WPS7366 Policy Research Working Paper 7366 Transport Policies and Development Claudia N. Berg Uwe Deichmann Yishen Liu Harris Selod Development Research Group Environment and Energy Team July 2015 Policy Research Working Paper 7366 Abstract This survey reviews the current state of the economic litera- assessed, especially those dealing with endogeneity issues in ture, assessing the impact of transport policies on growth, empirical studies. The specific implementation challenges inclusion, and sustainability in a developing country context. of transport policies in developing countries are discussed. The findings are summarized and methodologies are critically This paper is a product of the Environment and Energy Team, Development Research Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The corresponding author may be contacted at hselod@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Transport Policies and Development1 Claudia N. Berg, Uwe Deichmann, Yishen Liu, and Harris Selod2 Keywords: Transport, growth, poverty, sustainability JEL: O1, O2, R4, Q5                                                              1 The authors are grateful to Nathaniel Baum-Snow, Gilles Duranton, and staff from the World Bank’s Transport and ICT Global Practice for early discussions. Funding from DFID under the World Bank’s Strategic Research Program “Transport Policies of Sustainable and Inclusive Growth” (P151104) and under the World Bank’s Knowledge for Change Program “Impacts of Transport Infrastructure” (P133411) is gratefully acknowledged. Corresponding author: Harris Selod (email: hselod@worldbank.org, address: 1818 H Street, NW, Washington, D.C. 20433). 2 Berg, Deichmann, Selod: Development Research Group, The World Bank. Liu: The George Washington University. Transport Policies and Development 1. Introduction In developed economies, transport investments and improved transport technology over the last century have resulted in a continuous decline in transport costs, which in turn stimulated growth and economic development. In low- and middle-income countries, the current potential for transport investments and policies to boost sustainable and inclusive growth through declining transport costs also appears to be large. This is especially the case given significant backlogs of transport infrastructure investment in both rural and urban areas, weak governance and inadequate regulations in the transport sector, and rising social costs in terms of congestion, pollution and accidents, especially in emerging large cities. Transport investments can be very large and transformative in their nature, and their success depends on a variety of factors. Because these factors are not well understood and may not be taken into account by policy makers, there is often a risk that transport investments are not cost-effective and do not produce the range of expected outcomes. Thus, understanding and assessing how transport policies can produce growth-inducing effects and reduce social costs will be important for setting priorities in the strategic use of scarce resources. The objective of this paper is to review the broader direct and indirect benefits and costs of transport investments and policies in developing countries. The literature on transport impacts is large and covers a variety of issues, each study shedding light on a specific aspect. The largest share of papers looks at road and, to a lesser extent, at rail transport. Air and sea transport tends to be much less studied and this review reflects this bias in the literature. Recent surveys on this topic focus on the impact of transport infrastructure on the spatial distribution of economic activity, productivity and economic growth (see Redding and Turner, 2014, and Deng, 2013) but do not systematically review the existing evidence from developing countries. They also tend to overlook issues relevant to developing countries which have implications for transport policies. Policy makers may, for instance, wish to guide or accompany the ongoing process of structural transformation (i.e. the movement out of agriculture into industries and services) with interventions in the transport sector. Another such issue is the fast pace of urban growth currently occurring in 2 Africa and Asia, which requires massive transport infrastructure investments. Developing countries are also characterized by a low state of development of transport networks, potentially involving high returns on investment given decreasing returns to scale. The transport sector also often faces issues of weak governance and non-competitive practices that affect the construction of networks and the provision of transport services. Finally, the low density of public services in many developing countries also makes reliance on transport especially crucial. To account for the variety of mechanisms through which transport policies matter from a development perspective, we start in Section 2 by presenting a simple conceptual framework that details the potential links between the various interventions and impacts. In Section 3, we then present an econometric framework that discusses the identification challenges raised in the recent empirical literature. The following section summarizes the lessons learned from the literature, by distinguishing effects on growth, inclusion and sustainability, focusing to the extent possible but not exclusively on papers published on developing countries. Finally, Section 5 concludes by discussing the implementation challenges faced by policy makers.   2. Conceptual Framework There are three broad types of policies: direct transport infrastructure investments, price instruments and regulations. Investments may entail building new transport infrastructure (whether roads, rail, or airports), upgrading of existing links or of technology, or improvements of transport services. Price incentives may include subsidies or taxes to influence mode choice and transport behavior more generally (e.g. student fare reductions, tolls, parking fares, fossil fuel taxes, subsidies to clean transport). As for regulations, they may apply to a variety of approaches including rules to directly reduce emissions (e.g. fuel emission standards, driving restrictions) or to govern the organization of the transport sector (e.g. organization of freight, taxis, buses) or construction of infrastructure. Some policy interventions may affect supply, e.g. infrastructure investments, whereas others target demand, e.g. transport subsidies. 3 A useful categorization of the broader objectives of policies can be (i) to facilitate growth (e.g. through lower transport costs, which facilitates agglomeration effects, trade and structural change, and leads to higher productivity), (ii) to improve social inclusion (e.g. through better access to transport services, which can improve economic opportunities for the poor), and (iii) to promote sustainability (e.g. through reduced health and environmental externalities). The extent to which these broad objectives can be reached depends on the behavioral responses of firms and households to the particular policy interventions in terms of trade, location and mode choices. This framework is represented in Figure 1 below. Figure 1: Impacts of transport policies: the mechanisms  Policy    Focus of    Outputs    Responses    Outcomes  instruments  Intervention                    Physical  Δ Transport  Growth  infrastructure   costs  (Δ Production  (new, upgrading,  (incl. time costs)  Investments  and productivity)  Operations &  Trade  Maintenance )  Δ Access to  Price    Inclusion      transport  Location    instruments    (Δ Opportunity)  Transport  services /  services  connectivity  Transport use  Regulations  Sustainability  (Δ Environment  Technology  Δ Environmental  & quality of life)  externalities    The literature on the impact of transport policies covers a variety of interventions and outcomes at different geographic levels. Studies have analyzed many aspects from rural tracks and feeder roads all the way to global trade related transport. Given the variety of interventions, mechanisms and outcomes, a simple way to formalize the impact of transport policies is to consider how policies affect the welfare of individuals or groups, or profits of firms, through all possible channels. If we denote by i (=1…I) an individual/household (respectively a firm), and by j a location (=1…J), the indirect utility 4 of individual i in location j (respectively the profit function of firm i in location j) can be written as: , , , ,…, , (1) where (for e =1…E) are outcome functions which depend on the state of the transport system denoted by vector T, and on a vector of contextual variables that matter for outcome e denoted . Outcome functions may represent any argument entering the utility function of individuals or the profit function of firms. For individuals, outcome functions would include all relevant variables affected by transport such as employment status and wage (determined by productivity), access to amenities, prices of consumption goods, residential land rents, pollution externalities, health and education status, exposure to crime, etc.3 For firms, the outcome functions would include, in particular, the Marshallian externalities of learning, sharing and matching (see Fujita and Thisse, 2002) facilitated by the transport system. In this context, a transport policy can be written formally as a change in the transport system ∆T which induces a change in the indirect utility function of profit as given by:4 Δ ≡ Δ , , Δ , ,…, Δ , , , , ,…, , (2) It can be seen from (2) that, in general, several outcomes may happen simultaneously following a transport policy intervention. Many empirical studies are partial, usually focusing only on one of these outcomes (e.g. focusing on whether road investments increase wages) without considering the full set of effects. In addition, one should worry about the permanence of the effect and whether short-run effects hold in the long run. Another comment on (2) is that some effects can be desirable (e.g. an increase in wages from the perspective of worker) and others unwanted (e.g. pollution), with possible tradeoffs between different outcomes and agents affected by a given policy                                                              3 Individual indirect utilities mays be aggregated using a Social Welfare Function to characterize the welfare of a group. 4 For simplicity of exposure, we assume here that the contextual variables are not affected by the policy. The framework can easily be generalized by considering that the s are functions of . 5 intervention. For instance, better accessibility may improve access to jobs but can harm a subset of renters in a targeted area due to the capitalization of accessibility in housing values. Formula (2) allows for effects to be heterogeneous. Effects may indeed depend on the context ( ), for instance the initial state of the network. Policies may also affect heterogeneous households differently, for instance along the income distribution (as the poor, for instance, may respond differently to price incentives). When the policy objective is to reduce poverty, it will be important to be able to assess the impact of transport policies on the poor. Policies may also affect locations differently: an expansion of the transport network may induce relocation of activities from one place to another, with potential gains in one place and losses in the other. Therefore, it may be important to consider general equilibrium effects (including changes in prices) that can provide systemic insights into the effects of a given policy. Formally, policy makers need to consider the vector of changes for all agents in all locations: ∆ ≡∆ for i = 1,…, I and j = 1,…, J (3) where ∆ is given by equation (2). Depending on the focus, the literature has adopted various approaches ranging from equilibrium type models (spatial computable general equilibriums, structural estimation of trade and economic geography models5) to reduced form analysis, with different geographic scopes and identification strategies. In what follows, we summarize the main findings from the recent empirical literature, in light of relevant elements of theory as needed. 3. Empirical Framework Empirically estimating the causal impact of transportation is challenging. First, consistently with Figure 1, an assessment of policy impacts requires a clear understanding of the “treatment,” the outcomes to be evaluated (including the mechanisms by which the treatment affects each outcome), the impacted agents or areas, and the time horizon to be                                                              5 A particularly useful framework to assess the impacts of transport is the Eaton and Kortum (2002) model which yields structural equations for bilateral trade accounting for geographic barriers to trade, including high transport costs (see Burgess and Donaldson, 2012, Donaldson, forthcoming, and Donaldson and Hornbeck, 2013). 6 considered. Besides issues of data availability regarding outcomes, impacted agents, or time horizon, the choice of treatment and how it is measured deserves careful consideration. This is especially true in the case of infrastructure investment, where various measures are relevant (e.g. total expenditure, size of the network, quality of investments, etc.). Most empirical papers focus on improved accessibility associated with a measure of infrastructure investments. A second challenge is the implementation of an appropriate identification strategy given that the treatment variable of interest is likely to be endogenous. We discuss these two issues in the subsections below. 3.1 Choice of treatment Within the literature focused on evaluating the impacts of transport infrastructure, there are several treatment variables which are considered. In many instances, these variables can be thought of as indicators of access, ranging from proximity to transportation services to a variety of accessibility indicators accounting for locations of relevant markets and physical connections to those markets. A first approach is to consider the available local supply of transport infrastructure. Some authors use a dummy variable indicating the presence or absence of a road or railway within the geographic unit of observation (e.g. Atack and Margo, 2011; Atack et al 2010; Burgess and Donaldson, 2012; Chandra and Thomson, 2000; Datta, 2012; Faber, 2014; Jedwab and Moradi, forthcoming; and Michaels, 2008). Others consider the distance or time to the nearest highway (e.g. Faber, 2014; Emran and Hou, 2013; Ghani et al, 2012; Ahlfeldt and Wendland, 2011; Duranton et al, 2014, Duranton, 2014; Gibbons and Machin, 2005; and Banerjee et al, 2012). Still others take into account the road or rail density within a geographic unit of observation. For example, Duranton and Turner (2011) and Hsu and Zhang (2014) focus on lane- kilometers within metropolitan statistical areas. Baum-Snow (2007a, 2010) and Garcia- Lopez et al (2015) focus on the number of “rays” (radial highways) stemming from metropolitan areas. A second, more elaborate approach accounts for various measures of transport costs to markets. A difficulty with this approach is identifying the relevant market to consider. Dorosh et al (2012), for example, take distance to the nearest town in the context of agricultural goods. Ali et al (2015a, 2015b, 2015c) attempt to more accurately measure 7 the actual cost of reaching the market, specifically focusing on the cost of transporting one ton of goods to the least-costly-to-reach market.6 Several authors have used more sophisticated measures taking into account accessibility to multiple markets, improving upon the measure of market potential (Harris, 1954). For instance, Lall et al (2004) and Donaldson and Hornbeck (2013) focus on a “market access index”, which they note is superior to local road density measures as it allows one to consider how local accessibility is affected by changes occurring throughout the network. Other similar examples include Emran and Shilpi (2012), Hanson (2005), and Head and Mayer (2011). Sanches-Guarner (2012) and Gibbons et al (2012) study the impact of a local index of job accessibility.7 3.2 Identification challenges Considering an outcome function, , in equation (1), an econometric model of the impact of transport can be expressed in linear form as: (4) where is the outcome of interest for observation i (household, firm, area), is a measure of transport, and is a set of relevant exogenous variables. The parameter of interest is , which measures the impact of a change in on . Its estimation, however, will be biased in the presence of an endogenous . Sources of endogeneity In practice, endogeneity of the transport variable is very common. When measures transport infrastructure, a recurrent issue is non-random placement, which occurs when infrastructure is not placed independently of the outcome of interest. In practice, areas that attract transport infrastructure typically differ from those that do not. The bias in the estimates can go in either direction. For instance, in a regression of a measure of economic                                                              6 The authors calculated this cost using HDM-4 (Highway Development Management Model), a programming tool which takes into account road surface, condition, slope and distance.   7 These market access indices generally follow the same basic format, with minor variations: ∑ , where , market access at location i, is a function of , the “size” of market j, and the “distance” between location i and market j, with . a distance decay function. The measure of distance itself can be more or less elaborate, e.g. straight-line, distance through the network, time, or monetary costs. 8 activity on roads, if roads are built near economic growth centers, an estimation of equation (4) by ordinary least squares would overestimate . If instead, roads are built near disadvantaged areas, then the bias would be downwards. Another source of endogeneity relates to the issue of sorting across space which is inherent in spatial analyses. For instance, if equation (4) represents the impact of job accessibility ( ) on individual i’s employment status ( ), the fact that employed workers may decide to locate close to their jobs introduces endogeneity. Or if equation (4) represents the impact of market accessibility ( ) on rural incomes ( ), then the emergence of a market near areas of high productivity will also introduce a bias. Robust estimation of equation (4) requires a source of exogenous variation. We present below the most common identification strategies. Identification strategies Contexts where road placement is exogenous are scarce. A few authors restrict the sample to “inconsequential areas” (Redding and Turner, 2014) where they can defend that the infrastructure placement is exogenous. This is the approach adopted by authors who focus on rural places along roads that connect cities, excluding those endogenous nodes (Chandra and Thompson, 2000; Michaels, 2008). Restricting the analysis to specific areas, however, diminishes the external validity of findings. Alternatively, other authors have used quasi- or natural experiments as a source of exogenous variation for identification (such as a strike in public transport, Anderson, 2014, the opening of a metro system, Chen and Whalley, 2012, or the opening of metro stations, Gibbons and Machin, 2005). Other papers achieve identification by resorting to matching and difference-in-differences estimation strategy (Datta, 2012; Ghani et al, 2012). A number of papers correct for endogeneity by instrumenting the transport variable. In a two-stage least squares setting, which is most appropriate for presenting these issues, equation (4) is replaced by the following model: 4 5 9 where is the outcome of interest for observation (individual, household, firm, location) i; is a measure of connectivity; is a set of relevant exogenous variables; and is an exogenous instrumental variable. Because can be endogenous to the outcome of interest (e.g. due to reverse causality from to or unobserved variables in the residual, , correlated with both and ), the above identification strategy removes the source of bias in the estimation of the impact of on (second stage equation, 4) by instrumenting for with (first stage equation, 5). Since the instrument is by definition not correlated with the outcome, then only the exogenous sources of variations in are considered to identify the impact on the outcome of interest, . Choosing an appropriate instrument is key for a robust assessment of transport impacts. Several papers instrument existing transport networks with digitized old transportation networks or historical maps which are likely to be exogenous to the outcome of interest (see e.g. Baum-Snow, 2007a; Hymel, 2009; Donaldson, forthcoming; Duranton and Turner, 2012; Volpe Martincus, 2014; Duranton, et al, 2014; Baum-Snow et al, 2015; Garcia-Lopez et al, 2015; Ali et al, 2015c). Alternatively, some authors resort to the construction of an exogenous network based on straight lines between nodes (Ahlfeldt, 2011; Ahlfeldt and Wendland, 2011) and use it as an instrument. This addresses the endogeneity on the path between two nodes. Further refined approaches to address this issue of on-the-way endogeneity consider hypothetical networks that would best conform with geographic features that constrain existing networks (as in Faber, 2014; Emran and Hou, 2013; Ali et al, 2015a, 2015b). This is valid under the assumption that those geographic features are not correlated with the outcome of interest. Lastly, a few papers resort to the estimation of structural models (Roberts et al 2012), which reduces concerns about endogeneity issues. Table 1 in the appendix details a list of select empirical papers related to transport and their various identification strategies. 4. Lessons from the Literature In line with our conceptual framework (Section 2) and reviewing mainly studies that apply a robust identification strategy as presented in our empirical framework (Section 3), this section sequentially summarizes the links between transport policies and growth, inclusion, and sustainability. 10 4.1 Growth Investing in transport will reduce costs leading to an increase in productivity and shifting the economy to a higher growth equilibrium (see the Big Push Theory of Rosenstein-Rodan, 1943, further developed by Murphy et al, 1989, and Agenor, 2010). Empirical tests at the macroeconomic level confirm transport investments can have a significant impact on growth. Calderón et al (2015) estimate that the elasticity of output with respect to a synthetic infrastructure index—which includes transport along with electricity and telecommunications—ranges between 0.07 and 0.10. Similar impacts are found in the case of Sub-Saharan African countries.8 Other papers have looked at the different channels through which transport investments influence growth focusing on the impacts on firms’ decisions to trade and locate, on income generation and employment, as well as on the structural transformation of economies more generally. Trade A reduction in transport costs may stimulate the volume of trade, open up new markets, induce new industries to form, and thereby influence the patterns of trade. Trade costs are actually very high in much of the developing world. In Ethiopia and Nigeria, Atkin and Donaldson (2014) estimate that trade costs are four to five times larger than in the United States. A significant portion of trade costs is non-physical (e.g. costs and delays associated with border crossing, price markups of non-competitive transport firms, bribes), especially in African countries (Raballand et al, 2010). National borders have been estimated to reduce trade by 44 percent between the United States and Canada, and by 30 percent on average among other industrialized countries (Anderson and van Wincoop, 2003). Delays at borders and ports can last between 10 and 30 hours in Africa (Foster and Briceño-Garmendia, 2010). Reducing these costs will thus induce more intra- and inter-regional trade. Hummels (2007) shows that the decrease in transport costs over 50 years was a major driver of increase in international trade. At the regional level, Freund and Rocha (2011) find that a one day decrease in over-land travel time leads to a 7                                                              8 See Calderón and Serven, 2010, for the impact of a similar synthetic infrastructure index and Boopen, 2006, for an analysis of the impact of transport capital 11 percent increase in Africa’s exports. Simulations by Buys et al (2010) show that upgrading the primary road network connecting major cities would increase trade within Sub-Saharan Africa by $250 billion over 5 years (costing $20 billion for the initial upgrade plus $1 billion per year in maintenance). Reduction in trade costs also implies improved access to markets. Therefore, some papers directly measure the impact of better market access (proxying for better opportunities to trade) on various economic outcomes. In Sub-Saharan Africa, Dorosh et al (2012) find that reduced travel time to the nearest market (proxied by a city with at least 100,000 people) leads to an increase in agricultural production. Along the same vein, Bosker and Garretsen (2012) estimate that a 1 percent increase in a Sub-Saharan African country’s market access is associated with a 0.03 percent increase in its GDP per capita. Volpe Martincus et al (2014) find that Peru’s road improvement program led to a significant increase in firms’ average annual growth rate of exports (6.4 percent) and subsequently employment (5.1 percent). Emran and Hou (2013) show that both domestic and international market access stimulate per capita consumption in China. Reductions in transport costs may also reflect improvements in inter-city transport, which are likely to affect the diversification or specialization of cities. Anas and Xiong’s (2003) model predicts that low-cost transport infrastructure (e.g. railways and highways) should lead to more specialized cities. In an empirical paper, Duranton et al (2014) find that cities in the US that have more highways specialize in the export of heavy goods without seeing an impact on the value of goods traded. Duranton (2014) applies the same analysis in Colombia and finds, in contrast, that the effect of city roads is moderately stronger on the value of trade than it is on the weight of the goods traded, which is interpreted as roads shifting economic activities within cities towards somewhat lighter, tradable goods. Transport investments may have heterogeneous impacts across space through the stimulation of trade and specialization of industries. Chandra and Thompson (2000) find that industrial output in US counties benefited from connection to highways, but at the detriment of neighboring counties. In India, Donaldson (forthcoming) finds that colonial railways built in the 19th century lowered interregional trade costs and price gaps, increased trade flows, and increased real income per unit of land area. This in turn, increased incomes 12 within regions with railroads, but not necessarily in areas without railroads. In China, Faber (2014) finds that reducing transport costs can led to a reduction in industrial growth among connected peripheral regions relative to non-connected areas, as opposed to diffusing production from the metropolitan regions to the periphery. Firm location decision and agglomeration effects Lower transportation costs can lead to a re-allocation of manufacturing along the transport network. In this respect, Mayer and Trevien (2015) find that the initial construction of the regional express rail stations (RER) in the Paris area attracted manufacturing, business, and household service firms (especially foreign ones) in the municipalities where the stations were built. In Indonesia, Rothenberg (2013) finds that massive upgrades to the highway system during the 1990s led to the dispersion of manufacturing activities. The durable goods producers were more likely to disperse around cities while perishable goods producers were more likely to locate closer to their customers. Gutberlet (2013) finds that reduced transport costs led to a geographic concentration of industries in core regions of 19th century Germany, but that this occurred to the detriment of peripheral regions. Improved transport can also lead firms to be more productive. Better roads increases firm productivity through more intense vehicle use (Fernald, 1999 in the case vehicle-intensive industries) or less absenteeism (van Ommeren and Gutierrez-i- Puigarnau, 2011). The clustering of firms can further lead to increased productivity (due to agglomeration effects) or attract other more productive firms.9 In the case of India’s “Golden Quadrilateral,” Ghani et al (2012) find that there are higher entry rates of manufacturing firms near improved highways and that these firms have higher labor productivity and higher total factor productivity. Focusing on the same road improvement program, Datta (2012) finds that firms located near highway improvements are actually run more efficiently (as they keep inventories over a shorter period of time). At the scale of a city, transport investments can influence the location of economic activities within the city and affect residential patterns. In the US, radial highway investments caused suburbanization over the second half of the 20th century (Baum-Snow,                                                              9 Lower transportation costs may also lead to agglomeration effects by virtue of eased interactions between firms even if firms do not actually relocate. 13 2007a, 2007b). The suburbanization of jobs along with that of residences thus changed commuting patterns: instead of traveling to the central city, workers increasingly traveled within the suburbs (Baum-Snow, 2010).10 In Chinese cities as well, radial highways have led to the decentralization of both population and industries, (Baum-Snow et al, 2015). Applying similar methods, Garcia-López et al (2015) finds evidence of highways causing decentralization in Barcelona. For Spain as a whole, Garcia-López et al (2014) also find that highways caused suburbanization, but to a lesser extent than in the US. Economic Activity, Income, and Employment Increased trade and productivity results in greater production, improved labor market outcomes, and higher incomes. In the context of China, Banerjee et al (2012) find that infrastructure (roads and highways) have a positive effect on per capita GDP at the county level. Also in China, Roberts et al (2012) estimates a New Economic Geography model and finds that highways increased real incomes. In the case of Nigeria, Ali et al (2015a) find that reducing transport costs significantly increases local GDP but note that the full impact of transport costs on incomes may only emerge slowly over time. Jedwab and Moradi (forthcoming) measure the impact of railway access in Ghana (constructed around 1901) on economic development (as measured by nightlights) and on cocoa production. They show that railway access positively affects economic activity, both in the short and long run (path dependency).11 At the scale of Sub-Saharan Africa, Storeygard (2014) finds that cities close to a main port grow faster. In terms of employment, Duranton and Turner (2012) find that the initial stock of highways in the US caused higher city-level employment growth. The result is somewhat nuanced in a study on the UK in which Gibbons et al (2012) find that although roads affect firm entry and exit, they do not affect employment in existing firms. Regarding wages, Sanchis-Guarner (2012) highlights ambiguous possible effects from road construction resulting from two opposite forces: improved accessibility of firms that may result in higher productivity and thus higher wages, and increased competition among workers which can exert downward pressure on wages. As wages are found to                                                              10 Baum-Snow (2014) finds that radial highways caused more decentralization of population than jobs (and a heterogeneous response across sectors as some have strong incentives to remain spatially clustered). 11 A similar study for Kenya (Jedwab et al, 2014) finds further supportive evidence of those effects. 14 increase, the first effect seems to dominate the second. Gibbons et al (2012) using the same data and methodology find that accessibility changes induced by road improvements affects firm entry and exit but not the employment of existing firms. Structural transformation Improved transport networks may lead to the structural transformation of local economies. In rural communities, improved transportation may result in a shift from subsistence to commercial agriculture. In this respect, Ali et al (2015b) find that falling transportation costs lead to an increase in the production of high-input crops while crop production under a low-input regime is either unaffected or sees a decline in output. Further, the authors find that the likelihood that farmers use more modern techniques (e.g. mechanized agriculture) in their agricultural production increases with the fall in transport costs. Reduced transportation costs can also lead to a shift of production and labor away from the agriculture sector. Chandra and Thomson (2000) note that falling transportation costs led to a reallocation of labor from agriculture to manufacturing in the US. Van de Walle (2009) finds that the construction of a new road in Vietnam was followed by the emergence of new non-farm activities. Similarly, Gertler et al (2014) find that improved road quality in Indonesia increases job creation in the manufacturing sector and triggers an occupational shift of workers from agriculture to manufacturing. In Nigeria, Ali et al (2015a) find that falling transportation costs both decrease the probability of agricultural employment by households, and increase the likelihood of them being fully employed. This is indicative of a shift of workers from agricultural to non-agricultural jobs. When transport costs remain high, however, as in the rural areas of Sub-Saharan Africa, Gollin and Rogerson’s (2014) model shows that people remain located near the spatially diffused sources of food production, thus preventing structural transformation by hindering the movement of people out of subsistence into the modern sector. 4.2 Inclusion High transport costs may also have direct effects on various dimensions of poverty. The literature has focused on how poor transport can affect vulnerable groups through reduced trade, adverse labor market outcomes, poor education and health, as well as crime. 15 Rural Poverty In rural areas of developing countries, especially in Africa, there is a general lack of infrastructure investment given the major costs required, lack of available funds and political will (see, for instance, Blimpo et al, 2013 who show that there is underinvestment in roads in politically marginalized areas in four West African countries). This is all the more problematic as poor accessibility can be especially harmful for the rural poor. In Uganda, Kyeyamwa et al (2008) find that road high transport costs deter farmers from participating in local markets to sell their cattle, relying instead on farm gate sales. Similarly, in rural Kenya, land devoted to cash crops is only located close to markets (Omamo, 1998). There is therefore a large potential for transport infrastructure investments to improve the living standards of the rural poor through stimulation of economic activities. Different dimensions of poverty may be impacted. Jacoby and Minten (2009) find that lower transport costs increase remote household income in Madagascar. In rural China, Emran and Hou (2013) find that better access to domestic and international markets improve per capita consumption and the livelihoods of the poor. Focusing on Bangladesh, Khandker et al (2009) find that rural road investments reduce poverty, including through higher agricultural production, higher wages, and lower input costs, and higher output prices. They conclude that road investments are pro-poor, with gains that are proportionately higher for the poor than for the non-poor. This is consistent with the finding that benefits from rural road improvements are greater in poor communities, where levels of initial market development are lower (see Mu and van de Walle, 2011, on Vietnam). Impacts on inequality are more nuanced, as Jacoby (2000) finds that although roads improve welfare of poor rural households in Nepal, they do not reduce inequality. Urban Labor Markets Regarding the labor market impacts of transport costs, a large urban literature has focused on how poor physical connections between jobs and residences resulted in exacerbating the unemployment and low wages of unskilled workers. This so-called spatial mismatch literature originated in the US context, where jobs suburbanized over the past 70 years, whereas unskilled workers, especially among minorities, remained in inner cities 16 (Kain, 1968).12 However, this literature has much broader applicability, especially to sprawling cities in the developing world where the disconnection from jobs could be an important contributor to poverty (see, for example, Rospabé and Selod, 2006, in the case of Cape Town, South Africa). The mechanism of spatial mismatch (see Gobillon et al, 2007 and Gobillon and Selod, 2012, 2014 for reviews of this literature) revolves around the role of high transport costs in deterring the unemployed from accepting distant jobs, the harmful effects of long commutes on productivity or decreasing productivity, or high search costs that make the matching between unemployed workers and jobs less efficient. This is most relevant for unskilled workers who are disadvantaged on the labor market and for whom spatial connection to jobs matters a lot (see Selod and Zenou, 2006). Robust empirical studies have shown a causal effect of job decentralization on the unemployment of the less mobile inner city poor (see Zax and Kain 1996 and Weinberg 2000). In this context, there is an important role for transport policies to better connect people to jobs.13 Other vulnerable groups affected by spatial mismatch include women who have fewer transport choices than men (Blumenberg, 2004) and pay a large share of their income on transport, as compared to men (Babinard and Scott, 2011). Having fewer transport choices also induces women entrepreneurs to locate their business closer to their residences, thus missing out on the benefits of agglomeration in business districts (Rosenthal and Strange, 2012). In this context, improving access to transportation may have a significant impact on reducing unemployment of the targeted population. Sanchis-Guarner (2012) for instance finds that improved accessibility to work from road constructions in the UK increases wages and hours worked. In the US context, raising the car ownership rates of minorities would considerably narrow the differences in the unemployment rates between groups (Raphael and Stoll, 2001). Transport subsidies to unemployed workers have also been shown to reduce unemployment spells by facilitating search (see Philips, 2014, for a randomized control trial in the Washington, D.C. metro area).                                                              12 Although the original framing of the spatial mismatch hypothesis insisted on the role of barriers to residential mobility (mainly housing discrimination), the disconnection of the poor from job locations may also be the result of freely functioning markets in the context of spatial sorting. 13 In combination with other policies, either facilitating the residential mobility of workers (see Katz et al, 2001) or providing incentives to firms to locate in disadvantaged areas (Gobillon et al, 2011). 17 Education Upstream to the labor market, transport costs also have an impact on educational opportunities and choices. In the rural areas of Bangladesh, Khandker et al (2009) find that improving rural roads leads to higher rates of both boys’ and girls’ enrollment in school. Similarly, Jacoby and Minten (2009) finds a positive impact of road investments on higher secondary school enrollment in Madagascar. Health There are also potential food security and health benefits from better transportation. Blimpo et al (2013) find that areas with less transport access suffer from food security problems, as evidenced by stunting in West Africa. Food prices are indeed correlated with transport costs (related to road quality) as evidenced by Minten and Kyle (1999) in the case of the Democratic Republic of Congo. There is also evidence that penetration of rural areas by railroad helps poor communities be more resilient to negative agricultural productivity shocks threatening the food supply (see Burgess and Donaldson, 2012, on 19th and early 20th century India). More generally, lowering transportation costs significantly reduces the probability that a household is multi-dimensionally poor through improvements in health, education, and standard of living (see Ali et al, 2015a, who correlate the multi-dimensionally poverty index14 to transportation cost in the case of Nigeria). Crime Finally, as the poor are disproportionately exposed to crime, a trend of the literature has also focused on the relation between crime and transport. Crime in transport is a serious issue in poor developing countries. This is illustrated by the case of South African urban public transport which has high incidences of murder, rape, and assault (see Kruger and Landman, 2007). Women are especially vulnerable to security issues while traveling especially because they are more dependent on public transport (Babinard and Scott, 2011). Beyond crime repression, investment in physical infrastructure has an impact on reducing crime. In the case of Bogotá, Colombia, Marcelo (2013) finds that crime around Transmilenio (bus rapid transit) stations decreases, an effect the author attributes to better lighting at night. The construction of the cable cars in                                                              14 See Alkire and Foster (2011) and Alkire and Santos (2011). 18 Medellín, Colombia (Metrocable) was accompanied by neighborhood upgrading (new social housing, schools, and other infrastructure improvements) which resulted in a decline in violence (Brand and Dávila, 2011). The relation between transport availability and crime location is controversial and not well researched. In well-off US suburban areas, for instance, there is a perception that public transport would bring crime to the residential suburbs by making travel easier for potential criminals living in inner cities. This perception has been at the origin of political opposition to public infrastructure extension. Interestingly, Ihlandfeldt (2003) finds, on the contrary, that the construction of stations in Atlanta, Georgia decreased crime in the suburbs. The author argues that while rail access may have made neighborhoods more accessible to outside criminals, they may also have increased the opportunity cost of crime (as access to legitimate employment becomes easier).15 It should not be concluded, however, that improved transport will necessarily decrease crime. In the specific context of conflict areas in the Democratic Republic of Congo, Ali et al (2015c) have found that more accessible areas are more exposed to violent events, causing a decrease in population welfare. 4.3 Sustainability Transport may have important social costs, which need to be balanced against potential positive economic impacts. These costs involve negative externalities, ranging from congestion, accidents, impact on health caused through air pollution and the easier spread of epidemics, as well as direct costs to the environment such as deforestation, biodiversity loss, and more generally degradation of ecosystems. In the long run, some of these negative effects may even be harmful to growth. Transport policies thus have a role to play to minimize and mitigate negative impacts. We first describe the problems and then discuss the role of policies to address these issues. Negative Socio-Economic Impacts Congestion in transport is a major problem faced by economies—both developed and developing alike given its high opportunity cost (time spent in traffic could be spent in                                                              15 Criminals do not travel far (Chainey and Ratcliffe, 2013). 19 leisure or more productive pursuits). Long commuting times may also be detrimental to employment growth, as demonstrated by Hymel (2009) who estimates the causal impact of traffic congestion on aggregate employment growth in US metropolitan areas. Congestion is also problematic given its contribution to air pollution caused by vehicle emissions, especially by cars. In developing countries, although car ownership is only a fraction of that of the US, experience of traffic congestion and air pollution is often far worse (Sperling and Claussen, 2004). Fatalities and injuries on the road exert a toll, especially on the economies of developing countries. Cubí-Mollá and Herrero (2012) report that in 2004 over 50 percent of road fatalities worldwide were among economically active young adults, aged 15-44. Even when accidents are non-fatal, they can still be economically catastrophic, leaving people too disabled to work. It has been noted that transport (roads) can also help spread infectious diseases, as evidenced by AIDS following the road network in southern Africa (see Regondi et al, 2013). Mobility restrictions imposed during the recent Ebola epidemics onto affected countries, are likely to have had harmful economic impacts by interrupting the flow of trade (World Bank, 2014). Finally, transport infrastructure may disturb the ecosystem through deforestation, biodiversity loss, pollution, road kill, and blocking of seasonal migration patterns (see Chomitz and Gray, 1996; Laurance et al, 2009). 16 The negative impacts of roads may increase over time. In the Brazilian Amazon, for instance, the Belém-Brasília Highway, completed in 1970, today cuts a 400-kilometer-wide path through the rainforest (Laurance and Balmford, 2013). Designing Sustainable Transport Policies Policies to address these issues, whether through investment, pricing, or regulation, will influence either the supply of or the demand for transport. There are different policy options available, generally with nuanced impacts. Addressing congestion and air pollution through an increase in the supply of infrastructure, although intuitive, has been shown to be inefficient in the case of highways in the US. Duranton and Turner (2011) show that additional road investments in the US                                                              16 See Damania and Wheeler (2015) for an index measuring biodiversity losses. 20 actually increases traffic by either drawing traffic from alternative roads, stimulating commercial trucking, or attracting new drivers. Similarly, Hsu and Zhang (2013) replicate the analysis for Japan and find that road construction induces increased traffic. Along the same line, Anas and Timilsina (2009) estimate that although new road construction may reduce emissions of existing cars through increased travel speed, it may ultimately increase overall emissions by attracting additional drivers. Alternatively, a more efficient approach can be to target investments at public transport systems. Investments in rail in US cities was indeed found to draw commuters away from buses and, in some cases, away from cars (Baum-Snow and Khan, 2005). The public transit system in Los Angeles was found to significantly reduce congestion (Anderson, 2014). The opening in 1996 of the Taipei metro reduced carbon monoxide emissions between 5 and 15 percent (Chen and Whalley, 2012). Infrastructure investments also have a direct impact on the shape of cities, which in turn affects the level of congestion and emissions within cities. In particular, the pace of suburbanization (a historic trend associated with rising incomes) is accelerated by transport infrastructure investment and the subsequent fall in transport costs (Brueckner, 2000). There is indeed ample evidence that highways and railways have caused the decentralization of population at the city level (Baum-Snow, 2007, for the US; Garcia- López et al, 2013, for Spain; Ahlfeldt and Wendland, 2011, for Germany; and Baum-Snow, 2013, for China) as well as of jobs (Baum-Snow, 2010, for the US and Baum-Snow, 2013, for China). Gonzalez-Navarro and Turner (2014), using a panel of 138 cities around the world, confirm the role played by subways in decentralization. These evolutions have contributed to urban sprawl (Brueckner, 2000) where extensive car use increased carbon emissions (Glaeser and Kahn, 2003). For developing countries, especially those at early stages of urbanization and experiencing fast spatial expansion and population growth, there appears to be a role for transport policies to influence the evolving shape of the city in a sustainable way (Suzuki et al, 2013). Once cities have suburbanized, investment in public transit to mitigate car use is very costly. The impact of transport policies is also dependent on city shape and may have heterogeneous effects throughout a city. Using a core- periphery model calibrated on Beijing, China, Anas and Timilsina (forthcoming) show that more efficient transport in the core area would attract population and thereby reduce carbon 21 dioxide emissions, whereas lower transport costs in the periphery would encourage suburbanization, thus increasing carbon dioxide emissions. On the demand side, pricing, through subsidizing “good” behavior or taxing “bad” behavior, can also be a useful tool to alter incentives. Subsidizing public transport may reduce emissions by incentivizing people to switch from driving to using public transit.17 The price instrument, however, is seldom used in developing countries (see Timilsina and Dulal, 2011 on urban road transportation externalities). Subsidies are costly and taxation is politically difficult to implement. 18 Taxation (either a toll or fuel tax) can be efficient in reducing emissions by discouraging car use in cities (Anas and Timilsina, 2009). Graham and Glaister (2002) estimate for the US that a 10 percent increase in price, would reduce consumption by 3 percent in the short run and by 6 percent in the long run. These different price elasticities of fuel consumption suggest that there is a gradual adaptation of behavior. Consistently with these results for the US, Anas et al (2009), using a calibrated land-use and commuting- choice model for Beijing, China, find that a 10 percent increase in monetary cost of travel would reduce carbon dioxide emissions by 1 percent.19 In the case of France, Givord et al (2014) finds that both carbon and diesel taxes would slightly reduce carbon dioxide emissions in the short-term through fewer car purchases. Tolls should also reduce congestion and emissions by incentivizing drivers to switch to public transport (Anas and Lindsay, 2011).20 Rather than relying on price incentives to influence behavior, policy makers can mandate the desired behavior through a variety of regulations (e.g. catalytic converters, fuel efficiency standards, driving restrictions). In theory, fuel efficiency regulations may only have limited, short-term impacts on air pollution when they apply only to new cars and can only be gradually phased in (Anderson et al, 2011). Their impact may also be                                                              17 Note, however, that the effect may be offset by inducing pedestrians to use public transportation. 18 See Anas and Lindsay (2001) and Timilsina and Dulal (2011) for a review of externality pricing. 19 For other examples of calibrated models exploring the effects of externality pricing in developing country cities, including Mexico City (Anas and Timilsina, forthcoming; Anas et al 2009); Cairo (Parry and Timilsina, 2010); and Sao Paolo (Anas and Timilsina, 2009). These models are able to compare the efficiency of different taxes (e.g. gasoline tax vs. car tolls). 20 As for infrastructure investments, it is notable that pricing instruments may also have an impact on city shapes by influencing location and travel decisions (see Brueckner, forthcoming, in the case of cordon toll pricing). 22 offset by an increase in the vehicle-miles driven causing higher induced fuel consumption and emissions (see Small and Van Dender, 2007; Anas and Timilsina, 2009). Other types of regulations include driving restrictions during periods of high pollution generation. Quito, Ecuador implemented a vehicle use restriction program during peak driving (Pico y Placa) which was found to efficiently reduce emissions by between 9 to 11 percent during peak hours and by 6 percent during the day (see Carrillo et al, forthcoming). In Santiago, Chile, peak-hour vehicle-use restrictions led to a shift to public transport systems as well as a shift in arrival and departure times (De Grange and Troncosa, 2011). These studies, however, tend to focus on the relatively short-term. Bonilla (2012) argues that in the long run, households will respond to such programs by purchasing additional vehicles thus reducing the effectiveness of the program. This argument is consistent with the findings of Eskeland and Feyzioglu (1997) and Davis (2008) who study car use restrictions in Mexico (i.e. the “Hoy No Circula” program) and find that it does not improve air quality as it actually induces an increase in the number of vehicles in circulation, with people buying a second—often older and thus less fuel efficient—car. Other examples of regulations include more drastic measures such as those implemented by China in anticipation of the 2008 Olympic Games who combined plant relocation and furnace replacement with new emission standards and traffic controls. Chen et al (2013) find that pollution fell during and shortly after the games, but this effect was short lived given the temporary nature of the measures. Examples of potential regulatory tools aimed at reducing the negative environmental impact caused by road construction in forested areas are found in Damania and Wheeler (2015). They develop a novel index of biodiversity loss to guide the planning of road investments in ecologically rich areas. 5. Discussion This literature review presented the diverse mechanisms through which transport policies can induce beneficial outcomes, in terms of sustainable growth and inclusion, as well as the tradeoffs encountered. Transport infrastructure investments play a major role in the process of structural transformation, urbanization and city growth. Price instruments 23 and regulations can also be useful tools to affect behavior and address environmental externalities. In developing countries, however, transport policies are often poorly designed and implemented. In particular, there remains a large backlog of transport infrastructure investment especially in Africa (see Foster and Briceño-Garmendia, 2010) and price instruments are not widely used (Timilsina and Dulal, 2011). Implementation of transport policies in developing countries faces a number of challenges. Obtaining funding for transport infrastructure investments is difficult in contexts of scarce resources and limited fiscal capacity (Sperling and Claussen, 2004). This is exacerbated by the fact that construction of infrastructure is particularly costly in developing countries given non-competitive procurement (Estache and Iimi, 2009) or corruption (Collier et al, 2015). Today, in addition to using government resources and borrowing, there is an increasing emphasis on leveraging funds through public-private partnerships (see Dintilhac et al, 2015). However, this requires the presence of a private operator and appropriate coordination between the public sector (that sets the strategy) and the private sector (that provides funds, construction, and operation). An innovative funding mechanism (although dating back to the 19th century) is the capture of land value increases following an infrastructure investment (see Peterson, 2009).21 However, for this strategy to work, property rights must be well defined to allow for taxation, which is not the case in much of the developing world. When the infrastructure is in place, operation may also require scarce funding or be hindered by inefficient management (Bogart and Chaudhary, 2012; Lowe, 2014), non- competitive market structures of service providers (Lall et al, 2009) or excessive regulations (Button and Keeler, 1993; Combes and Lafourcade, 2005), which further drives up user costs. These non-physical costs may represent a significant share of total transport cost, in particular in African countries (Teravaninthorn and Raballand, 2008). Another challenge, although not restricted to developing countries, revolves around the politics that underpin transport investment choices and the resulting allocative inefficiencies. Ethnic favoritism and political clientelism influence the placement of road                                                              21 Investments that improve a location accessibility will raise land values. For empirical studies in an urban context, see Baum-Snow and Kahn (2000), Boarnet and Chalermpong (2003), Gibbons and Machin (2005) and Gonzalez-Navarro and Quintana-Domeque (2014). For increased accessibility and land price increases in rural areas, see Atack and Margo (2011) and Donaldson and Hornbeck (2013). 24 investments (see Burgess et al, forthcoming, and Blimpo et al, 2014). This occurs at the detriment of investment in other areas where the return of investment could have been greater, or that could have a greater poverty-reduction impact (for example, in light of the “pro-poor” effect of rural feeder roads, see Jacoby, 2000, van de Walle, 2002). There can be complementarities between transport policies. For instance, the effectiveness of price incentives in shifting people towards greater public transport use also depends on the quality and capacity of that network (Anas and Lindsay, 2011, Sperling and Claussen, 2004). Transport policies may also need to be combined with non-transport policies to produce the intended effects. For instance, better road connections will have an amplified impact on the development of commercial agriculture when combined with expanded cell phone coverage (Casaburi et al, 2013), agriculture extension services, or access to credit (Ali et al 2015b). Similarly, urban renewal programs may require both transport investments and neighborhood investments (Brand and Dávila, 2011). Connecting unemployed workers to jobs may produce little effect in the absence of training to reduce the skill mismatch. At the macro scale, transport investment will have transformative impacts provided an enabling environment is present that allows for the development of a manufacturing sector and trade. 25 References Agénor, P. (2010) A theory of infrastructure led development, Journal of Economic Dynamics and Control, 34, 932-50. Ahlfeldt, G.M. (2011) “The train has left the station: do markets value intracity access to intercity rail connections?” German Economic Review, 12(3), 312-335. Ahlfeldt, G. and N. Wendland (2011) “Fifty Years of Urban Accessibility: the Impact of the Urban Railway Network on the Land Gradient in Berlin 1890-1936,” Regional Science and Urban Economics, 41(2), 77-88. Ali, R., A.F. Barra, C.N. Berg, R. Damania, J. Nash, and J. Russ (2015a) “Transport Infrastructure and Welfare: An Application to Nigeria,” Policy Research Working Paper, WPS 7271, World Bank. Ali, R., A.F. Barra, C. Berg, R. Damania, J. Nash, and J. Russ (2015b) “Agricultural Technology Choice and Transport,” Policy Research Working Paper, WPS 7272, World Bank. Ali, R., A.F. Barra, C. Berg, R. Damania, J. Nash, and J. Russ (2015c) “Infrastructure in Conflict-Prone and Fragile Environments: Evidence from the Democratic Republic of Congo,” Policy Research Working Paper, WPS 7273, World Bank. Alkire, S. and J.E. Foster (2011) “Counting and multidimensional poverty measurement.” Journal of Public Economics, 95(7) 476–487. Alkire, S. and M.E. Santos (2011) “Acute multidimensional poverty: A new index for developing countries,” OPHI Working Paper No. 38. Anas, A. and R. Lindsey (2011) “Reducing urban road transportation externalities: Road pricing in theory and in practice,” Review of Environmental Economics and Policy, 5(1), 66-88. Anas, A. and G. Timilsina (forthcoming) “Offsetting the CO2 locked-in by roads: Suburban transit and core densification as antidotes,” Economics of Transportation, in press. Anas, A. and G. Timilsina (2009) “Impacts of Policy Instruments to Reduce Congestion and Emissions from Urban Transportation: The Case of Sao Paulo, Brazil,” Policy Research Working Paper No. 5099. The World Bank, Washington, D.C. 26 Anas, A., G. Timilsina, and S. Zheng (2009) “An Analysis of Various Policy Instruments to Reduce Congestion, Fuel Consumption and CO2 Emissions in Beijing,” Policy Research Working Paper No. 5068. The World Bank, Washington, D.C. Anas, A and K. Xiong (2003) “Intercity Trade and the Industrial Diversification of Cities,” Journal of Urban Economics, 54, 258-276. Anderson, M.L. (2014) “Subways, Strikes, and Slowdowns: The Impacts of Public Transit on Traffic Congestion,” American Economic Review, 104(9), 2763-96. Anderson, S. T., I.W. Parry, J.M. Sallee, and C. Fischer (2011) “Automobile fuel economy standards: Impacts, efficiency, and alternatives,” Review of Environmental Economics and Policy, 5(1), 89-108. Arnott, R.J. (1979) “Unpriced transport congestion,” Journal of Economic Theory, 21, 294-316. Atack, J., F. Bateman, M. Haines and R. Margo (2010) “Did Railroads Induce or Follow Economic Growth? Urbanization and Population Growth in the American Midwest, 1850-1860,” Social Science History, Vol. 34, No. 2, pp. 171–197. Atack, J. and R. Margo (2011) “The Impact of Access to Rail Transportation on Agricultural Improvement: The American Midwest as a Test Case, 1850-1860,” Journal of Transport and Land Use, 4(2), 5-18. Atkin, D. and D. Donaldson (2014) “Who’s Getting Globalized? The Size and Implications of Intranational Trade Costs,” mimeo, Yale University. Babinard, J. and K. Scott (2011) “What Do Existing Household Surveys Tell Us About Gender and Transportation in Developing Countries?” Women's Issues in Transportation, 2, 213-224. Banerjee, A., E. Duflo, and N. Qian (2012) “On the road: Access to transportation infrastructure and economic growth in China,” (No. w17897). National Bureau of Economic Research. Baum-Snow, N. (2007a) “Did Highways Cause Suburbanization?” Quarterly Journal of Economics, 122(2) 775-805. Baum-Snow, N. (2007b) “Suburbanization and Transportation in the Monocentric Model,” Journal of Urban Economics, 62(3): 405-423. 27 Baum-Snow, N. (2010) “Changes in Transportation Infrastructure and Commuting Patterns in U.S. Metropolitan Areas, 1960-2000,” American Economic Review Papers & Proceedings, 100(2) 378-382. Baum-Snow, N. (2014) “Urban Transport Expansions, Employment Decentralization, and the Spatial Scope of Agglomeration Economies,” mimeo. Baum-Snow, N. and M. Kahn (2005) “Effects of urban rail transit expansions: Evidence from sixteen cities, 1970-2000,” Brookings-Wharton Papers on Urban Affairs: 1(4):147–197. Baum-Snow, N., L. Brandt, J.V. Henderson, M.A. Turner, and Q. Zhang (2015) “Roads, railroads and decentralization of Chinese cities,” Processed, University of Toronto. Blimpo, M., R. Harding, and L. Wantchekon (2013) “Public Investment in Rural Infrastructure: Some Political Economy consideration,” Journal of African Economies, 22(2), 57-83. Blumenberg, E. (2004) “En-gendering Effective Planning: Spatial Mismatch, Low- Income Women, and Transportation Policy,” Journal of the American Planning Association, 70(3), 269-281. Boarnet, M. and S. Chalermpong (2003) New Highways, House Prices, and Urban Development: A Case Study of Toll Roads in Orange County, CA, Housing Policy Debate, 12(3), 575-605. Bonilla, J. (2012) “Effects of the Increased Stringency of Driving Restrictions on Air Quality and Car Use in Bogota,” Department of Economics, University of Gothenburg. Boopen, S. (2006) “Transport Infrastructure and Economic Growth: Evidence from Africa Using Dynamic Panel Estimates,” The Empirical Economic Letters, 5(1) 37–52. Bosker, M. and H. Garretsen (2012) “Economic geography and economic development in Sub-Saharan Africa,” The World Bank Economic Review, 26(3), 443-485. Brand, P. and J. Dávila (2011) “Mobility innovation at the urban margins: Medellín’s Metrocables,” City, 15(6), 647-661. Brueckner, J. (forthcoming) “Cordon Tolling in a City with Congested Bridges,” Economics of Transportation, in press. 28 Brueckner, J.K. (2000) “Urban Sprawl: Diagnosis and Remedies,” International Regional Science Review, 23(2), 160-171. Burchfield, M., H.G. Overman, D. Puga, and M.A. Turner (2006) “Causes of Sprawl: A Portrait from Space,” The Quarterly Journal of Economics, 121(2), 587-633. Burgess, R. and D. Donaldson (2012) “Railroads and the Demise of Famine in Colonial India,” mimeo. Burgess, R., R. Jedwab, E. Miguel, A. Morjaria, and G. Padró i Miquel (forthcoming) “The Value of Democracy: Evidence from Road Building in Kenya,” Forthcoming in the American Economic Review. Button, K. and T. Keeler (1993) “The Regulation of Transport Markets,” The Economic Journal, 103(419), 1017-27. Buys, P., U. Deichmann, D. Wheeler (2010) “Road network upgrading and overland trade expansion in sub-Saharan Africa,” Journal of African Economies, 19(3), 399-432. Calderón, C., E. Moral-Benito, and Luis Servén (2015) “Is Infrastructure Capital Productive? A Dynamic Heterogeneous Approach,” Journal of Applied Econometrics, 30, 177-198. Calderón, C. and L. Serven (2010) “Infrastructure and Economic Development in Sub- Saharan Africa,” Journal of African Economics, 19, AERC Supplement 1, i13–i87. Carrillo, P., A. Malik and Y. Yoo (forthcoming) “Driving Restrictions That Work? Quito’s Pico y Placa Program,” Canadian Journal of Economics, in press. Casaburi, L., R. Glennerster, and T. Suri (2013) “Rural Roads and Intermediated Trade: Regression Discontinuity Evidence from Sierra Leone,” mimeo. Chainey, S. and J. Ratcliffe (2005) GIS and Crime Mapping, Chichester: John Wiley & Sons Ltd. Chandra, A. and E. Thompson (2000) “Does Public Infrastructure affect Economic Activity? Evidence from the Rural Interstate Highway System” Regional Science and Urban Economics, 30, 457-90. Chen, Y. and A. Whalley (2012) “Green infrastructure: the effects of urban rail transit on air quality,” American Economic Journal: Economic Policy, 4(1), 58-97. 29 Chen, Y., G. Jin, N. Kumar, and G. Shi (2013) “The Promise of Beijing: Evaluating the Impact of the 2008 Olympic Games on Air Quality," Journal of Environmental Economics and Management, 66(3), 424–443. Chomitz, K.M. and D.A. Gray (1996) “Roads, Land Use, and Deforestation: A Spatial Model Applied to Belize,” The World Bank Economic Review, 10(3), 487-512. Collier, P., M. Kirchberger, and M. Söderbom (2015) “The cost of road infrastructure in low and middle income countries,” mimeo. Combes, P.P. and M. Lafourcade (2005) “Transport Costs: Measures, Determinants, and Regional Policy Implications for France,” Journal of Economic Geography, 5, 319-349. Cubí-Mollá, P. and C. Herrero (2012) “Quality of life lost due to non-fatal road traffic injuries,” Health Economics, 21(5), 528-50. Damania, Richard and David Wheeler (2015) “Road Improvement and Deforestation in the Congo Basin Countries” World Bank Policy Research Working Paper 7274. Datta, S. (2012) “The impact of improved highways on Indian firms,” Journal of Development Economics, 99(1), 46-57. Davis, L. (2008) “The Effect of Driving Restrictions on Air Quality in Mexico City,” Journal of Political Economy, 116(11) 38-81. De Grange, L. and R. Troncoso (2011) “Impacts of Vehicle Restrictions on Urban Transport Flows: The Case of Santiago, Chile,” Transport Policy, 18(6), 862-869. Dementyeva, M., P.R. Koster, and E.T. Verhoef (forthcoming) “Regulation of road accident externalities when insurance companies have market power,” Journal of Urban Economics, in press. Deng, T. (2013). Impacts of Transport Infrastructure on Productivity and Economic Growth: Recent Advances and Research Challenges. Transport Reviews, 33, 6, 686- 699. Dintilhac, C., F. Ruiz-Nunez, and Z. Wei (2015) “The Economic Impact of Infrastructure and Public-Private Partnerships: Literature Review,” The World Bank, mimeo. Donaldson, D. (forthcoming) “Railroads of the Raj: Estimating the impact of transportation infrastructure,” American Economic Review, in press. Donaldson, D. and R. Hornbeck, R. (2013). “Railroads and American Economic Growth: A ‘Market Access’ Approach,” NBER Working Paper 19213. 30 Dorosh, P., H.G. Wang, L. You, and E. Schmidt (2012) “Road connectivity, population, and crop production in Sub‐Saharan Africa,” Agricultural Economics, 43(1), 89-103. Duranton, G. (2014) “Roads and Trade in Colombia,” mimeo. Duranton, G., P. Morrow, and M.A. Turner (2014) “Roads and Trade: Evidence from the US,” Review of Economic Studies, 81(2), 681-724. Duranton, G. and M.A. Turner (2011), “The Fundamental Law of Road Congestion: Evidence from the US,” American Economic Review, 101(6), 2616-2652. Duranton, G. and M.A. Turner (2012) “Urban growth and transportation,” The Review of Economic Studies, 79(4), 1407-1440. Eaton, J. and S. Kortum (2002) “Technology, Geography and Trade,” Econometrica, 70(5) 1741-1779. Emran, M. S. and Z. Hou (2013) “Access to markets and rural poverty: evidence from household consumption in China,” Review of Economics and Statistics, 95(2), 682-697. Emran, M.S. and F. Shilpi (2012) “The extent of the market and stages of agricultural specialization,” Canadian Journal of Economics, 45(3) 1125-1153. Eskeland, G.and T. Feyzioglu (1997) “Rationing Can Backfire: The ‘Day without a Car’ in Mexico City,” World Bank Economic Review, 11(3), 383-408. Estache, A. and A. Iimi (2009) “Joint Bidding, Governance and Public Procurement Costs: A Case of Road Projects,” Annals of Public and Cooperative Economics, 80(3), 393-429. Faber, B. (2014) “Trade integration, market size, and industrialization: Evidence from China’s National Trunk Highway System,” The Review of Economic Studies, 18, 1046- 1070. Fernald, J. (1999) Roads to prosperity? Assessing the link between public capital and productivity. American Economic Review, 89 (3), 619–638. Foster, V. and C. Briceño-Garmendia (2010) Africa’s Infrastructure: A Time for Transportation, The World Bank, Africa Development Forum. Freund, C. and N. Rocha (2011) “What Constrains Africa’s Exports?” World Bank Economic Review, 25(3), 361-386. Fujita, M. and J.F. Thisse (2002) Economics of Agglomeration. Cambridge University Press. 31 Garica-López, M.A. (2012) “Urban spatial structure, suburbanization and transportation in Barcelona,” Journal of Urban Economics, 72, 176-190. Garcia-Lopez, M-A, A. Holl, and E. Viladecans-Marsal (2015) “Suburbanization and Highways: when the Romans, the Bourbons and the First Cars still shape Spanish cities,” Journal of Urban Economics, 85, 52-67. Ghani, E., A.G. Goswami, and W.R. Kerr (2012) “Highway to Success: The Impact of the Golden Quadrilateral Project for the Location and Performance of Indian Manufacturing,” NBER Working Paper 18524. Gibbons S., T. Lyytikäinen, H. Overman and R. Sanchis-Guarner (2012) “New Road Infrastructure: the Effects on Firms,” SERC Discussion Paper 117. Gibbons, S. and S. Machin (2005) “Valuing rail access using transport innovations,” Journal of Urban Economics, 57(1), 148–169. Gertler, P.J., M. Gonzalez-Navarro, T. Gracner, and A.D. Rothenberg (2014) “The Role of Road Quality Investments on Economic Activity and Welfare: Evidence from Indonesia’s Highways,” mimeo. Givord, P., C. Grislain-Letrémy, and H. Naegele (2014) “How does fuel taxation impact new car purchases? An evaluation using French consumer-level data,” DIW Discussion Paper 1428. Glaeser, E.L. and M.E. Kahn (2003) “Sprawl and Urban Growth,” NBER Working Paper 9733. Gobillon, L. and H. Selod (2012) “Spatial Mismatch,” in Smith, S., M. Elsinga, L. Fox O’Mahony, O. Seow Eng, S. Wachter, and R. Green (eds), International Encyclopedia of Housing and Home, Vol. 7. Oxford: Elsevier, 2012, 5-12. Gobillon, L. and H. Selod (2014) “Spatial Mismatch, Poverty, and Vulnerable Populations,” in Fischer, M. and P. Nijkamp (eds), Handbook of Regional Science, Springer, 93-107. Gobillon, L., H. Selod, and Y. Zenou (2007) “The Mechanisms of Spatial Mismatch,” Urban Studies, 44(12), 2401-2427. Gobillon, L., T. Magnac, and H. Selod (2012) “Do unemployed workers benefit from enterprise zones? The French experience,” Journal of Public Economics, 96(9-10), 881–892. 32 Gollin, D. and R. Rogerson (2014) “Productivity, transport costs, and subsistence agriculture,” Journal of Development Economics, 107, 38-48. Gonzales-Navarro, M. and C. Quintana-Domeque (2014) “Paving Streets for the Poor: Experimental Analysis of Infrastructure Effects,” mimeo. Graham, D. and S. Glaister (2002) “The Demand for Automobile Fuel: A Survey of Elasticities,” Journal of Transport Economics and Policy, 36, 1-25. Gutberlet, T. (2013) “Railroads and the Regional Concentration of Industry in Germany 1846 to 1882,” mimeo. Gutierrez-i-Puigarnau, E. and J. van Ommeren (2010) Labour Supply and Commuting, Journal of Urban Economics, 68(1), 82-89. Hanson, G.H. (2005) “Market potential, increasing returns and geographic concentration,” Journal of International Economics, 67, 1-24. Harris, C.D. (1954) “The market as a factor in the localization of industry in the United States,” Annals of the Association of American Geographers, 44, 315-348. Head, K. and T. Mayer (2011) “Gravity, market potential and economic development,” Journal of Economic Geography, 11, 281-294. Hsu, W-T. and H. Zhang (2013) “The fundamental law of highway congestion: Evidence from national expressways in Japan,” Journal of Urban Economics, 81, 65–76. Hummels, D. (2007) “Transportation Costs and International Trade in the Second Era of Globalization,” Journal of Economic Perspectives, 21, 131-154. Hymel, K. (2009) “Does Traffic Congestion Reduce Employment Growth?” Journal of Urban Economics, 65(2), 127-135. Ihlanfeldt, K. (2003) Rail Transit and Neighborhood Crime: The Case of Atlanta Georgia, Southern Economic Journal, 70(2), 273-94. Jacoby, H.G. (2000) “Access to markets and the benefits of rural roads,” The Economic Journal, 110(465), 713-737. Jacoby, H.G. and B. Minten, B. (2009) “On measuring the benefits of lower transport costs,” Journal of Development Economics, 89(1), 28-38. Jedwab, R. and A. Moradi (forthcoming) “The Permanent Effects of Transportation Revolutions in Poor Countries: Evidence from Africa,” Review of Economics and Statistics, in press. 33 Kain, J.F. (1968) “Housing Segregation, Negro Employment and Metropolitan Decentralization,” The Quarterly Journal of Economics, 82(2), 175-197. Katz, L.F., J.R. Kling, and J.B. Liebman (2001) “Moving to Opportunity in Boston: Early Results of a Randomized Mobility Experiment,” The Quarterly Journal of Economics, 116(2), 607-654. Kruger, T. and K. Landman (2007) “Crime and Public Transport: Designing a Safer Journey,” Proceedings of the 26th Southern African Transportation Conference (SATC 2007), Pretoria South Africa Kanemoto, Y. (1976) “Cost-benefit analysis and the second-best land use for transportation,” Journal of Urban Economics, 4, 483-503. Khandker, S.R., Z. Bakht and G.B. Koolwal (2009) “The poverty impact of rural roads: evidence from Bangladesh,” Economic Development and Cultural Change, 57(4), 685- 722. Kyeyamwa, H., S. Speelman, G.V. Huylenbroeck, J. Opuda-Asibo, W. Verbeke (2008) “Rising offtake from cattle grazed on natural rangelands in Sub-Saharan Africa: a transaction cost economic approach,” Agricultural Economics, 39, 63-72. Lall, S.V, U. Deichmann and Z. Shalizi (2004) “Agglomeration economies and productivity in Indian industry”, Journal of Development Economics, 73, 2:643-673, 2004. Lall, S. V., T. Munthali, and H. Wang (2009) “Explaining High Transport Costs Within Malawi: Bad Roads Or Lack of Trucking Competition?” Policy Research Working Paper No. 5133, The World Bank, Washington, D.C. Laurance, W.F. and A. Balmford (2013) “A global map for road building” Nature, 495, 308-309. Laurance, W.F., G.R. Clements, S. Sloan, C.S. O’Connell, N.D. Mueller, M. Goosem, O. Venter, D.P. Edwards, B. Phalan, A. Balmford, R.Van Der Ree, and I.B. Arrea (2014) “A global strategy for road building,” Nature, doi. 10.1038/nature13717 Laurance, W.F., M. Goosem, and S.G.W. Laurance (2009) “Impacts of roads and linear clearings on tropical forests,” Trends in Ecology and Evolution, 24(12), 659-669. Marcelo, D. (2013) “Impact evaluation of the Transmilenio infrastructure on street robberies,” mimeo. 34 Mayer, T. and C. Trevien (2015) “The Impacts of Urban Public Transportation: Evidence from the Paris Region,” CEPR Discussion Papers 10494. Michaels, G. (2008) “The effect of trade on the demand for skill: evidence from the interstate highway system,” The Review of Economics and Statistics, 90(4), 683-701. Minten, B. and S. Kyle (1999) “The Effects of Distance and Road Quality on Food Collection, Marketing Margins, and Traders’ Wages: Evidence from the Former Zaire,” Journal of Development Economics, 60, 467-495. Mu, R. and D. van de Walle (2011) “Rural Roads and Local Market Development in Vietnam,” The Journal of Development Studies, 47(5), 709-34. Murphy, K.M., A. Shleifer, and R.W. Vishny (1989) “Industrialization and the Big Push,” The Journal of Political Economy, 97 (5), 1003-1026. Omamo, S.W. (1998) “Transport Costs and Smallholder Cropping Choices: An Application to Siaya District, Kenya,” American Journal of Agricultural Economics, 80(1), 116-123. Parry, I. and G. Timilsina (2010) “How Should Passenger Travel in Mexico City be Priced?” Journal of Urban Economics, 68, 167-182. Parry, I. and G. Timilsina (2012) “Demand Side Instruments to Reduce Road Transportation Externalities in the Greater Cairo Metropolitan Area,” Policy Research Working Paper No. 6083, The World Bank, Washington, DC. Parry, I. and J. Strand (2012) “International Fuel Tax Assessment: An Application to Chile,” Environment and Development Economics, 17(2), 127-144. Peterson, G.E. (2009) Unlocking Land Values to Finance Urban Infrastructure, The World Bank, Washington, D.C. Philips, D. (2014) “Getting to Work: Experimental Evidence on Job Search and Transportation Costs,” Labor Economics, 29, 72–82. Raballand, G., P. Macchi, and C. Petracco (2010). Rural Road Investment Efficiency: Lessons from Burkina Faso, Cameroon, and Uganda. The World Bank, Washington, D.C. Raphael, S. and M. Stoll (2001) “Can Boosting Minority Car-Ownership Rates Narrow Inter-Racial Employment Gaps?” Berkeley Program on Housing and Urban Policy, Working Paper Series qt4k4519pw, Berkeley Program on Housing and Urban Policy. 35 Redding, S.J. and M.A. Turner (2014) “Transportation Costs and the Spatial Organization of Economic Activity,” NBER Working Paper 20235. Regondi, I., G. George, N. and Pillay (2013) “HIV/AIDS in the transport sector of southern Africa: Operational challenges, research gaps and policy recommendations,” Development Southern Africa, 30, (4-5), 616-628. Roberts, M., U. Deichmann, B. Fingleton, T. Shi (2012) “Evaluating China's road to prosperity: A new economic geography approach,” Regional Science and Urban Economics, 42(4), 580-594. Rosenstein-Rodan, P.N. (1943) “Problems of Industrialisation of Eastern and South- Eastern Europe,” The Economic Journal, 53(210/211), 202-211. Rosenthal, S. and W.C. Strange (2012) “Female Entrepreneurship, Agglomeration, and a New Spatial Mismatch,” The Review of Economics and Statistics, 94(3): 764-788. Rospabé, S. and H. Selod (2006) “Does city structure cause unemployment? The case of Cape Town,” in: Bhorat, H. and R. Kanbur (ed) Poverty and policy in post-apartheid South Africa, Chapter 7, HRSC Press, Cape Town, 262-287. Rothenberg, A. (2013) “Transport Infrastructure and Firm Location Choice in Equilibrium: Evidence from Indonesia’s Highways,” mimeo Sanchis-Guarner, R. (2012) “Driving Up Wages: the Effects of Road Construction in Great Britain,” SERC Discussion Paper 120. Selod, H. and Y. Zenou (2006) “City Structure, Labour Discrimination and Job Search. Theory and Policy Implications,” Economic Journal, 116, 1057-1087. Small, K. and K. Van Dender (2006) “Fuel Efficiency and Motor Vehicle Travel: The Declining Rebound Effect,” Energy Journal, 28: 25-52. Sperling, D. and E. Claussen (2004) “Motorizing the Developing World,” Access, 24, 10- 15. Storeygard, A. (2014) “Farther on down the road: transport costs, trade and urban growth in sub-Saharan Africa,” Revise and Resubmit at the Review of Economic Studies. Suzuki, H., Cervero, R. and K. Iuchi (2013) Transforming Cities with Transit: Transit and Land Use Integration for Sustainable Urban Development, The World Bank. Teravaninthorn, S. and G. Raballand (2008) Transport Prices and Costs in Africa: a Review of the Main International Corridors, World Bank, Washington, D.C. 36 Timilsina, G.R. and H.B Dulal (2011) “Urban road transportation externalities: costs and choice of policy instruments,” The World Bank Research Observer, 26(1), 162-191. van de Walle, D. (2002) “Choosing rural road investments to help reduce poverty,” World Development, 30(4), 575-589. van de Walle, D. (2009). Impact evaluation of rural road projects. Journal of Development Effectiveness, 1(1), 15-36. van Ommeren, J. and E. Gutierrez-i-Puigarnau (2011) “Are workers with a long commute less productive? An empirical analysis of absenteeism,” Regional Science and Urban Economics, 41(1), 1-8. Volpe Martincus, C., J. Carballo, and A. Cusolito (2014) “Routes, Exports, and Employment in Developing Countries: Following the Trace of the Inca Roads,” mimeo. Weinberg, B. (2000) “Black residential centralization and the spatial mismatch hypothesis,” Journal of Urban Economics, 48, 110-134. Wheaton, W.C. (1998) “Land use and density in cities with congestion,” Journal of Urban Economics, 43, 258-272. World Bank (2014) The Economic Impact of the 2014 Ebola Epidemic: Short- and Medium-Term Estimates for West Africa. Zax, J. and J. Kain (1996) “Moving to the suburbs: do relocating companies leave their Black employees behind?” Journal of Labor Economics, 14, 472-504. 37 Table 1 –Focus and identification strategy in select transport research Study General Focus Outcome of Treatment or Identification Source of Exogenous Variation or IV Interest Explanatory Problem Variable of Interest Ahlfeldt and Impact of metro and Changes in land Travel time to Endogeneity of Instrument: distance to straight-line Wendland railways on decentralization value gradient CBD metro and railway counterfactual transport network (2011) (Berlin, Germany, 1890- placement between the CBD and neighboring 1936) municipalities Anderson Impact of public transit on Highway delay Public transit Endogeneity of Regression discontinuity using public (2014) traffic congestion (Los ridership public transit transit workers’ strike Angeles, US) provision Atack and Impact of railroad access on Cultivated land and Dummy for access Endogenous Instrument: Straight-line hypothetical Margo (2011) agricultural land use and land values to railroad network railway placement network connecting selected nodal cities value (US, 1850-1860) (identified for transport investment in the 1820s) Atack et al Impact of railroad access on Urbanization rate at Dummy for access Endogenous Instrument: Straight-line hypothetical (2010) urbanization (US, 1850- the county level to railroad network railway placement network connecting selected nodal cities 1860) (identified for transport investment in the 1820s) Banerjee et al Impact of transport Level and growth of Proximity to Endogeneity of Argue exogeneity of regressor by using (2012) accessibility on economic local GPD per capita transport infrastructure distance to straight line network development (China) infrastructure placement (linking 19th century historical cities and ports) Baum-Snow Impact of highways on Central city Number of radial Endogenous Instrument: Number of rays in historical (2007a) decentralization in cities population highway (“rays”) highway placement highway plan (1947 National System of (US, 1950-1990) Interstate Highways plan) Baum-Snow Impact of highways on Commuting patterns Number of radial Endogenous Instrument: Number of rays in historical (2010) decentralization in cities highway (“rays”) highway placement highway plan (1947 National System of (US, 1960-2000) Interstate Highways plan) Burgess and Impact of railways on Local prices, total Dummy for Endogenous Claim that network built for military Donaldson famine (India, 1861-1930) output, and District penetrated railway placement reasons is exogenous (2012) mortality by rail Casaburi et al Impact of road Crop prices Road improvement Endogeneity of Regression discontinuity using (2013) improvement on trade selection of roads assignment rule for road improvement (Sierra Leone) for rehabilitation 38 Study General Focus Outcome of Treatment or Identification Source of Exogenous Variation or IV Interest Explanatory Problem Variable of Interest Chandra and Impact of Interstate Real earnings by Dummy for Endogenous Focus on “inconsequential” non- Thompson highways on economic industry highway going highway placement metropolitan areas (2000) activity in rural areas (US) through rural counties Chen and Impact of urban rail transit Carbon monoxide Metro ridership Endogeneity of Regression discontinuity using opening Whalley (2012) on air pollution (Taipei, emission metro station of new metro Taiwan, China) placement Datta (2012) Impact of highways on firm Stock of input Dummy for area Endogeneity of Focus on inconsequential areas and use efficiency (India) inventories benefitting from highway network of difference-in-difference highway network improvement improvement (more or better infrastructure) or distance to highway network Donaldson, Impact of railways on trade Trade costs, Distance (speed Endogenous Claim that network built for military forthcoming (India, 1870-1930) interregional price along least cost railway placement reasons is exogenous gaps, trade flows, path) income Donaldson and Economic impact of Agricultural land Market access Endogenous Instruments: Hornbeck railways (US) value index railway placement - counterfactual market access measure (2013) based on fixed initial population figures - waterway only market access Dorosh et al Impact of roads on Crop production Travel time to Endogeneity of Instrument: Terrain roughness in (2012) agriculture (Sub-Saharan nearest city road placement production area (difference between max Africa) and min elevation) Duranton (2014) Impact of roads on trade Weight and value of Road travel time or Endogeneity of Instrument: (Colombia) goods distance roads - colonial route (1938)     39 Study General Focus Outcome of Treatment or Identification Source of Exogenous Variation or IV Interest Explanatory Problem Variable of Interest Duranton and Impact of road supply on Traffic (vehicles- Lane-kilometers of Endogeneity of Instruments: Turner (2011) congestion (US) kilometers traveled) roads road placement - historical interstate highway plan (1947) - historical railroad routes (1898) - exploration routes (1835-1850) Duranton et al Impact of interstate Weight and value of Length of interstate Endogeneity of Instruments: (2014) highways on trade between goods highways between road placement - historical interstate highway plan cities (US) two MSAs (1947) - historical railroad routes (1898) - exploration routes (1835-1850) Emran and Hou Impact of domestic and Per capita Travel distance by Endogenous - Instrument: straight-line distance from (2013) international market access consumption road or railway to placement of roads village to nearest coastline and navigable on rural poverty (China) domestic business and railways river center or - Identification by heteroskedasticity international seaport Faber (2014) Impact of highways on Value added in Proximity to Endogenous Instruments: trade, market size, and manufacturing and highway (dummy highway placement Hypothetical road networks between industrialization (China) agriculture; local for less than 10km, nodes (cost-minimizing or straight line) government or exact distance to revenue; GDP; nearest highway population segment) Garcia-López Impact of changes in Change in Change in distance Endogenous Instruments: straight-line distance to the (2012) transportation on location population density to nearest highway placement of nearest network (Roman roads built for patterns within cities at census-tract level ramp, and distance highways and military purposes, pre 19th century roads (Barcelona, Spain, 1991- to nearest railroad railways designed to improve communications; 2006) station and 19th century railroad lines) Garcia-López et Impact of highways on Change in central Change in number Endogenous Instruments: Rays or kilometers from al (2015) decentralization in cities city population and of highway rays in highway placement historic roads (either Roman roads built (Spain) suburban population central cities for military purposes, or Bourbon roads growth built for political reasons).     40 Study General Focus Outcome of Treatment or Identification Source of Exogenous Variation or IV Interest Explanatory Problem Variable of Interest Ghani, Impact of highways on Entry of firms, labor Distance of non- Endogenous Focus on “inconsequential” non-nodal Goswami, and economic activity (India) productivity, and nodal district to highway placement districts Kerr (2012) TFP highway system Gibbons and Impact of transport Change in property Change in distance Endogeneity of Diffs in diffs within differentially treated Machin (2005) accessibility on land prices values to nearest railway station placement areas: within 2km and further away (London, UK) station Gibbons et al Impact of road construction Number of firms Local index of job Endogeneity of Comparison of effects at various (2012) on economic activity (UK) and employment accessibility to jobs road construction distances within a narrow band and of worker (avoiding selection issues in the location comparison) Instrument: counterfactual accessibility index in the absence of relocation Hsu and Zhang Impact of road capacity on Traffic (vehicles- Lane-kilometers of Endogeneity of Instrument: (2013) traffic (Japan) kilometers traveled) roads road placement - historical expressway extension plan (1987) Hymel (2009) Effects of congestion on Employment growth Annual aggregate Endogeneity due to Instruments: employment growth (US, in US metropolitan time lost in traffic reverse causality - Number of radial road-miles per capita 1982-2003) areas at metropolitan (employment in historical highway plan (1947 level growth causing National System of Interstate Highways congestion) plan) - Number of transportation committee members in the House of Representatives (influencing congestion through road construction in their constituencies) Jedwab and Railroad construction on Cocoa production, Dummy for Endogeneity of Instrument: distance from straight lines Moradi economic development population, urban proximity to railway placement between two main seaports (forthcoming) (Ghana, 1891-2000) growth railway Placebo regressions     41 Study General Focus Outcome of Treatment or Identification Source of Exogenous Variation or IV Interest Explanatory Problem Variable of Interest Khandker and Impacts of rural roads on Per capita Rural road Endogeneity of First differences Koolwal (2011) household outcomes expenditure, school improvement road placement and Instruments: lagged outcome variables (Bangladesh) enrollment, non- improvement agricultural employment, consumption prices Khandker et al Impacts of rural roads on Per capita Rural road Endogeneity of First difference controlling for initial (2009) household outcomes expenditure, school improvement road placement and conditions (Bangladesh) enrollment, wages, improvement hours worked, input and output prices Mayer and Impact of public transport Local share of firms Construction of Endogeneity of Focus on non-consequential stations Trevien (2003) extension on local economic (out of metropolitan suburban express station placement between “new towns” and metropolitan activity (Paris, France) area total) railway station core Michaels (2008) Impact of interstate Earnings in the Dummy for Endogenous Instruments: highway on labor trucking and highway going highway placement - Dummy whether a route was planned market/trade (US) warehousing through rural for in 1944 industries counties - Angle with closest city (correlated with (indicative of trade probability of connection to highway activity) grid) Sanchis-Guarner Impact of road construction Wages and hours Local index of job Endogeneity of Comparison of effects at various (2012) on labor market outcomes worked accessibility to jobs road construction distances within a narrow band (UK) and of worker (avoiding selection issues in the location comparison) Instrument: counterfactual accessibility index in the absence of relocation Storeygard Impact of transport cost on City light output Price of oil Endogenous road Argues exogeneity of oil prices (2014) trade and urban growth interacted with placement (sub-Saharan Africa) distance along road to primate city     42 Study General Focus Outcome of Treatment or Identification Source of Exogenous Variation or IV Interest Explanatory Problem Variable of Interest Van de Walle Fungibility of aid money for Length of road Aid money for road Endogenous Difference-in-difference matching and Mu (2007) road construction and improvements or improvement assignment of aid improvement (Vietnam) new roads money to project areas Volpe Impact of road expansion on Exports and Road infrastructure Endogeneity of Difference-in-difference. Martincus et al exports and employment employment expansion program road placement Instrument: distance to and along Inca (2014) (Peru) Road   43