Exploring Alternative Methods to Estimate the Agglomeration Economies of Urban Rail Projects

Exploring Alternative Methods to Estimate the Agglomeration Economies of Urban Rail Projects


E Gwee, Monash University/Land Transport Authority of Singapore, AU/SG; G Currie, J Madden, Monash University, AU; J Stanley, Sydney University, AU


To test robustness of agglomeration economies of urban rail, new estimation methods are developed including secondary data, employer surveys, travel modelling & a CGE model. Results are applied to a case study & compared to conventional method.


Agglomeration economies in production are positive impacts on productivity associated with an increasing concentration of firms or employment. In central cities these have been associated with major urban commuter rail projects which are said to increase agglomeration of this type and achieve agglomeration benefits. Where economic benefits associated with agglomeration economies are included in rail project evaluation (e.g. UK Cross Rail), they have substantially improved the performance of large rail projects compared to the use of conventional economic benefit methods. Yet much uncertainty surrounds the use of economic benefits associated urban agglomeration effects in transport project impact appraisal. Only a few countries include them in their national guidelines and methods of estimation have been limited.

This paper aims to better inform the debate regarding the validity of urban rail agglomeration benefits by developing new methods to estimate their value and comparing the benefits they provide with conventional methods. The various approaches are applied to a case study, the CBD of Melbourne, Australia where the agglomeration disbenefits of not increasing rail capacity in a city with an overloaded rail system are estimated.

The paper presents the background to the theory of agglomeration and describes conventional approaches to their estimation. The study case study context is also described since this is the basis for the implementation of modelling estimates. The new estimation approaches are described which include the use of secondary data of city density and rail provision, transport modelling using commute trip suppression as a surrogate to employment impacts, a survey of employers to establish their view on employment suppression impacts of traffic congestion and the adoption of a computable General Equilibrium Model (ORANI) to estimate economic impacts of productivity shocks associated with rail capacity suppression.

A series of short term and long term impact scenarios of the economic disbenefits of not providing higher capacity rail services is then developed using the alternative methodologies. Two measures are used to estimate agglomeration impacts; the potential CBD employment suppression and the change in CBD's accessibility. The agglomeration effects are estimated based on the impact on gross regional product of a 10% rail capacity constraint.

Results show a surprisingly narrow range of employment suppression impacts resulting from alternative estimation methods. A 10% rail capacity restraint results in a 0.69%-1.07% employment suppression using travel demand modelling, a 0.84%-1.1% range from a survey of CBD employers and 1.06% range using analysis of secondary city data. The lower bound of short range estimates for all methods including those based on current methodologies (such as defined in the UK national guidelines) all provide remarkably close results (between -0.08% of GRP to -0.09% of GRP). However high range estimates for short run effects are much wider. Some (-0.13 to -0.16% of GRP) for employment suppression measures which are than higher than those estimated from conventional methods and those using the CGE modelling approach.

Long range estimates were far more diverse. CBD proximity based methods provided the lowest long term measures (-0.08 to -0.09 of GRP). These were from the conventional method and also the net effects of the CGE modelling. The latter included long term redistributional impacts of employment changes on the regional economy. Higher long term impacts results from the employment suppression based approaches (-0.13 to -0.25%) notably for the high end estimated based on secondary city data.

A major finding of the study is that there is much consistency in employment impacts of transport capacity suppression on employment from alternative methods. In the short term there is also consistency in results regardless of alternative methods of estimation. However long term estimation methods have more diverse results. There are advantages in modelling these impacts using CGE model since it can account for flow-on effects of employment changes in one region to another. This suggests that direct employment changes in one area are compensated for in another. It suggests short and long run effects are about similar after long range flow-on effects are accounted for.


Association for European Transport