this category, people work-ing at home and telecommuting); motorized self-operated vehicles (SOVs),
CITIES AND CLIMATE CHANGEincluding motorcycles and private cars (car pools included); and transit mode (minibuses, buses, bus rapid transit [BRT], light rail, subways, and suburban rail). e types of vehicles used in the last two modes vary enormously in emission performance. In addition, within each mode—SOV and transit—each city has a eet of vehicles, which have a wide range of GHG emissions performance. Comparisons between vehicles o en di er by orders of magni-tude depending on technology, maintenance, age of vehicle, energy source, and load (the average number of passengers per vehicle). To see more clearly the impact of di erent transport strategies on the reduction of GHG emis-sions, we have built a simple model linking the various vehicle eet parameters to GHG emissions per commuter. e model is limited to analyzing CO2 emis-sions from commuting trips, which are still the most common motorized trips in low- and middle-income cities. For each mode, the inputs of the model are the following:1. e percentage of commuters using the mode2. e average commuting distance (in kilometers)3. e CO2 equivalent (CO2e) emission per vehicle kilometer traveled (VKmT), calculated for full life cycle when data available4. e load factor per type of vehicleNumerous publications provide GHG emissions expressed in grams of CO2 per PKmT (table 4.1). However, the data assume a passenger load to calculate the CO2 per PKmT. Because the load is a crucial parameter in the model, it has been necessary to calculate the CO2 emissions per VKmT. However, fuel consumption may vary for the same vehicle, depending on the load; there-fore, load and fuel consumption are not completely independent variables. We have therefore slightly adjusted the energy consumption values by VKmT to re ect this. A more sophisticated model would establish more accurately the relationship between load and fuel consumption for each type of vehicle. For demonstration purposes of the proposed methodology, results were found to be robust enough to allow this simpli cation. e equation used in the model showing the daily GHG emissions as a function of the number of passengers using di erent modes, with di erent average commuting distances, load factor, and engine fuel performance, is presented in the annex.Based on the equation given in the annex, it can be shown that trying to reduce the average commuting distance per day (variable D)—de facto reducing labor mobility—would not provide much e ect on Q (GHG emissions per day) compared with a change in vehicle eet performance (variable E), a mode shi (variable P), or an increase in the load factor (variable L). As seen in table 4.1, the possible values taken by E vary by a factor of four between a hybrid diesel
GHG EMISSIONS, URBAN MOBILITY, AND MORPHOLOGY
95and an SUV, and by a factor of two between the New York City subway and a Toyota Prius! By contrast, land-use changes might, at best, reduce average com-muting distance D by 5 to 10 percent within a minimum period of 20 years. is model, which could be used as a rough policy tool, was tested for parameters for New York City and Mexico City. e inputs and outputs of the model using New York City parameters in 2000 are shown in table 4.2. e model shows the di erence of performance in terms of GHG emissions between transit and cars in New York City: Emissions per car passenger per year are nearly six times more than the emissions per transit passengers. e model allows testing of the impact of alternative strategies; for instance, what would be the impact of an increase of hybrid cars over the total number of cars, everything else staying constant? Or what would be the impact of an increase in transit passengers, or in the load factor of buses, and so on? Table 4.3 shows the impact of two alternatives in reducing GHG emissions.Table 4.3 demonstrates the potential impact in New York City of a change in the composition of the car eet and, alternatively, a mode shi from cars to transit. e changes concern only the value of variable P in the model’s equation. e current situation in 2005 is shown in column A. In column B, an increase from 0.5 to 19 percent in the number of commuters using hybrid cars, repre-senting about one out of ve cars used by commuters, bring a 28 percent reduc-tion in GHG emissions. In column C, a mode shi from car to transit, raising the share of transit from 36 percent of commuters to 46 percent, decreases GHG TABLE 4.1 GHG Emissions for Various Vehicles with Various Passenger Load AssumptionsVehicle typeGrams of CO2 per passenger mileGrams of CO2 per passenger kilometerSUV 416 258Average U.S. car 366 227Motor buses 221 137Light rail 179 111Commuter rail 149 93Hybrid gas 147 91Toyota Prius 118 73Hybrid diesel 101 63Metro 94 58New York MTA 73 45New York subway 58 36Source: Demographia 2005; EIU 2008; O’Toole 2008.Note: MTA = Metropolitan Transportation Authority.
CITIES AND CLIMATE CHANGETABLE 4.2 Input and Output of GHG Emissions for New York CityAverage distance per passenger per commuting tripNumber of commuters per modePercent of commuters per modeGrams of CO2e per VKmTLoad factor of vehicleLoad factor as % of total vehicle capacityGrams of CO2e per PKmTTotal tons of CO2 emitted by commuters per dayUnitsKm people % gr CO2e people % gr CO2e T CO2eMode / SymbolD Pn P E L L/Ca % QiWalk 2.5 470,000 5 — 1 100 — —Cycle 5.0 94,000 1 — 1 50 — —Car (gasoline) 19.0 4,324,000 46.0 375 1.63 33 230 37,802Car (diesel) 19.0 47,000 0.5 256 1.63 33 157 281Car (hybrid) 19.0 47,000 0.5 105 1.63 33 64 115Car (electric) 19.0 — 0 163 1.63 33 100 —Motorcycle 2-stroke 8.0 94,000 1 119 1.1 55 108 163Minibus gasoline 20.0 — 0 720 7 58 103 —Minibus diesel 20.0 — 0 600 7 58 86 —Bus diesel 20.0 564,000 6 1,000 30 50 33 752Bus natural gas 20.0 1,128,000 12 1,200 30 50 40 1,805Rail transit 20.0 2,632,000 28 3,950 110 73 36 3,7819,400,000 100 Tons per day Q = 44,698 Total transit 46
GHG EMISSIONS, URBAN MOBILITY, AND MORPHOLOGY
97Number of people in New York City MSA14,687,500 Kg/per year per commuter 1,240E/P ratio (%) 64 Kg/year by transit passenger 38Number of commuting days per year261 Kg/year by Car passenger 2,217Source: Authors’ analysis.Note: Total number of commuters (T) = 9,400,000. Figures in italics are input of the model, other ﬁ gures are output. MSA = metropolitan statistical area; PKmT = passenger kilome-ter traveled; VKmT = vehicle kilometer traveled.
CITIES AND CLIMATE CHANGEemissions by 13 percent. Further reductions could be achieved by introducing hybrid buses or increasing loads of both cars and transit. e use of the model allows a back-of-the-envelope calculation of the impact of potential changes in technology and transport mode on GHG emissions. e model does not have anything to say about the feasibility or the probability of such a change to occur. Although the rough calculations shown imply that the combined impacts of technology change and mode shi could be large, how to achieve these changes remains the main problem to be solved. Most of the vehicle technology, such as hybrid engines, that reduces fuel consumption has been around for at least 10 years. Rail transit using electricity has been common in large cities for more than 100 years. e fact that in many cities the use of transit represents a minority mode raises important questions about Table 4.3 Potential Impact of Vehicle Shift and Mode Shift on GHG Emissions in New York City Metropolitan Area(A) (B) ©iMode / Symbol P PChange in CO2e emissions PChange in CO2e emissions 1 Walk 5% 5% 5% 2 Cycle 1% 1% 1% 3 Car (gasoline) 56% 37.5% −33% 46% −18% 4 Car (diesel) 0.5% 0.5% 0.5% 5 Car (hybrid) 0.5% 19.0% n.a. 0.5% 6 Car (electric) 0% 0% 0% 7 Motorcycle 2-stroke 1% 1% 1% 8 Minibus gasoline 0% 0% 0% 9 Minibus diesel 0% 0% 0%10 Bus diesel 5% 5% 6% 20%11 Bus natural gas 10% 10% 12% 20%12 Rail transit 21% 21% 28% 33%Tons per day 51,545 36,918 44,698 −13%Kilograms per year per commuter 1,418 1,024 −28% 1,240 −13%Source: Authors’ analysis.Note: n.a. = not available.
GHG EMISSIONS, URBAN MOBILITY, AND MORPHOLOGY
99consumer preferences for urban transport. e transport mode split for New York City in 2005 shown in table 4.2 represents a state of equilibrium. It is important to know what factors could change this equilibrium to a new state that would be more favorable for GHG reductions.Consumers’ Demand for Transport e loss of transit share over the past few decades in most of the world’s major cities has to be acknowledged. Even in Singapore transit mode share declined from 55 percent of commuters in 1990 to 52.4 percent in 20003 (Singapore Department of Statistics 2000). is decrease is striking because Singapore has had the most consistent transport policy over two decades favoring tran-sit, including strict limits on car ownership, and has been a world pioneer for congestion pricing using advanced technology. In addition, Singapore has always had excellent coordination between land use and transport investments. Although the preceding section has shown that there is an overwhelming case for increasing transit mode to reduce GHG emissions, consumer choice seems to follow the opposite trends. It is therefore important to understand why tran-sit is losing ground in so many cities and what alternative strategies exist and in which type of cities the trend could possibly be reversed.Consumers’ decisions to use one mode of transport over others depend on three main factors:1. Cost2. Speed3. Convenience, as determined by frequency and reliability of service and comfortFor low-income commuters, the cost of transport is the major consideration. For very low-income commuters, walking is o en the only a ordable option, which signi cantly lowers their ability to take advantage of the large labor mar-ket o ered by large cities. In Mumbai, for instance, about 4 million people walk to work every day (about 45 percent of the active population). Middle- and low-income users above extreme poverty are the prime customers for transit, as buying and maintaining a car is beyond the means of most of these, although subsidized fares frequently exist to make transit more a ordable. However, in numerous middle- and high-income countries, some cities retain a signi cant number of transit users who are middle or high income—for instance, Hong Kong, London, New York City, Paris, and Singapore, among others. How these cities have managed to maintain a high use of transit among a uent house-holds will be described in the next section.
CITIES AND CLIMATE CHANGEIn an increasing number of cities in low- or middle-income countries, the dispersion of employment makes it inconvenient to use transit, because no transit route goes directly to their location of employment. For those commuters who cannot a ord to use individual cars or motorcycles, the most convenient options are collective taxis or minibuses. Commuting by microbuses at the expense of transit has become the dominant transport mode in Gauteng, Mexico City, and Tehran, for instance. As households’ income increases, the speed of transport and convenience become more important factors than cost, or rather, higher-income commuters give a higher value to the time spent commuting than do lower-income ones. Speed of transport is limited in most transit system by fre-quent stops and the time required for transfers. In city structures where a car is a feasible alternative mode of transport, commuters who can a ord the cost would normally switch to individual cars. e exhaustive study conducted by Pisarski (2006) on commuting char-acteristics in U.S. cities gives an order of magnitude of the speed di erence between transit and individual cars in those cities ( gure 4.3). e average commuting distance is about the same between the di erent modes except for walking, cycling, and rail transport. One can see that in spite of the congestion prevalent in most U.S. cities, commuting time by transit requires about double the time required by individual cars. Travel time for car pooling when involving Figure 4.3 Average Travel Time in U.S. Cities by Transport ModeSource: Pisarski 2006.WalkBikeTaxiMotorcycleDrive aloneCar 2 peopleCar pool 3 peopleCar pool 4 peopleMinibus 5 or 6 peopleStreetcarBusMinibus 7+ peopleSubwayFerryRailroad1219202224273134394446474866710102030Minutes40 50 60 70 80
GHG EMISSIONS, URBAN MOBILITY, AND MORPHOLOGY
101more than four people becomes similar to transit. is explains in great part the loss of transit share in U.S. cities in the past two decades.In Singapore, with one of the most e cient transit systems in the world, the ratio of transit travel time to car driving time is lower than in U.S. cities. However, the di erence in travel time is signi cant enough (see table 4.4) to indicate that transit would not be a rst-choice transport mode for people who can a ord an alternative. e high speed of car commuting is, of course, part of the success of Singapore’s transport strategy. Congestion pricing, constantly adjusted to facilitate uid tra c, ensures high speed for all car commuters who can a ord the high premium paid for car ownership and for congestion tolls. e challenge is to propose urban transport strategies that would result in reducing GHG emissions while maintaining mobility as re ected by commut-ers’ mode preference. ese di erent strategies would have to be adapted to dif-ferent spatial forms of urban growth—monocentric, polycentric, high and low densities—and to a context of increasing urban income and a decreasing cost of car acquisition. ese strategies will have to rely on the three tools available to urban managers: pricing, regulations, and land-use policy.Energy Pricing, GHG Emissions, and Market-Based IncentivesAs discussed earlier, a signi cant reduction in GHG in urban transport could be achieved in two ways: technological change to reduce carbon content per VKmT and transport-mode shi from private car to transit. As alluded to ear-lier, the pricing of energy based on its carbon content is an indispensable policy instrument to trigger these changes to reduce GHGs in the long run. e pric-ing of energy based on carbon content could be achieved through a carbon tax or through “cap and trade.” e merit of each approach is discussed next.TABLE 4.4 Singapore: Travel Time by Transport ModeModeMedian travel time (minutes)Distance(kilometers)Speed (kilometers/hour)Car 27 29.2 65Metro 41 11.5 17Metro + bus 51Bus alone 38 Source: Singapore Department of Statistics 2000.
CITIES AND CLIMATE CHANGEIn each city the current use of low-carbon technology and the ratio between transit and car commuting is re ecting an equilibrium state between supply and demand. Any change in technology or transport-mode share will require a move to a new state of equilibrium in the economy of transport. Signi cantly higher gasoline prices, as experienced in 2008, temporarily modi ed this state of equilibrium. Demand for transit increased and VKmT decreased. However, as long as renewable energy sources were not available at a competitive price, the high price of oil made it cheaper to generate electricity from coal or shale oil. Electricity is used mostly as a source of energy for rail transit, but electri-cal cars that would recharge their batteries from the electricity grid will use it increasingly. Electricity produced by coal-burning power plants generates twice as much GHG per kilojoule than power plants using natural gas. Without a system of pricing energy based on its carbon content, higher oil and natural gas prices could increase GHG emissions rather than reducing them by shi ing electricity generation to coal-fueled power plants.However, carbon pricing cannot be decided at the local level and is depen-dent on national policy and increasingly on international agreements. It must be acknowledged that these policy instruments will have a limited impact in the absence of carbon pricing.Various policy instruments are currently available to reduce GHG emis-sions due to urban transport. eir e ectiveness is o en limited by the qual-ity of national and local governance, as well as a city’s income distribution and spatial structure. Policy instruments can be divided among three princi-pal categories:1. Regulatory instruments, such as limitations on the number of vehicles on the road on a given day (for example, Beijing, Bogota, and Mexico City pico y placa (peak and [license] plate) and limitation on the number of cars reg-istered in the city (for example, Singapore car quota system)2. Pricing instruments modifying relative prices between private car and transit modes, such as road pricing: xed tolls and congestion pricing (for example, London, Singapore, and Stockholm); a fuel tax, which needs to be compared with an increase in the price of a barrel of oil due to oil market evolution (for example, Bogota, Singapore, Chicago, and most other U.S. cities); transit fare subsidies (for example, Los Angeles and San Francisco); and pricing and taxing of parking (for example, Edinburgh, New York City, Peterborough, and She eld) 3. Investment in transport infrastructure in order to increase and improve the supply of transit modes (for example, Bogota, Jakarta, and Singapore)
GHG EMISSIONS, URBAN MOBILITY, AND MORPHOLOGY
103Regulatory InstrumentsRegulatory instruments aiming at mode shi from car to transit are generally not e ective because the choice of a transport mode must be demand driven. Regulatory instruments aiming to limit or reduce car ownership and car usage could seriously limit mobility in the absence of adequate investments in tran-sit to replace the decrease in car trips. e example of Singapore in xing a quota for car growth is rather unique. It could have been very disruptive to the economy if the government had not simultaneously been able to nance and develop a very e ective transit system consistent with its land-use policy. is important aspect will be developed later.In countries with high economic inequality (such as Colombia or Mexico), policies such as pico y placa4 create an incentive for higher-income households to buy a second car. is second car is o en a secondhand car with worse engine performance than most recent models. As a result, the pico y placa policy has o en resulted in worse pollution and higher GHG emissions than the status quo ante. e availability of a new type of low-cost car—the Tata Nano, for example—could make this policy even more ine ective.Pricing InstrumentsPricing instruments are normally aimed at pricing transport at its real eco-nomic price (Button 1993; Goodwin, Dargay, and Hanly 2004). When this can be achieved, it removes the distortions that hidden subsidies introduce in resource allocation. Congestion pricing and parking pricing, for instance, aim at adjusting the price of using a highway or of a parking space to re ect its real economic value, including externalities due to congestion (Luk 1999). e aim of economic pricing is not to be punitive but to seek a more e cient allocation of resources. Pricing instruments also include subsidies, which have a di erent aim than economic pricing. Subsidies aim at being redistributive. For instance, most transit fares are heavily subsidized.5 Transit-fare subsidies are aiming at increasing the mobility of low-income households, allowing them to fully par-ticipate in a uni ed metropolitan labor market.It is tempting also to use transit-fare subsidies as a nancial incentive to convince car commuters to switch to transit. is is not a very e ective way to increase transit-mode share in the long run. e subvention for transit opera-tion and maintenance o en comes from local government budget allocation. e larger the number of users, the larger the subsidies required. is works as a reverse incentive for the transit operator to improve services. In the long run, the subsidies paid by the government to the transit authority usually fall short of the real cost of operation and maintenance, resulting in a deterioration
CITIES AND CLIMATE CHANGEof service. An example of this problem came to light during the latest nancial crisis in the United States. Local governments, because of increasing de cits, were obliged to scale down transit services, including frequency, right at the moment when the high price of fuel and declining households’ income were forcing some commuters to switch from car to transit commuting.Transit-fare subsidies, when they exist, should be targeted to low-income households or to the unemployed. Transit-fare subsidies directed to the a u-ent are in fact a transfer payment made by government to commuters for not polluting instead of charging car commuters for the externalities they cause.Pricing instruments re ecting real economic costs have a value in them-selves because they contribute to better allocation of resources. However, they do not necessarily change consumer behavior. For instance, a toll charge on a highway may not reduce congestion if it is set too low. Congestion pricing, as practiced in Singapore, involves increasing tolls until the desired decrease in congestion is achieved. Congestion pricing consists of increasing or decreasing prices until equilibrium between supply and demand is reached. Congestion pricing does not aim at recovering the cost of a highway, but at limiting tra c volume to obtain a desired speed.Pricing parking at the market price is equivalent to congestion pricing: e operator will increase the price of parking until all the parking spaces are lled. In New York City, the municipality taxes a private parking space at 18 percent of the daily rate paid (in addition to the property tax and business tax). In this way, the municipality recovers a share of the private market rate without having to set a municipal parking rate. e transaction cost of recovering the rate from consumers and adjusting it to the market price is paid by the private operator. Taxing privately operated parking garages might be a more e ective way of recovering an area-wide congestion fee than the way it is currently recovered in London.Congestion pricing is not always possible. It requires technology investment that may be expensive to install and operate, and the high transaction cost may greatly reduce the income of the operator. In some cases, congestion pricing is not politically acceptable. For instance, it would be di cult to increase or decrease the transit fare every hour depending on the number of commuters boarding at any given time.In the case in which congestion pricing is not feasible, the e ectiveness of increasing or decreasing prices (that is, changing prices to increase or decrease demand) depends on the price elasticity of demand. e price elasticity of demand depends on numerous factors and can be measured from empirical experience, but it cannot be calculated in advance without empirical data. Vari-ous factors a ect how much a change in prices impacts travel demand for a given travel mode: type of price change, type of trip, type of traveler, quantity
GHG EMISSIONS, URBAN MOBILITY, AND MORPHOLOGY
105and price of alternative options, and time period (short term [one year] and long term [5–10 years]).Nearly all studies assume that the e ects of a reduction are equal and oppo-site to the e ects of an increase or, in other words, that elasticity is “symmet-rical” (Goodwin, Dargay, and Hanly 2004). Empirical evidence suggests that this assumption might not be true. However, because of the number of factors a ecting elasticity, it is o en di cult to extrapolate with certainty results from one city to another in the absence of an empirical local database. With this caveat, available data from the literature on the price elasticity of demand in urban transport are reviewed. e current literature on price elasticity in trans-ports could be summarized as follows:• Long-run elasticities are greater than short run ones, mostly by factors of 2to 3 (Goodwin, Dargay, and Hanly 2004).• Fuel consumption elasticities to fuel price are greater than tra c elasticities, mostly by factors of 1.5 to 2.0 (Goodwin, Dargay, and Hanly 2004).• Motorists appear to be particularly sensitive to parking prices. Compared with other out-of-pocket expenses, parking fees are found to have a greater e ect on vehicle trips, typically by a factor of 1.5 to 2.0 (Gordon, Lee, and Richardson 2004): A $1 per trip parking charge is likely to cause the same reduction in vehicle travel as a fuel price increase that averages $1.50 to $2.00 per trip.• Shopping and leisure trips elasticities are greater than commuting trip elas-ticities. Although we can reduce or avoid travel or the need to travel for shopping, we are more likely to continue traveling to commute.• Road pricing and tolls e ects depend on the pricing mechanism design. Luk (1999) estimates that toll elasticities in Singapore are −0.19 to −0.58, with an average of −0.34. Singapore may be unique; the high cost of car ownership constitutes a very high sunk cost, which may tend to make travel less sensi-tive to price.• Transit price e ects are signi cant: Balcombe and others (2004) calculate that bus fare elasticities average around −0.4 in the short run, −0.56 in the medium run, and 1.0 over the long run, whereas metro rail fare elasticities are −0.3 in the short run and −0.6 in the long run. Bus fare elasticities are lower during peak (−0.24) than o -peak (−0.51).Carbon-Based Investment in Transport InfrastructureCarbon-based investments in transport infrastructure face three main barriers: nancial, institutional, and political. Carbon markets have been positioned as an economically e cient market-based incentive for answering these three barriers.
CITIES AND CLIMATE CHANGEToday, however, their usage for cities, and even more for urban transportation, is limited for several reasons:• Cities’ participation in carbon markets is limited to exibility mechanisms such as o set, voluntary, or Clean Development Mechanism (CDM)/Joint Implementation projects.• ese markets have been rarely used for promoting a more energy- and car-bon-e cient urban transportation pattern: To date, 1,224 CDM projects have been registered by the UN Framework Convention on Climate Change Exec-utive Board, and only two have been transportation projects, representing less than 0.13 percent of total CDM projects (the Bogota BRT TransMilenio and the Delhi subway regenerative breaking system).• Carbon markets favor low-hanging fruit projects, which do not have the greatest potential to reduce GHG emissions: e majority of the CDM transportation projects accepted or proposed claim their emission reduc-tions through switching fuels used. Some entail improvements of vehicle e ciency through a di erent kind of motor or better vehicle utilization. Few projects deal with modal shi , and none involves a reduction of the total transportation activities.Given these barriers, two questions must be addressed. e rst is, How and why are carbon markets biased against projects targeting urban transportation? Sev-eral explanations can be explored:1. CDM and transport projects di er widely in terms of challenges and oppor-tunities. ere is a scale gap between the two realities in which the main leaders of each project evolve: a. (Local) transport projects aim to change the city and make it economically attractive. Challenges include involving all stakeholders in thedecision-making process. b. (International) challenges for CDM projects are technical (convincing CDM executive boards and international experts) and nancial.2. Di use emissions, such as in the transportation sector, are costly to aggre-gate, thus the CDM “act and gain money” incentive has rather limited e ects.3. Classic CDM challenges are particularly vexing for the transport sector: a. De ning project boundaries, because of complex up- and downstream leakages. b. Establishing a reliable baseline, when behavioral parameters are key. c. Implementing a reliable monitoring methodology, because data genera-tion is costly.
GHG EMISSIONS, URBAN MOBILITY, AND MORPHOLOGY
107 e consequences of this bias are that transport and CDM projects are con-ducted in parallel; without interaction, cities outsource CDM projects to inter-national experts and organizations without much involvement; and CDM project-based design is missing the main GHG reduction opportunities. us, within their existing framework, carbon markets can be used as a source of funding signi cant only at the local level to do the following:• Subsidize (and reduce) transit fares.• Finance intermodality infrastructures and thus facilitate modal shi .• Finance well-bounded technology-oriented CDM projects, such as changes in fuels and technology, optimization of the balance between bus supply and demand, tra c-light systems, and more generally, new information technol-ogies for vehicle or system operations. ese well-bounded, technology-oriented CDM projects could be levered by bundling them through the newly existing programmatic CDM. e second question asks: How could the design of carbon markets evolve to be more “urban transportation friendly”? In the perspective of the post-2012 transportation sector, a unanimous call is heard for changes in the carbon mar-kets’ design. Many important opportunities for transportation emission reduc-tions would not easily t into an individual CDM project. Various propositions are under discussion:1. A sectoral policy-based approach crediting new green policy or enforce-ment of standards. A sectoral approach would not reduce methodological di culties. Its advantages would rather be to scale activities up to a level that is equal to the scale of the challenges faced in redirecting transport into a more sustainable direction.2. Cities’ commitment to reduce GHG emissions and a “No Loose Target” approach.3. Registries including National Appropriate Mitigation Actions for cities and the urban transportation sector.4. Integrate Global Environment Fund and O cial Development Assistance in CDM funding, notably to nance transaction costs, to fund capacity-building activities, and to generate data.In brief, a broader and exible approach, based on a bottom-up mechanism, would do the following:• Foster cities to take the lead on GHG emissions reduction strategies ( nan-cial and electoral motivations)• Give cities incentives to act for the short term (low-hanging fruits) as well as for the long term and, thus, change the urban development trajectory
CITIES AND CLIMATE CHANGE• Leave intact their ability to create and implement solutions that are relevant and palatable with local speci cities—for example, to implement land-use policies that increase the oor area ratio (FAR) in CBDs or transport policies that modify the relative prices of di erent transport modesUrban Spatial Structures and Transport ModePrice and speed are not the only determinant of consumers’ choice for trans-port mode; urban spatial structures play a major role in determining the type of transport that is likely to be the most convenient. Urban structures are de ned by the spatial distribution of population densities within a metropolitan area and by the pattern of daily trips. Depending on a city’s spatial structure, com-muters may be able to switch from car to transit, or their choices may be limited between individual cars, minibuses, and collective taxis. In high-density cities, sidewalks and cycle lanes could be designed in such a way as not to discour-age walking and cycling. Although urban structures do evolve with time, their evolution is slow and can seldom be shaped by design. e larger the city, the less it is amenable to change its structure. However, it is important for urban managers to identify the opportunities present in their city and to take full advantage of them to reduce GHG emissions with transport strategies consis-tent with their spatial structures. Identi ed next are the most common types of spatial structures and the transport strategies that would have the most chances of success for each type of spatial structure.Type of Urban Spatial Structures and Choice of Transport ModesUrban economists have studied the spatial distribution of population densities intensively since the pioneering work of Alonso (1964), Mills (1970), and Muth (1969, 1985), which developed the classical monocentric urban density model. Empirical evidence shows that in most cities, whether they are polycentric or monocentric, the spatial distribution of densities follows the classical model predicted by Alonso, Muth, and Mills (Bertaud and Malpezzi 2003). e density pro le of most large cities shows that the traditional monocen-tric city model is still a good predictor of density patterns. It also demonstrates that markets remain the most important force in allocating land, in spite of many distortions to prices due to direct and indirect subsidies and ill-conceived land-use regulations. e pro le of the population densities of 12 cities on four continents ( gure 4.4) shows that in spite of their economic and cultural dif-ferences, markets play an important role in shaping the distribution of popula-tion around their centers. All the cities shown in gure 4.4 follow closely the
GHG EMISSIONS, URBAN MOBILITY, AND MORPHOLOGY
109Figure 4.4 Distribution of Population Densities in 12 CitiesRio de Janeiro Metropolitan Area1 3 5 7 9 111315171921232527290501001502002503003501 3 5 7 9 11131517192123252729050100150200250300350Paris 19901 3 5 7 9 11131517192123252729050100150200250300350Bangalore (1990)0501001502002503003501 3 5 7 9 11131517192123252729Atlanta1357911131517192123252729Buenos Aires50100150200250300350-1 3 5 7 9 11131517192123 25272950100150200250300350-Stockholm0501001502002503003501 3 5 7 9 11131517192123252729Jakarta (Jabotabek)1 3 5 7 9 11131517192123 252729050100150200250300350Los Angeles1 3 5 7 9 11131517192123 25272950100150200250300350-Mexico City1 3 5 7 9 11131517192123 25272950100150200250300350-Barcelona0501001502002503003501 3 5 7 9 11131517192123 252729Bangkok 19880501001502002503003501 3 5 7 9 11131517192123 252729New York metropolitan areaDistance from city center (kilometers)people/hectareSource: Bertaud 2006.
CITIES AND CLIMATE CHANGEnegative sloped gradient predicted by the classical monocentric urban model, although several cities in the samples are de nitely polycentric (Atlanta, Mexico City, Portland, and Rio de Janeiro). e density pro le indicates that some parts of metropolitan areas are incompatible with transit. In areas where residential densities fall below 50 people per hectare, the operation of transit is ine ective.Land use and the transport network determine the pattern of daily trips taken by workers to commute to work. As income increases, noncommuting trips—trips to shopping centers, to take children to school, to visit relatives, or to take leisure trips—become more important. e proportion of commuting trips in relation to other types of trips is constantly decreasing.Figure 4.5 illustrates in a schematic manner the most usual trip patterns in metropolitan areas. In monocentric cities ( gure 4.5A) where most jobs and amenities are concentrated in the CBD, transit is the most convenient transport mode because most commuters travel from the suburbs to the CBD. e origin of trips might be dispersed, but the CBD is the most common trip destination. Small collector buses can bring commuters to the radials, where BRT or an underground metro can bring them at high speed to the CBD. Monocentric cities are usually dense (density more than 100 people per hectare).In polycentric cities ( gure 4.5B), few jobs and amenities are located in the center, and most trips are from suburbs to suburbs. Although a very large Figure 4.5 Urban Trip Patterns in Monocentric and Polycentric CitiesSource: Bertaud 2006. AThe most common urban spatial structuresBCDlowhighDensitiesThe Classical Monocentric Model- strong high-density center with high concentration of jobs and amenities- radial movements of people fromperiphery toward centerThe “Urban Village” Model- people live next to their place of employment- people can walk or bicycle to work- this model exists only in the mind of planners, it is never encountered in real lifeThe Polycentric Model- no dominant center, some subcenters- jobs and amenities distributed in a near uniform manner across the built-up area- random movement of people across the urban areaThe Composite Model- a dominant center, some subcenters- simulateneous radial and random movement of people across the urban area
GHG EMISSIONS, URBAN MOBILITY, AND MORPHOLOGY
111number of travel routes are possible, most will have few passengers per route. e trips have dispersed origins and dispersed destinations. In this type of city structure, individual means of transportation or collective taxis are more con-venient for users. Mass transit is di cult and expensive to operate because of the multiplicity of destinations and the few passengers per route. Polycentric cities usually have low densities because the use of individual cars does not allow or require much concentration in any speci c location.Figure 4.5C shows the so-called urban village model that is o en shown in urban master plans but does not exist in the real world. In this model, there are many centers, but commuters travel only to the center that is the closest to their residence. is is a very attractive model for urban planners because it does not require much transportation or roads and it dramatically reduces VKmT and PKmT and, as a consequence, GHG emissions. According to this model, everybody could walk or bicycle to work even in a very large metropolis. e hypothesis behind this model is that urban planners are able to perfectly match work places and residences! is model does not exist in reality because it con-tradicts the economic justi cation of large cities. Employers do not select their employees on the basis of their place of residence, and specialized workers in large cities do not select jobs on the basis of their proximity from their resi-dence (with the exception of the very poor who walk to work and are limited to work within a radius of about 5 kilometers from their home). e “urban vil-lage model” implies a systematic fragmentation of labor markets, which would be economically unsustainable in the real world. e ve satellite towns built around Seoul are an example of the urban vil-lage conceit. When the towns were built, the number of jobs in each town was carefully balanced with the number of inhabitants, with the assumptions that these satellite towns would be self-contained in terms of housing and employ-ment. Subsequent surveys are showing that most people living in the new sat-ellite towns commute to work to the main city, and most jobs in the satellite towns are taken by people living in the main city. e “composite model” shown in gure 4.5D is the most common type of urban spatial structure. It contains a dominant center, but a large number of jobs are also located in the suburbs. In this type of city most trips from the suburbs to the CBD will be made by mass transit, whereas trips from suburb to suburb will use individual cars, motorcycles, collective taxis, or minibuses. e composite model is, in fact, an intermediary stage in the progressive trans-formation of a monocentric city into a polycentric one. As the city population grows and the built-up area expands, the city center becomes more congested and progressively loses its main attraction. e original raison d’être of the CBD was its easy accessibility by all the workers and easy communication within the center itself because of spatial concentration.
CITIES AND CLIMATE CHANGEAs a city grows, the progressive decay of the center because of congestion is not unavoidable. Good tra c management, timely transit investment, strict parking regulations and market price of o -street parking, investments in urban environment (pedestrian streets), and changes in land-use regulations allowing vertical expansion would contribute to reinforce the center, to make it attractive to new business, and to keep it as a major trip destination. ese measures have been taken with success in New York City, Shanghai, and Singa-pore, for instance. However, the policy coordination between investments and regulations is o en di cult to implement. is coordination has to be carried out consistently for a long period to have an impact on the viability of urban centers. Failure to expand the role of traditional city centers through infrastruc-ture and amenities investments weakens transit systems in the long run because the number of jobs in the center becomes stagnant or even decreases while all additional jobs are created in suburban areas. e comparison between the distributions of population in Jakarta (Jabo-tabek) and Gauteng ( gure 4.6) explains why Jakarta is able to successfully Figure 4.6 Spatial Distribution of Population in Jakarta and Gauteng Represented at the Same Scale Source: 2001 census.Gauteng: 8.7 Million peopleJakarta (Jabotabek) 16 Million peopleScale100,000 people50 km0
GHG EMISSIONS, URBAN MOBILITY, AND MORPHOLOGY
113implement a BRT network in addition to the existing suburban rail network, whereas in Gauteng suburban rail is carrying barely 8 percent of commuters, and the great majority of low-income commuters rely on microbuses. e dispersion of population in Gauteng is due in part to its history of apartheid. In the past 10 years, a very successful subsidized housing program has contributed to fur-ther disperse low-income people in distant suburbs while signi cantly attenu-ating the extreme poverty created by apartheid. e comparison as seen on the three-dimensional representation of population densities between the resulting city structure of Gauteng and that of Jakarta is striking. A BRT is being planned for the municipality of Johannesburg (one of the municipalities in the Gauteng metropolitan region), but the current urban structure will make it di cult to operate for a long time. In addition, the violent opposition of microbus opera-tors is making the project politically di cult. A change in transit mode involves a new equilibrium of transit types, which creates losers as well as winners. is is not an easy process, even when the nal long-range outcome seems desirable for all. e structure of cities is path dependent. Once a city is dominantly polycen-tric, it is nearly impossible to return to a monocentric structure. Monocentric cities, by contrast, can become polycentric through the decay of their tradi-tional center. e inability to adapt land-use regulations, to manage tra c, and to operate an e cient transit system are the three main factors that explain the decay of traditional CBDs.Transport Strategies Need to Be Consistent with Cities’ Spatial StructuresFindings concerning the relationship between urban spatial structures and transit can be summarized as follows:• Transit is e cient when trips’ origins are dispersed but destinations are concentrated.• Individual transport and microbuses are more e cient when origin and destinations of trips are both dispersed and for linked trips if amenities are dispersed.• Mode shi toward transit will happen only if price and speed are competi-tive with other modes.• Trips toward dense downtown areas (more than 150 people/ha) should be prevalently made by transit. Failure to provide e cient transit service to the CBD and to regulate tra c and parking would result in a dispersion of jobs in suburban areas, making transit ine cient as a primary means of trans-port in the long term.