This is the first version of our rolling stock costs database, to complement our primary urban rail construction costs database, excluding rolling stock. The new database looks at the procurement costs of metro trains, commuter rail electric multiple units (EMUs), and some light rail stock. The patterns we observe are different from those of the infrastructure cost database.
The primary cost metric we have chosen to use is the purchase cost per meter of train length. This is motivated by seeing that the figures, so normalized, vary relatively little—less than the cost per kilometer of urban rail infrastructure construction. Moreover, across the types of train sets we have investigated, there is no apparent pattern in other kinds of capacity calculation: in particular, wider commuter trains do not cost more per meter of length than narrower metro or light rail vehicles, and if anything they tend to cost less. We do, however, include capacity numbers measured in seats where these are available, for eventual comparison of single- and double-deck mainline trains.
The data largely studies three separate supply chains: the Chinese supply chain, comprising largely metros but also a handful of medium-speed commuter trains with a top speed of 160-200 km/h; the European one, comprising streetcars or light rail vehicles, metros, and commuter trains; and the North American one, comprising light rail vehicles, metros, and some commuter rail EMUs. Purchases within each of these are largely disjoint, and it is best to compare them in two steps: within each supply chain, it’s fruitful to make comparisons between cities, types of vehicles, or even individual orders; then the three supply chains can be compared as aggregated averages.