High Speed 2 is (Partly) Canceled Due to High Costs
It’s not yet officially confirmed, but Prime Minister Rishi Sunak will formally announce that High Speed 2 will be paused north of Birmingham. All media reporting on this issue – BBC, Reuters, Sky, Telegraph – centers the issue of costs; the Telegraph just ran an op-ed supporting the curtailment on grounds of fiscal prudence.
I can’t tell you how the costs compare with the benefits, but the costs, as compared with other costs, really are extremely high. The Telegraph op-ed has a graph with how real costs have risen over time (other media reporting conflates cost overruns with inflation), which pegs current costs, with the leg to Manchester still in there, as ranging from about £85 billion to £112 billion in 2022 prices, for a full network of (I believe) 530 km. In PPP terms, this is $230-310 million/km, which is typical of subways in low-to-medium-cost countries (and somewhat less than half as much as a London Undeground extension). The total cost in 2022 terms of all high-speed lines opened to date in France and Germany combined is about the same as the low end of the range for High Speed 2.
I bring this up not to complain about high costs – I’ve done this in Britain many times – but to point out that costs matter. The ability of a country or city to build useful infrastructure really does depend on cost, and allowing costs to explode in order to buy off specific constituencies, out of poor engineering, or out of indifference to good project delivery practices means less stuff can be built.
Britain, unfortunately, has done all three. High Speed 2 is full of scope creep designed to buy off groups – namely, there is a lot of gratuious tunneling in the London-Birmingham first phase, the one that isn’t being scrapped. The terrain is flat by French or German standards, but the people living in the rural areas northwest of London are wealthy and NIMBY and complained and so they got their tunnels, which at this point are so advanced in construction that it’s not possible to descope them.
Then there are questionable engineering decisions, like the truly massive urban stations. The line was planned with a massive addition to Euston Station, which has since been descoped (I blogged it when it was still uncertain, but it was later confirmed); the current plan seems to be to dump passengers at Old Oak Common, at an Elizabeth line station somewhat outside Central London. It’s possible to connect to Euston with some very good operational discipline, but that requires imitating some specific Shinkansen operations that aren’t used anywhere in Europe, because the surplus of tracks at the Parisian terminals is so great it’s not needed there, and nowhere else in Europe is there such high single-city ridership.
And then there is poor project delivery, and here, the Tories themselves are partly to blame. They love the privatization of the state to massive consultancies. As I keep saying about the history of London Underground construction costs, the history doesn’t prove in any way that it’s Margaret Thatcher’s fault, but it sure is consistent with that hypothesis – costs were rising even before she came to power, but the real explosion happened between the 1970s (with the opening of the Jubilee line at 2022 PPP $172 million/km) and the 1990s (with the opening of the Jubilee line extension at $570 million/km).
Pedestrian Observations 
The CPRE wanted HS2 to follow the M1 and the head of HS1 wanted it to follow the M40. In either case you would have had less public opposition as it would follow an existing ugly and noisy piece of infrastructure.
And if the route in between was truly necessary HS2 should have added at least one station in Buckinghamshire as well as spending significantly more money on community projects. I don’t believe they are funding any community projects in Aylesbury.
Spending £100k a parish on the community for the parishes close to the line really doesn’t add much project cost – but means all your most powerful local opponents are on side – and means they are arguing what to spend the money on.
The M40 through the Chilterns is not suitable for a 150km/h railway to run alongside it, let alone one designed for at least twice that. A simple drive on the hilly and twisty route route would make that clear, but here’s a quote from the 2009 route selection:
“The M40 passes through comparatively hilly terrain, and would require much of the route to be in tunnel or viaduct. Given the geometry of the M40, little of the route would have been sufficiently close to the motorway to be described as within its corridor. This formed a route that was not pursued because it was found to perform least well in comparison with the other options, with major adverse impacts on landscape, biodiversity and water resources.”
Looking at the options from the 2012 route selection report (which was post initial-consultation where people had asked the same questions as you do about putting the railway next to a motorway instead of taking what is pretty much the most direct and flattest route between London and Birmingham that the railway is taking), the motorway routes (which having been rejected early before as not practical, had been reduced to 186mph design speed to try and get the railway to sit alongside the motorway more) had more tunnelling than the selected route due to passing alongside major population centres:
“As with the M40 corridor, a surface alignment along the M1 route would encounter a much greater number of major population centres than the consultation route would. This would result in unacceptable impacts on communities through major demolitions, severance and noise impacts and therefore this route would require substantial sections of tunnelling. This makes it a substantially more complicated and expensive option than the consultation route.”
The M1 route had twice the number of residential demolitions as the consultation route – despite the tunnels, and moving the line away from the motorway in places (and so isolating 14 communities, stuck between railway and motorway with limited access across both). It would have tunnelled under about 18 times as many dwellings as the consultation route (and that was before that was refined to not tunnel under a load of listed houses in Old Amersham).
The M1 route was found to have similar landscape impacts as the consultation route through the heart of the AONB, though impacted less on protected sites as they were all designated after the M1 made them less worth protecting! The M40 route was deemed to only have ‘slight’ environmental benefits vs the pre-existing route – and most of that was due to needing to be in tunnel for far more of its time across the Chilterns than the selected route.
Going via the motorways would have cost more than, had very little environmental benefit over, and taken longer to travel between London and Birmingham than, the Misbourne route. Opposition would have been worse.
CPRE would still have been fuming at the environmental cost – the M1 route idea was so they could look like they weren’t just opposed to the whole idea.
The HS1 guy had sour grapes over not getting to take part in the design HS2’s route – and then also their being a ‘mistakes of HS1’s route that we don’t want to make again’ section in one of the documents on route selection – he just wants to attack what HS2 do in an attempt to get some love for his work that HS2 wasn’t impressed with. Even if they had picked the M40 route, then he’d have said the M1 was better.
The bulk of the opposition would have hated a slower and more expensive route – given their straw man battle cry of “£100bn to get to Birmingham 15 minutes quicker”. And while the current route goes through a great bastion of NIMBYism, annoying an order of magnitude more people simply as they live near to the line would have a lot more bite than the bark that the Amersham/Missenden set gave.
Interesting stuff!
A problem is how much credibility should be given to those studies.
Further, a lot seems to rest on the notion of building it on the ground. Why? There seems a lot of resistance to making it elevated most of the route — to remove interference with the motorway or anything else (like turning villages into exclaves! or requiring far more house demolitions!). Of course there will be more cost but it looks preferable to the alternative, as well as giving intrinsic flexibility in coping with gradient changes (Alon always points out that HSR can cope with much higher climbing grades than regular rail, and that they used this property to keep the costs down on Paris-Lyon).
Again, the higher environmental impact assumes it is at ground level beside the motorway rather than above it. The Beijing–Shanghai HSR has 87% of its 1140km elevated with the world’s #1 (164km, almost London to Birmingham!), #2 (114km) bridges/viaducts. Estimated cost $32b at $28 per km.
Of course the environmental impact of HSR is grotesquely exaggerated by the irresponsible and ill-informed who are essentially BANANAs. I don’t know if this is strictly correct but the Wiki on LGV-Sud-Est says “The line has a surface area of 16 km2 (6.2 sq mi)—in comparison Charles de Gaulle Airport occupies 32 km2 (12 sq mi) …”.
Was the M40 or M1 study a genuine one or a Clayton’s report? How would it look in retrospect?
As with the M40 corridor, a surface alignment along the M1 route would encounter a much greater number of major population centres than the consultation route would.
That encountering greater numbers of major population centres was deemed as a purely negative point without any recognition of the obvious benefits of providing additional cities with HSR seems like a fundamental weakness of their analysis.
Encountering a population center you don’t stop at is bad, not good.
And no HSR system in the world stops at “population centers” of this size and spacing (size 50k-300k, spacing every ~20km) – they would be served by legacy rail, which would be more frequent with long distance trains off the tracks.
The French TGV definitely stops at towns of 50k.
And you would get millions of rides out of a HS2 stop in Buckinghamshire for sure. I cant really imagine less than 2 million.
I consider 50km the be the minimum stop spacing on a HSR line, and 100km ideal. Which means I will sometimes skip a population of 300k, while elsewhere some town of 2k gets a station. Even though nobody would use the station in a tiny town, it gives me a good story for people in rural areas or those visiting them. 100km station spacing means you can run hourly trains overnight with timed overtakes closing one track (only possible if you have built your tracks with enough separation that someone can safely work on the closed track!). Or on marginal routes that is a good place to do timed overtakes and you can save money by only building on track.
Of course if you have a population of 300k only 20km from a station that implies a very dense population and I expect they have great local transit options to get to the station.
“no HSR system in the world stops at ‘population centers’ of this size and spacing (size 50k-300k, spacing every ~20km)”
The Shinkansen very much fits that description.
Looking at the Hokuriku Shinkansen:
– Annaka Haruna: 18.5 km from previous station, municipal population 53,000
– Karuizawa: 23 km, pop. 20,000
– Sakudaira: 18 km, pop. 138,000 in a twin-city catchment area
– Ueda: 25 km, pop. 151,000
Or the Kyushu Shinkansen:
– Kurume: 7 km from previous station, pop. 300,000
– Chikugo-Funagoya: 16 km, pop. 48,500
– Shin-Omuta: 18 km, pop. 106,000
– Shin-Tamana: 21 km, pop. 62,000
Or an older example in the Tokaido Shinkansen:
– Atami: 21 km from previous station, pop. 32,000
– Mishima: 16 km, pop. 105,000
– Shin-Fuji: 26 km, pop. 241,000
A route following roughly the M1 could have a stop in either Milton Keynes (pop. 290k) or Northampton (pop. 240k) — or both for that matter. From what I can tell, the 90 km distance from Milton Keynes to Birmingham currently takes an hour or more by train. Northhampton is 20 km closer and still takes that long as well. HSR would be an enormous boon for these cities without requiring much deviation from the shortest path between Birmingham and London.
In Germany two small cities immediately spring to mind that are ICE stops: Hildesheim and Göttingen. Both have populations of only around 100k and are only about 40 km from much larger cities along the route (Hanover and Kassel, respectively). There are quite a few others than these two examples.
Milton Keynes and Northampton already have good legacy service that would become even better when the tracks are free of long distance trains. No need for HSR in addition.
They aren’t going to improve the train services to Milton Keynes much with the construction of HS2, plus the main people affected are Aylesbury for whom Milton Keynes isn’t super close anyway.
Fundamentally across the Chilterns there are three main objections in no particular order, the cost, the lack of station and the route.
Now sure you are never going to be able to make everyone happy with the route, but at least with a station there are some benefits from the project – even if the route is one’s main issue. Plus you change the views of the community as a whole from some people being against and a bunch of people not caring either way, to a more balanced viewpoint.
And if you live in Aylesbury and the station is at Calvert on East West rail, well that is closer than Milton Keynes or Oxford so that’s an advantage of the route they have picked.
They let people that have origins or destinations or both, outside of the municipality, use the station.
You have to consider the catchment area.
Deciding “every 50 km” is good when when you are giving SimCity specifications. The real world won’t conform. And example I can think of: Amtrak is not high speed rail. Hudson N.Y. has a tiny population in a county that doesn’t have much either. In fiscal year 2019 it was the third busiest Amtrak station in New York State. I speculate on why but that’s not relevant at the moment. The real world won’t conform to ideals.
Average station spacing on the Shinkansen from Tokyo to Hakata is 31km, this ranges from a low of 6.8km between Tokyto-Shinagawa to a high of 68.1km Maibara-Kyoto.
The express stops an average of every 106km, although this is distorted by multiple closely spaced stops around Tokyo, Kyoto-Osaka-Kobe, and Fukuoka. Spacing ranges up to 317.5km Yokohama-Nagoya and 192km Hiroshima-Kokura.
The limited stops an average of every 66km, up to 123km Hiroshima-Yamagachi and 134km Nagoya-Kyoto.
Britain is not coastal Japan, and the goal isn’t to replicate a distance but a pattern, however along the M1 similar service would have looked like:
Express: Euston-Old Oak Common-Birmingham
Limited: Eus.-OOC-Milton Keynes-Birm.
Regional: Eus.-OOC-Luton-Milton K-Northampton-Coventry-Bir. Airport-Birm.
@henrymiller74, the point about being able to use only one track is still useful in emergencies – even if not particularly for maintenance.
If you have a suspected stroke 5 minutes out of Old Oak Common being able to get to an ambulance after 15 minutes at a station isn’t terrible frankly – probably gets the person into an ambulance faster than anything else.
And if the medical people take longer than expected you don’t risk other passengers getting impatient and walking on the track because you stopped in the middle of nowhere.
Britain is not coastal Japan
Neither is anyplace else. My evaluation of that may be hopelessly out of date.
Station spacing is a compromise. If acceleration was infinite to top speed and total stop time (for people to get on/off) was 1 second at each station I’d stop every 1km. However in the real world acceleration and time stopped at a station takes time. While you can maybe get this down to 15 seconds on a metro, most HSR passengers have bags, and the trains are going to choose to have less doors on the car and more seats. All this means that a HSR train is typically stopped longer in stations than a much slower metro. More stops are bad for anyone who doesn’t want to get on/off at that station.
Sometimes you decide that the real world just isn’t ideal and so you have stations closer than ideal. You must respect the real world which has poor local transit connections to some places, difficult geography, weird politics, and other such things. I’m going to stand by 50km stop spacing in dense metro areas, and 100km in more rural areas is ideal. However in the real world you will never see exactly those numbers. In the real world you place stations in the middle of your biggest, densest cities, then (respecting geography and politics) draw the line on a map, place stations where politics require it, then add in other stations where there is more than 125-200km between stations at whatever town is in the middle.
(Rescued from spamfilter.)
Upon quick reflection, HS2 probably shouldn’t have a Limited service pattern, like the Hikari/Sakura in Japan. The main Shinkansen route is a single line, and although Tokyo is clearly dominant, the size of Osaka/Kyoto makes it largely double ended, even to some extent southwestern portion, with Fukuoka larger than anything between it and Osaka.
In Britain, however, things are undeniably one-sided with regards to London, the route needs to branch (at least to Liverpool, Glasgow, Edinburgh, Newcastle, maybe to Sheffield as well), and cities get smaller the farther from the center (Birm. > Manchester/Leeds > Glasgow/Newcastle). Service should thus be cascading expresses: at its far origin a train runs regional to all stops until the next major city, then it runs express. So trains starting in Scotland stop at Carlisle, Preston, etc. until Manchester, then only stop at Brimingham. Trains starting in Manchester stop for (not necessarily in) Crewe, Stoke, etc. then run express from Birm. Newcastle trains would make stops until Leeds, then only again at Birm. Trains starting at Leeds would stop Sheffield, etc. until Birm., then express.
In a hypothetical world with an M1 line, there would be an hourly train London to Birmingham stopping at Coventry, Luton, etc., pulling over frequently to let all the trains from farther away pass at speed.
That’s complicated. Most railfans don’t do complicated.
The new high speed line between Hamburg and Hannover is supposed to run next to the highway A7 and NIMBYs are screaming about. It’s very hard or impossible to accommodate NIMBYs. So you mainly have to ignore them.
Are the rich suburbs getting a station? Or some other quid-pro-quo?
Because if not there’s your problem.
@Seb: “It’s very hard or impossible to accommodate NIMBYs. So you mainly have to ignore them.”
Yes, but it is much easier for politicians to appear reasonable and to convince all the others that the freeway route is the least disruptive. It is a winnable PR war. But in the UK if you propose a completely new ROW that touches (barely) some trees, all NIMBYs around the country go BANANAs.
@Matthew Hutton: “Are the rich suburbs getting a station? Or some other quid-pro-quo?”
I don’t live in the UK anymore but from everything I read, the real case for HS2 has not been properly explained, as difficult as that is in todays age with everyone obsessed with TikTok etc. It is to provide a separate route to take the main pax load between the major cities. It cannot be squeezed into existing rail infrastructure. This will take pressure off the existing mainlines which desperately need maintenance and upgrading which currently requires a lot of closure (mostly during weekends) and bustitution; and is not making much headway. Given the new route’s purpose there is no point in having other stops en route and it makes sense to make it HSR (since in any other country there is not much cost difference when constructing new rail; there is an idiot focus on saving 20 minutes or whatever and it was never about that). It is win-win on all fronts because total capacity is increased hugely and minor cities will get an improved service, and yes new stops may be possible (and wouldn’t cost much, and wouldn’t disrupt main city-to-city service).
Lots of so-called smart people including those PPEs like Simon Jenkins and George Monbiot (that Borners believes I lurve) have never understood this, or refuse to acknowledge it for their own reasons. It’s reached Brexit levels of divisiveness and miscomprehension, exacerbated by the absurd carpetbagging by those who can (in the UK that invariably means Lord and Lady Muck).
HSR generally doesn’t stop in suburbs, it stops in the city center. Maybe for geographical reasons you stop in suburb on the edge before continuing to the next city, but that is a compromise that is bad for the city in question. You cannot appease NIMBY with HSR stations in general as all those extra stops add enough extra time that the line isn’t HSR anymore. You can maybe give the NIMBYs a station on a different line that gets express service to where the HSR station is (this is probably a good idea but of course you have to look at the city system as a whole)
Yeah but then all you are doing is encouraging people to go by car instead of by train.
Once you get beyond the tube the official time from (say) Great Missenden to Manchester is at least 3h40 and can be considerably longer. Even with the speed up with HS2 if built fully to Manchester and it will still be quicker to drive.
And adding 4 minutes to journey times by having an extra stop isn’t a big deal in the scale of things.
Do agree that service on another line to connect with the HSR line out of the city centre is also a decent option to avoid having too many extra stops.
Adding one stop isn’t significant, but you have a lot of NIMBYs who all want a stop close. HSR won’t even be up to full speed before it has to slow down for the next stop, and that adds up to making the whole line too slow to be called HSR.
You could still have one stop in Buckinghamshire and one at Coventry with every other train stopping at either for HS2 and one each in the Hamburg and Hannover suburbs with every other train stopping at one for example.
These things don’t need to get silly.
As far as i can tell nobody makes a switch that allows HSR to trains to pass at full speed. So while you save the time stopped in the station your train would still have to slow down a lot around the stations you don’t stop at.
“nobody makes a switch that allows HSR to trains to pass at full speed […] your train would still have to slow down a lot around the stations you don’t stop at”
That’s a rather bizarre claim when Shinkansen, THSR, and CRH services (among others) regularly pass through minor stations at full line speed, running over switches at either end of the platforms.
It stops in the suburbs so there isn’t a parking desert in downtown.
It depends on the metro area and the demand along the line.
Not to dispute your main point, but the Chilterns are definitely not flat. I spent a lot of time walking up, down and around there during quarantine times so I can say from personal experience. 🙂
Perhaps there is excessive tunneling, but it’s not trivial terrain (unlike, say, Aylesbury area just north of the hills, which really is flat; I have some great photos overlooking the plains from Coombe Hill).
They’re flatter than what’s between Paris and Lyon.
OK, and its more hilly than anything in Holland, lets play the geography game. Significant tunneling was added because of Tory voting folk not wanting the value of their overinflated property to decrease too much. End of.
As for following existing motorways, sounds good in theory, I mean they could have done what WMATA did for the Dulles Rail link and put it in the central median running at 50mph, but that would rather negate the whole purpose.
The man who ran the HS1 project believes that the M40 was a viable corridor and the CPRE believe the M1 was too. I mean maybe the CPRE were wrong – but I doubt they both were.
@Tunnelvision
I suppose Alon’s point is that Paris to Lyon didn’t require a single tunnel, including about half the route which was completely new ROW (as I remember) which went thru prime Burgundy land so … But, as we know, in the UK the reason is not the hilliness or otherwise. It’s the half of the land owned by less than 1% of the population.
For motorways, surely the way is to build them elevated all the way like the Chinese, Japanese and most Asians (India, Indonesia), and for the same reason: too much ground-level stuff. And it can’t be more expensive than tunnelling. It would appease the treehuggers even though the line was disrupting hardly any. And as is clear the actual cost of building stuff is not why it goes so ridiculously over budget. All too late for Phase 1.
Like the renewable energy naysayers like to point out, that the sun doesn’t shine at night, I want to point out to treehuggers that rich complicated biodiverse ecosystems don’t exist under elevated highways or elevated rail lines.
HSL Zuid has the third highest construction costs of any actually-open line in Europe per km, after Bologna-Florence (almost entirely in tunnel) and HS1 (UK). It has a gratuitous tunnel in the middle just to avoid going through Holland farmland.
“As for following existing motorways, sounds good in theory, I mean they could have done what WMATA did for the Dulles Rail link and put it in the central median running at 50mph, but that would rather negate the whole purpose.”
No one is saying to lay tracks on top of a motorway lane. “Follow the M1” means take roughly the same path as the expressway, depending on the needs of HSR (next to it here, a km or so north of it there, several km to the south farther on). There would be no need to tunnel to preserve nearby untouched land if you are building on land that has already been touched by a multi-lane divided road.
Of course there was really no need to tunnel anywhere as it was, so there’s that.
Why not? They put other motorway lanes on top of motorway lanes. Sometimes 5 stacks in LA. I still remember my first (and last) visit to Hawaii with its highway from the airport into the city being double-decked the entire way. Lots of bridges are double decked (even retrospectively like G. Washington). Can’t quite visualise it but isn’t the Airport Express in Hong Kong above the freeway most of the way (certainly on the double-decked Tsing Ma bridge)? I’d be surprised if it hasn’t happened on some of China’s 25,000km of HSR track.
The design speed for motorways is much lower, it’s for vehicles with rubber tires and they squiggle all over. The squiggly part probably ruins whatever you are imagining.
Yeah, this. In Europe there are a bunch of high-speed lines that are twinned with motorways, but they have small deviations, and consequently never run in the median, always on one side. There’s a deep cut post of mine about it from nine years ago.
Once again that is all about using medians (or sides) at grade. It substantially changes if elevated the whole route (or most of it). There is no reason why the train can’t swing from one far side to the other far side which is at least 100m of most freeways (ie. including the verges etc) so it could considerably smooth any curves. But even so, if it has to go off piste at, say 50%, of curves, motorways usually have quite long straight stretches. How could it not achieve what one wants to avoid: creating a completely new ROW like HS2 has done. IMO that is the heart of the problem with the cost explosions.
Do you know if the Chinese have done this much? Like that Beijing-Shanghai HSR? They built their freeway network long before they began building HSR and they are not averse to building it elevated …
I also quite like re-using the Chiltern Mainline corridor – perhaps with a tunnel under High Wycombe.
That is pretty straight as it is – beyond Wycombe the only challenge is the curve north of Princes Risborough.
Ah yes, have a 300kph alignment follow a 70mph alignment. Such a brainless idea, you cant.
The Frankfurt-Colonge high speed line follows the 3 Autobahn.
“There is no reason why the train can’t swing from one far side to the other far side which is at least 100m of most freeways (ie. including the verges etc) so it could considerably smooth any curves.”
Ha! HSR requires curve radii of thousands of meters (min 7,200m on HS2) while highways have a min curve radius of perhaps 600m at the high end (>110kph).
Once again, “to follow the highway” means “take the same general route through already impacted terrain, instead of cutting new ROW through national parks and untouched rural areas”. It does not mean “follow or use any highway ROW or re-use any civil infrastructure used by a highway”.
The cost explosion has little to do with the ROW. It has to do with:
1) Cost exploding in the US/UK for all transportation projects for decades, as Alon has repeatedly noted.
2) The decision to tunnel to avoid spoiling the Chiltern AONB. This was a choice not a requirement, the route could have been at grade. Following the M1/M40 would help only in that it would allow an at grade ROW through non-protected terrain, thus not needing to make the choice.
3) Pour routing decisions from poor operational design, specifically the decision not to through-route trains at Birmingham/Manchester. I’ll discuss that elsewhere when I have a chance.
@Onux you should start a blog
Yes, this.
“@Onux you should start a blog”
“Yes, this.”
I’m not sure if these are compliments of the quality of my comments or criticism of their length….
The former, not the latter. (tl;dr comments are a long tradition on transit blogs. Reminder: people were telling me and Adirondacker that we’re arguing like an old married couple, in 2009 or so.)
The population density near HS2 in the Chilterns is much higher though.
There’s probably more people living in Aylesbury alone than within 20km of LGV Sud Est.
The population density in Belgium is the same as in England and the costs there are not much higher than in France. Germany has notably higher tunneling, but:
1. German NBSes are generally built in more difficult terrain, because the paradigm here is that long stretches of 160-200 km/h are good enough, and so there’s ~no NBS construction on routes like Berlin-Halle (not on any plan), Hamburg-Hanover (canceled again), Berlin-Dortmund (only Hanover-Bielefeld, the slowest section of this route, has an NBS planned).
2. The routes here are sometimes politicized to produce more tunnels, like the dogleg via Erfurt.
3. The ruling grade here is usually gentler than in France, because of a two-step process to mollify NIMBYs: promise a combined passenger-freight line so that freight trains can be removed from the NIMBYs’ local rail line, then don’t actually run freight trains on the line because there’s no demand and because they interfere with fast passenger trains.
4. Even with all of this, the construction costs per km in Germany are a fraction of those of HS2.
I don’t want to defend HS2 in the slightest. Even if you got costs down to raw HS1 costs you’d be looking at a reduction of 2.5x.
I also think that while the London section certainly cannot straightforwardly use the legacy approaches that costs for the first 100km should cost maybe as much as a similar amount to HS1 – with the remaining sections costing something more comparable to our European peers – depending on what approaches into Birmingham, Manchester and Leeds are appropriate. Most realistically you should have been able to achieve something like a 70-80% cost reduction for the project – which is still pretty good.
That all said if you had an upper middle class hilly region like the Chilterns that the trains had to travel through you’d be tunnelling it in France, Switzerland, Belgium etc – and it wouldn’t be $20m/km to do that. Additionally if the Belgians who have done things at low costs could really do Antwerp Centraal-Brussels Zuid for much less than the HS1 costs per km then it would have happened already – rather than them prioritising lower benefit routes elsewhere in the country.
Belgium is a lot like Germany in that it just leaves value on the table. In Germany, at actual German costs, the benefits of a lot of high-speed rail routes as well as U- and S-Bahn tunnels are very large, but they’re not being built – I rant about the lack of an extension of U8 to Märkisches Viertel on a habitual basis. In Belgium, even at the cost blowouts they’re seeing for the Brussels Metro, it would clearly be worth it to build a Leuven bypass and built a Brussels-Antwerp HSL, but they’re not doing so, nor do they seem to mind that their intercity rail speeds are about on a par with German regional express train speeds.
Alon, if a Brussels Zuid to Antwerpen Centraal high speed line costs more than $60-70/km it might well be worth building. But then it’s difficult to see how HS1 which is pretty the only other high speed line in Europe which includes full new high-speed urban approaches is bad value.
So therefore that means the cost blowouts with HS2 are really a post-Blair thing and if we get a Labour government costs we can build good value high speed rail again.
And also to be fair at least some of the tunnelling through the Chilterns seems like something every other country in Europe would do in that situation so it’s hard to be too upset about the costs of a 300km/h tunnel there. It appears the Germans didn’t even try and build a high speed line from Bamburg to Nuremburg, presumably to avoid upsetting Bamburg and Erlangen, both of which I believe are wealthy.
On the Bamberg-Nürnberg topic: That line is getting improved as part of the same project that built the high speed line through Erfurt (i.e. VDE8). It’s getting quad-tracked right through all the suburbs and downtowns and that doesn’t seem to bother anyone. It might have been better to do a high speed bypass, but I don’t think they decided against that because of NIMBYism. Even for Hamburg-Hannover the opponents were mostly Nabu and the villages along the line, and not so much the wealthy suburbs.
To be fair improving an existing right of way is always less controversial.
And the places along the way will get faster and more reliable express trains and more local trains with the four tracking so everyone benefits.
Bamberg-Nuremberg was because they ran out of money after all the tunnels between Erfurt and Bavaria, no?
All I could find is this:
1) In the 90s it was decided to go through Erfurt (against recommendations by the TU Berlin). 2) 1999-2002 the project was put on hold due to high costs. I couldn’t find whether the project was changed because of this (although it is telling that the baseline cost by itself delayed the project so much). After all a 230kmh ABS isn’t that bad, and from how the curves look nothing should stop higher speeds like on the ABS Berlin-Hannover.
And with high speed rail given the complexities HS1 was OK. So if HS2 was delivered as well as HS1 it would have come in at $20-40bn PPP at the rate for your transit costs project.
The weird thing is that tunneling does not need to be that expensive. Once you have purchased the TBM and supporting equipment, say $40 to 60m, that is essentially a fixed cost and gets spread across the length of the tunnel. Tunneling in that kind of easy ground is a repetitive activity and the segments being installed for the lining are identical and produced in a factory. The crew size is fixed and on HS2 the TBM’s were pretty much cutting edge with as many activities as possible automated, thereby reducing the crew size. The cross passages and vent shafts should not be that expensive unless you want the vent shaft head houses to resemble ancient barns……. or you could have gone for the Channel Tunnel configuration with a service tunnel and piston relief ducts and avoided any surface facilities. Point is just because tunneling is added does not mean it should be more expensive than a surface version of the same route, with embankments, cuttings, bridges, noise barriers etc. The longer the tunnel, the cheaper in theory it should get as the fixed cost is amortized over a longer distance and longer time.
Sure NIMBYism has played a massive part, some of the design changes have been driven by the very same people now whining about the costs, maybe these sort of planning decisions should be removed from local planning authority to avoid the betterments being so substantial.
As for managing the work, the trend towards Project Delivery Partners and additional layers of management t manage the managers is getting a little ridiculous BUT is not really driving the cost increase you see on HS2. Truth is there is not and never was any Government agency in the UK remotely capable of delivering HS2. So on HS2 you have the HS2 management, are essentially Government Employees, then you have the Phase 1 Engineering Delivery Partners who would have produced the reference designs that were bid on, this included SENER to bring experience of the Spanish successes in building high speed rail, then there is the Procurement Delivery Partner. Then there is the Design Builders themselves who have the liability for designing and building the project and somewhere in there you have the CAT3 checkers whose job it is to run a parallel design effort to check that the designs are engineeringly sound. IN this organization HS2 is in of itself a Government owned entity, but if you believe that the Government would be able to recruit and supply sufficient staff to support this project, dream on, they cant even employ enough livestock inspectors for border checks let alone qualified engineering staff.
Lets face it the problems in the construction industry now in the UK with cost overruns, are to some extent because of the procurement models being adopted, which came out of the Latham review in the1990’s that culminated in the New Engineering Contract, and a move away from the Institute of Civil Engineers form of contract which was a little more confrontational in its language. The use of collaborative design and construction processes, the move to design build form of construction but reimbursed at actual costs with only a target costs to aim for partly explains some of the overruns, as any deviation from the reference design on which the contracts were tendered likely constitutes a change and hence additional payment.
But none of the above should be driving a 3x cost increase.
“Point is just because tunneling is added does not mean it should be more expensive than a surface version of the same route, with embankments, cuttings, bridges, noise barriers etc.”
Among the many ignorant comments that get made on this blog and all around the internet, this is easily one of the most ignorant.
NO WHERE IN THE WORLD is tunneling cost any where CLOSE to being the same cost as a surface line. It is ALWAYS more expensive. Period. There might be a few outlaying cases where an easy tunnel might be equal in cost per meter to a difficult bridge, but taking into account a full line the tunnel costs more, every time.
This is so obvious it shouldn’t require stating, but lets address a few points, so that no one who reads this ever believes a shred of what you said is true.
“Tunneling in that kind of easy ground is a repetitive activity”
Surface ROW (roads, railroads, etc.) are also a repetitive activity, with bulldozers, pavers, ballast/track machines, etc. doing the same thing over and over in a linear route, just like a TBM, but much cheaper.
“the segments being installed for the lining are identical and produced in a factory.”
The same is true for precast elements for most civil works (segmented retaining walls, viaduct beams, etc.) but the difference is that the tunnel requires ~10+m3 of concrete for the lining EACH AND EVERY METER as opposed to surface works (bridges, cuts) where concrete is required only at certain places. Tunnels also require lights, fire alarms, ventilation systems, and all kinds of things that surface lines don’t need, all of it expensive compared to dirt, rock and concrete used for cut and fill.
“or you could have gone for the Channel Tunnel configuration with a service tunnel and piston relief ducts and avoided any surface facilities.”
The third tunnel and the piston ducts are all more expensive than vent shafts, and MUCH MORE expensive than OPEN AIR, so this suggestion makes your “tunnels can be cheap” argument even more rubbish.
The Chunnel still has the cross shafts, they are required for evacuation safety, you can’t get rid of their cost.
“The longer the tunnel, the cheaper in theory it should get as the fixed cost is amortized over a longer distance and longer time.”
The longer the tunnel the cheaper the cost-per-linear-meter (relative cost) but the more expensive overall (absolute cost). Have you ever gone to the store and bought 1kg of pasta for less than 500g because the store amortized its rent and utility cost over a larger amount? The longer tunnel isn’t ‘cheaper.’ Plus for all but the shortest tunnels the cost of the tunnel is so much higher than the fixed cost it doesn’t make a difference, we’re talking hundreds of millions (dollars, euro, pounds, doesn’t matter) per km versus a few tens of millions once.
What’s more, a tunnel incurs higher operations and maintenance costs than surface track FOREVER, to power and repair all of those lights, fans etc. I am not going to list them all now, instead go look at the comment on 08 Mar at 13:02 here: https://caltrain-hsr.blogspot.com/2019/03/build-dumbarton-rail-tunnel.html. Not only is a tunnel more expensive in the short term to build it is more expensive in the long term as well.
And I am going to disagree with your moronic ill informed and offensive statement. And no, tunnel costs do not always have to be more expensive than surface options. There is nothing inherent in the process of designing and constructing tunnels that should make them more expensive. If your so right please demonstrate why that statement is incorrect, if you can with some real examples. If you cant, STFU.
There is vastly more earth to remove to create a tunnel in almost all cases. Sure you can find pathological exceptions (I’m sure putting the English tunnel on earth works would be more than a tunnel even ignoring what that does to shipping and the environment), but in general putting track on the ground is a lot less earth works.
Which does not answer the question I posed.
There are many, many, many things “inherent in the process of designing and constructing tunnels” that does make them more expensive, lets review:
1) Tunnels require more expensive equipment. A TBM at $40-60M costs more than all of equipment that would be used for surface construction (a several dozen pieces of equipment at $0.5-$1M each). You must buy the TBM custom for a tunnel, but for surface work firms bring their own equipment then take it to the next job. The tunnel still needs surface equipment to build the launch box and ventilation/access shafts, etc. Equipment to work inside the tunnel, like mining equipment for cross shafts, is specialized and more expensive, while surface work uses normal bulldozers, etc.
2) More concrete used, as I already mentioned 10m3 or more for each meter of route just for the lining.
3) More earth moved. That same tunnel needs ~50m3 of excavation for each meter of progress. In contrast a 2m high sloped berm with 4.5m track centers needs 20m3 of fill for each meter forward. As with concrete this is just for the base tunnel, before accounting for cross tunnels or ventilation shafts.
4) Earth has to be moved farther. First you have to move the earth a few thousand meters to a portal or lift the earth a few tens of meters to the surface (or both!), then you have to move it tens of kilometers or more to a dump. On the surface you frequently move the dirt only a few hundred meters, as spoil from a cut becomes fill for a berm.
5) Tunnels require lights, surface track does not.
6) Tunnels require ventilation equipment to circulate fresh air, surface track does not.
7) Tunnels require fire alarm equipment (smoke detectors, horns, etc.), surface track does not.
8) Tunnels require drains, drain piping, and pumps to move water out of the tunnel to the surface or elsewhere. Surface tracks requires much simpler trench drains or ditches (or in some terrain no drainage) and never pumps.
9) Tunnels require radio relay systems for emergency communications. On the surface there is nothing to block radio waves so you don’t pay to buy or install this.
10) Tunnels require electrical equipment for #5-9 above, beyond the system to power the trains. That means cables, transformers, electrical panels, etc. Surface track does not need electrical power for systems that do not exist.
11) Tunnels require fireproof doors in cross shafts, evacuation signage, security cameras, ladders to climb down to the track from the walkways, and a host of other items surface track does not.
12) Tunnels require a fire sprinkler system, surface track does not.
13) Installing everything from #5-12 takes longer because access is limited and installers can’t just drive their truck up and get to work. Thus installation cost per unit is higher than for a surface building due to lower productivity.
14) Tunneling work is a rare activity performed by limited firms/unions/personnel. Surface civil infrastructure is common and performed by many grading/paving/concrete firms. The supply/demand mismatch means you pay more for tunneling.
Since you need things explained in a very simple fashion, I will tell you that you pay for things by volume, distance, and time. Pouring 10m3 of concrete is twice as expensive to pour 5m3. Moving 1m3 of dirt 10km is more expensive more than moving it 500m. If a pipefitter can install 20m/shift of sprinkler in a building, but only 16m/shift underground, then your labor cost will be 25% more expensive. I am explaining this so simply because you said that henrymiller74’s point about moving more earth “does not answer the question” when most school children understand that doing more costs more. Thus every item above is a “real example” while tunnels are always more expensive than surface options, and that your original statement remains astoundingly ignorant, regardless if you choose to disagree with me or not.
If by real examples you are asking for a list of tunnel projects that are more expensive than at-grade construction, the list consists of every tunnel every built and every surface line ever built (accounting for inflation). In every country, at every time, the cost for a tunnel is more than track on the surface. It might be possible that the most expensive stretch of surface track in the world (somewhere US/UK) might be slightly more than the cheapest subway (Spain? Turkey? Sweden?) but a tunnel in the US/UK costs more than surface, every time, and surface track in the Spain/Turkey/etc. is cheaper than tunnels, every time.
Surface civil infrastructure is common
There are design and structural considerations. It’s a bit more complex. The difference between building a highway and building a rail road is that the highway has a thin layer of pavement on top and the railroad has rails. The contractors – most of them – can switch between the two and probably do.
The chancellor Jeremy Hunt has clearly seen your research 😀
Or just listened to his mandarins, who’ve always hated the scheme and forced it to do non-standard practices in an attempt to sink it by making it cost a lot more. For instance, demanding costs be P95, rather than P50, doubled the public price tag overnight!
Though, of course, the man himself isn’t blameless – *unt’s ‘austerity’ program has added a good few billion to the price tag – saving a few billion this year by pausing construction but ending up paying a lot more in future instead. I guess that’s the actions of a government that knows it can make a mess this year and force the next lot to clean it up when they come in next year.
Hunt’s pretty much an empty vessel like Cameron. In a less bankrupt political culture he’d be a modest normie centre-right politician. But the British state has such weak structural legitimacy that it can’t actually governand relies on being able to extract revenue from Southeast England remaining agglomerations.
HS1 was built for 300kmph. The railways HS2 is being compared to were also built for 300kmph? That gives quite a bit more flexibility in route selection…
The claim is repeatedly advanced by its supporters that it’s really about capacity but not speed, but if it was then those benefits would be realised without going so fast. If you read the reports on its benefits, it’s pretty clear it’s about speed — most of the ‘economic benefit’ comes from assigning monetary value to the hypothetical time savings. Which don’t, of course, translate into higher GDP; the country doesn’t get back the money it spent making things quicker for a small set of commuters.
I wrote a comment on that issue but it was swallowed/disappeared by WP.
The primary driving force behind building it was in fact capacity. The mainlines are increasingly crowded, needing maintenance that required weekend closures etc. and there was no spare space to squeeze in extra lines. And they can’t even take duplex trains to relieve the demand. So a completely new (and necessarily independent) ROW was required. Now, if you are going to do that and the inevitable NIMBY and eco battles, are you really going to not make it HSR? Of course you’re going to make it HSR, especially as the price differences when building new is not so great.
So the concept was a new HSR line that would take a lot of traffic off the existing mainlines which could then be maintained (ie. closed for proper upgrading). Ultimately it would be win-win for everyone. The large inter-city traffic would get an express fast service while the old mainline would serve all the smaller cities with a better service. Ditto for freight.
Of course some PR gurus may have chosen the wrong strategy of selling it to the public by stressing the high-speed and the faster city-to-city times, modest as the improvements were. No one ever seems to explain it properly and even smart people, like PPEs Simon Jenkins and George Monbiot (who Borners is convinced I lurve) never mention this in their opposition to HS2. They’re wrong.
Yeah, because the spamfilter got aggressive and swallowed nine comments, right after an LLM-generated spam comment got past it and I spammed it, which I think trained the filter to be worse at telling these apart. Sorry.
Speaking of cost increases:
It was announced today that BART to Downtown San Jose
is now estimated at $12.2 billion for six miles.
The cost is really bad but changing the completion date of 2026 to 2036 is even worse. Why are we going to expect people to abandon cars in favor of transit when it takes so long to build transit?
In related American news, the six mile BART extension to downtown San Jose is now expected to take a decade more to build and will be double the cost. This means that it won’t be ready to open until 2036 and will cost over 12 billion. Building more transit and getting more people on transit is an important part of combatting climate change but English speaking countries in particular seem pathetic at this.
I thought I was used to the incompetence of US construction, but the $12 billion number is enough to make a grown man cry.
Just… stupid. Malice, stupidity, terrible priorities, take your pick. Criminally expensive, and for what?
Now a metaphorical gun has been pointed at their heads the Euston section is 10km for £3bn including a terminus station. Which actually isn’t bad.
In what year dollars?, is what I don’t get; if it’s in 2032 dollars or something like that then it might well be less bad than the worst non-US subway I know of (the last phase of the Circle line in Singapore).
Yeah, okay, this is bad.
The £3bn includes tunnelling from Old Oak Common to Euston with 400m long surface station in central London presumably with 8-10 platforms. £3bn really isn’t terrible for that.
Omiya-Tokyo was ~¥26bn/km at 2007 prices – and Japan can turn trains around much more efficiently at Tokyo than we realistically can in Europe with our lower reliability and more leisure travel.
More billable hours for white collar professionals ! ! !
Speaking of cost increases (again):
The 2.2 mile extension of Caltrain into Salesforce SF is now estimated at $8.25 billion.
It makes BART to Downtown San Jose look cheap ….
I wish this were in constant dollars. A lot of these cost overruns are largely rebasing to a future year with higher expected inflation. I hate it when this kind of accounting is done in nominal and not real dollars.
Elon saw the new cost increase.
Musk says his Boring Company can provide the transportation for 1% of the Caltrain number.
Not sure who or what to believe …
https://www.benzinga.com/news/23/10/35478196/tesla-ceo-elon-musk-offers-to-solve-8-2b-sf-downtown-rail-extension-problem-for-just-1-of-cost
What $82m for 3km underground? That feels a bit low…
Alon,
You write a bit about construction costs. How about operating costs? CAHSR just announced a labor agreement with 13 unions to operate the initial Central Valley segment. They estimate 400 jobs. Does that sound like too many? Too few?
https://hsr.ca.gov/2023/11/17/news-release-california-high-speed-rail-partners-with-labor-on-operating-nations-first-high-speed-rail-project/