What drives sub-way costs - imitate, don't innovate!

Alex Tabarrok has a very good post linking to the work of Alon Levy in compiling and analysing the data on construction costs of over 205 metro rail projects in 40 countries. The data set is very rich and impressive. Sample this,

New York is planning to have spent around $35 billion between 2005 and 2030 on subway and commuter rail expansion. But it’s only getting 15 km of new tunnel! Paris is spending a similar amount over the same period: Euro 40 billion, for a total of 228 km, 187 underground. Madrid, a much smaller city, spent Euro10 billion in 1995-2015 on 234 km, around 180 underground.

This is also perhaps a great example of how academic research can add significant value to real world issues. Levy's main point is that lack of awareness and siloed thinking in structuring contracts and choice of construction techniques is the main contributor to the exorbitantly high subway costs in New York compared to other countries. It is an example of how US can learn to imitate from Europeans and Koreans.

This is interesting,

Levy is to be lauded for his pioneering work on this issue yet isn’t it weird that a Patreon supported blogger has done the best work on comparative construction costs mostly using data from newspapers and trade publications? New York plans to spend billions on railway and subway expansion. If better research could cut construction costs by 1%, it would be worth spending tens of millions on that research. So why doesn’t the MTA embed accountants with every major project in the world and get to the bottom of this cost disease? (See previous point). Perhaps the greatest value of Levy’s work is in drawing attention to the issue so that the public gets mad enough about excess costs to get politicians to put pressure on agencies like the MTA.

This is something for perhaps the World Bank to take forward. In particular, comparing across emerging technologies, and also life-cycle costs of various systems can be very useful decision-support for governments planning new metro systems. 

The main takeaways from the study

1. Geographic and geological factors are not responsible for the wide variations in project costs of sub-way systems. 

2. Station costs are a big factor in the variations. 

3. In terms of total cost, cut-and-cover method is cheaper than bored tunnel with cut-and-cover stations, by a factor of 1.5 to 2, which in turn is cheaper than bored tunnel with mined stations. 

4. The cost of mined stations is often 4-8 times that of cut-and-cover stations. So the best approach is shallow cut-and-cover construction, disrupting the street for a period of time, and refrain from mining except at under-crossings. Further, avoid mezzanines, with all circulation, including fare barriers to be on the platform level or at street level. And have an island platform, ideally accessed from a street median, to avoid duplicating elevators, stairs etc. Finally standardised station designs.

5. City centre tunnels cost more because cut-and-cover is not possible and because of number of crossing of older lines.

6. Contracts should avoid design-build approach and should be flexible and strictly separate design and construction. They should also not be lump-sum construction contracts, and the bills should be itemised to limit litigation during renegotiations and changes. 

7. Contract awards should be quality and cost-based, rather than purely cost-based ones, awarded to the lowest bidder. Madrid used technical score (50%), speed (20%) and cost (30%)

8. Finally, instead of relying on consultants, use an in-house team to oversee the design and project management. The in-house team costs 25-33% less than consultants and are also likely more effective and constructive. 

The Madrid metro is a good example of a project which incorporated these principles into its construction, and one of the cheapest in the world.

Update 1 (5.01.2020)

This is a good Citylab article by Alon Levy explaining why it is so expensive to build in New York.