Australian Construction Costs
There’s a report just released by the Grattan Institute called Megabang for Megabucks, talking about high construction costs in Australia. Our transit costs project is quoted as an international comparison, pointing out that Australia is near the global high end. I encourage people to read the report itself, which says interesting things about problems with Australian construction and procurement. I am especially happy to see that the recommendations for the most part accord with what we are learning from other cases – of course, our Boston case is out and the report authors have likely read it, but the recommendations are in line with things we see from yet-unpublished cases, so this is not just me looking at a mirror.
The issue of competition
Australian megaproject contracts have insufficient competition. Only three firms are Tier One, the largest infrastructure contractors in Australia; those get most contracts for the largest infrastructure projects, and when mid-tier firms bid, it’s often in partnership with a Tier One company. Moreover, in the largest size category, higher than $1 billion, even the Tier One firms often partner with one another, leading to monopoly.
International firms do access the Australian market, but it is inconsistent. Australia overweights the importance of local experience, and has some unusual rules, such as requiring firms to engage in more prior design than is typical.
This is consistent with what I’ve seen in Israel. In short, the electrification contract in Israel was won by Spanish contractor SEMI, which had extensive European experience but none in Israel. This was criticized domestically, and some people blamed it for the schedule slips on the electrification project, but such blame is unfair. The bulk of the delays are not the fault of SEMI but come from a lawsuit launched by Alstom, which competed for the contract and lost out on price; Alston employed industrial espionage to create FUD about the bid, and the lawsuit delayed works by three years. Despite this, the costs have not run over much, and the absolute per-km costs remain on the low side, net of extras like Haifa’s demand for a trench. Thus, even in a situation of extensive domestic complaints about the winning bidder’s lack of local experience, said lack did not materially create problems.
This is also consistent with lessons from Turkey. In Turkey, there must be a minimum of three bidders. If there are only one or two, the state or municipal government must rebid. Absolute costs in Turkey are low and so are cost overruns; the extensive competition helps discipline the contractors, as does the political consensus in favor of rapid infrastructure construction, credibly promising firms that there will be more work in the future and if they behave they will get some of it.
The study discusses different contracting regimes. It does not talk about the design-build issue; I do not know whether it is as prevalent in Australia as in Canada, and regrettably there is no cost history, thus no way for me to confirm my suspicion that Australia resembles Canada and Singapore in only having had a cost explosion in the last 20 years. However, it does talk about change orders.
Change orders are a notable problem in California. Low bids followed by renegotiation are common there; Tutor Perini is notorious for this behavior. The study goes over strategies to deal with this issue, though it does not talk explicitly about itemization as in Spain and Italy, where the unit prices are public and then if more is needed (e.g. more labor due to slower progress) then the change is already pre-agreed, avoiding litigation. Sweden avoids litigation as well.
Finally, the study talks about rushing. This was an issue in Boston, so this may be me learning from a mirror, but, in brief, American funding for infrastructure encourages agencies to rush the preliminary design to apply for federal funding early. This leads to compromised designs and premature commitment, since there is no ongoing funding for long-term design.
Learning from good examples
I think the one drawback of the study is the list of comparisons. Sourced partly to us and partly to Read-Efron, they say,
The empirical evidence is incomplete, but what there is shows that rail construction costs in Australia are in the top quarter of 27 OECD countries studied. They are higher than in numerous other rich countries: 26 per cent higher than in Canada, 29 per cent higher than in Japan, and more than three times as high as in Spain (Figure 1.2 on the following page). And road and rail tunnels cost more in Australia than elsewhere in the world, according to an international study.
The comparison with Canada has a problem: the Canadian costs in our database go back 15-20 years, and back then, costs were much lower than today. The latest costs do not show an Australian premium over Canada – Toronto is more expensive to build in than Sydney and almost as much as Melbourne. It is critical to understand that high costs are really a pan-Anglosphere phenomenon, and thus Australia should learn from Continental European and East Asian examples (except very high-cost Hong Kong), and not from countries that in the last 10 years have had the same problems as Australia or worse. Spain is always good, as are common features to low-cost Spain, Italy, Turkey, South Korea, and the Nordic countries, and even common features to those and medium-cost countries like France, Germany, China, and Japan.
Isn’t Hong Kong an honorary member of the Anglosphere? I mean, it’s only been what, 20 years since the lease was terminated and it was “given” back?
I don’t really know much about Israel (despite having a sister who lived on a kibbutz there for a while) but they kinda seem, from a cursory glance, to be more “American” than they would probably care to admit …
If any of this is true, it means that they are cost outliers relative to the world, but not to the Anglosphere.
I sometimes nursed a pet conspiracy theory that China was happy to let it retain its Anglosphere infra and housing problems in stark contrast to China’s massive building boom. Certainly seems like a superb tactic to me.
*if it was intentional that is. I have no evidence that is was, and perhaps it could also be a natural result of opposing factions at loggerheads about most things.
China is the only place in the world that I can think of where real estate speculation can support such staggeringly high housing vacancy rates with zero hope of anyone ever living in many places. I don’t think the building boom is really something for China to be proud of, more like something to be nervous about.
Not really, as major Chinese cities have worse housing problem than Hong Kong nowadays as they learn from Hong Kong’s mode of funding the government budget by selling lands expensively.
As of 2020, Shenzhen’s income-to-house-price level is 48.1, meaning an average citizen earning average income need to work 48.1 years without spending a single cent of money they have earned for food or any other things, in order to afford their own housing.
Even worse for residents in Mainland Chinese cities is that, while Hong Kong have 50% housing available as social housing, in Mainland China there are only token numbers of them in each city. The result is in cities like Shenzhen, half of city live in subpar places that are technically illegal, as opposed to Hong Kong’a 2%.
And from the other perspective, in fact the current trend of Hong Kong property prive was set on late 1980s to early 1990s. The Sino-British Joint Declaration heavily restricted land sale to developer in the decade before handover of Hong Kong from UK to China as Chinese government fear the British would try to make quick buck overselling Hong Kong’s land reserve and that caused explosive rise of Hong Kong property price until 1997. Following 1997 the Hong Kong government tried to correct the problem by massively increasing land supply, and when combined with Asian Financial Crisis and numerous negative economical events aroumd the world, such expectes huge increase in supply and huge drop in demand mean the price dropped drastically. The huge fluctuation have damaged economy and social stability, with many who habe bought their apartments before 1997 now going bankrupt amd property owners jumping from building have become daily news. Seeing such significant impact, the government walked back their housing target, opt to completely freeze the develop of public housing and even converted some social housing into private housings, stopped periodically selling lands to property developers, cancelled development of new towns, and property developers have also lobbied for a reclamation ban in Hong Kong’s harbor to limit supply. These policies sustained for more than a decade, despite global economy have started picking back up after 2003 and into overheat in 2008, and despite the rapid growth of Chinese economy, as well as the development of free trade agreement and freedom of travel for individual between Mainland China and Hong Kong as started from year 2003 have caused eapid increase in economic growth and thus demand in the city. The result of all these factors combined is that the coty’s property price only barely stretched a bit in 2008 and continues to rocket till today. Starting from around 2014 the government (both Hong Kong domestic government and the Chinese Central government) started to recognize the unaffordibility of housing in Hong Kong as a severe problem and a destabilizing factor, and now they’re trying to convert all sort of lands in the city into housing development and plan to build a new big artifical island which is said to capable of housing 1 or 2 million people in the next decade.
Does Mainland housing affordability really hold constant outside of Shenzhen though? The big Tier 1s (Shenzhen, Guangzhou, Shanghai, Beijing) are generally perceived as being “expensive” cities by Chinese standards (although it really applies to the 1st three primarily), but does this still apply to New Tier 1, Tier 2, and Tier 3 cities? Particularly ones not in the south or near special economic zones?
Many “New Tier 1” and Tier 2 cities still often have average housing price equal to 10 years salary of an average workers, exceeding the rate of even many American cities. Although one can argue income inequality is a big reason for such relatively expensive housing price
It’s not just income inequality, a lot of it is driven by CCP capital controls and the lack of remotely trustworthy domestic alternatives to real estate.
I don’t think hot money is really that big of a factor when it come to property price..?
It is in Communist China. Due to the untrustworthiness of other Chinese investments like stocks, and the restrictions on getting the money out of China to put into more trustworthy foreign investments, spare Chinese money just goes straight into Chinese real estate. Prices keep going up despite staggeringly high vacancy rates, and afaik, it’s the only place in the world where property speculation can support such absurdity. Not even the very famous Japanese bubble got that bad, and prices for residential land in Tokyo peaked and started going down before the Nikkei fell off a cliff.
I think you could extend the scare quotes to “back” as there is good argument that the Qing dynasty has not been legitimately succeeded by the CCP
Even for Anglosphere I don’t think it is normal to anticipate a construction cost of US$2.5-4 Billion for a ~15 km long, mostly at grade, bus or tram system?
lol, talk about “success problems”: learning from a mirror.
The issue with lack of competition was interesting to me, since in Taiwan (with an almost identical population) there are also only three major infrastructure firms and it’s been a pretty big problem in recent years. Many projects have seen delays simply because there were no bids the first time the RFPs went out; this has been the case for the Taoyuan Airport Terminal 3 project, as well as almost every new metro line over the past decade.
(When this happens railfans in Taiwan tend to blame the government for lowballing cost estimates and being penny-wise/pound-foolish. There really is very little awareness of cost control issues here.)
The Taoyuan T3 project ended up going to a joint bid by Samsung and a local firm, so maybe Taiwan will see some progress in opening up the market, but it’ll probably always be more difficult for Australia due to the language barrier.
With regards to designs being “rushed” I’d have thought the bigger issue was the public sector not being prepared to hire a ~5-10 person permanent design team where they are given a competitive salary and conditions compared to the private sector.
I mean in the UK that would cost maybe £1m/year including all costs. Compared to the cost overruns of HS2 the cost would be chicken feed.
Yeah, the US has the same problem. The cost premium per MBTA station over having standardized design is larger than the cost of hiring an in-house team to develop a standardized design, and yet this isn’t done. South Coast Rail includes an item, I believe low single-digit millions, for standardized station designs along the line, outsourced to consultants of course; the planners hope to be able to reuse these designs for high-platform projects on other lines.
As I’ve long ago learned, reusing design is not popular. Everyone thinks THEIR station is unique and so it cannot be standard. I’m not talking about each station needing a unique color paint and a different sign, which should be done and is cheap, I’m talking about needing an all new ground up redesign to reflect their own cities “unique” personality.
This isn’t unique to rail – schools all look about the same, yet each has a unique design because God forbid that we recognize that schools all look the same. The state education department could design a few standard schools and that is all they allow new schools to choose from. It isn’t hard, you need 1, 2, 3, or 4 classrooms per grade, and then between 2 and 7 grades in a school. (high school is a bit more complex, but not much. The cost of unused rooms if there is something unique is less than the cost of a new design – and I’m sure each school will find a use for an unused room. I have no doubt that you can do the same for lots of other things – schools are just the most visible.
Aren’t architecture fees around 15% of materials? Yes it’s extra cost, but 15% extra cost is a small amount for these projects.
Isn’t the biggest effect that economies of scale and speed of construction are lost when every project is a new design? It also gives the contractors less scope to pull their “unexpected complications” stunt to adjust their fees.
Make it 20% for one, and then pay some extra to build more jigs. The second will be a lot cheaper.
There is a lot of room to standardize as well without forcing exactly one station design. Single family houses are as cheap as they are even though each is difference because every room has studs that are 92-5/8 or 104-5/8 that come pre-cut from the sawmill. A ton of labor is automated in the factory and architects carefully design as many rooms as possible to fix the standard size. Likewise you can get your doors in one of 5 different widths (I don’t recall the exact number), all exactly the same height – you can get other sizes if you want but the cost will be much higher. Your stair spindles come in 3 different decorative styles, choose one. Each time you standardize something you save a lot of costs down the line.
Thus it might be better to think here is the standard station choices: 8, 12, or 16 cars long. With each you can move the entrances around (because each city will have a different ideal location). Standard modules for the gift shop (everyone thinks they need these, though I don’t know why), ticket booth (hopefully mostly automated), restrooms. The 16 car long station has 3 restroom modules, but they modules are the same. We probably need something for 2 or 4 tracks, and probably a standard city bus shelter attached…. the point here is to limit choices to something that any non-engineer can move around until they find a layout they like and then give it to the contractor who has already agreed on the price for each module.
The above will cost as much for the first utilitarian station (though because of the standard modules there may be room for unique “art”) as one oversized palace station, but the next station is almost free.
15% is a HUGE amount! Most construction projects have profits margins of only a few %. 15% on a billion dollar project is $150 million. Cutting 15% of design costs would mean that for every stations you are planning to build you could build an eighth for the same budget!
That may be somewhat how the Chinese do it, but when it comes to subway construction, soil conditions and such would make a large difference, no? I’d like to see a comparison between Chongqing (yes, I know, an extreme case) and the median Chinese city to see how much such “modularization” can help.
And the Chinese even already have standardized rolling stock design – MTR or ex-USSR specifications, pick one. A similar thing could have happened in the US between WMATA, MARTA and some of the other contemporary systems. NYC might be able to join too had they kept the 75 footers.
@yuuka Tunneling in different soil is well understood and different costs come down to mostly labor which isn’t very expensive. NYC where they insist on twice as many people to run a TBM, and pay them too much is an exception, but even there the tunnel itself isn’t too far out of line. Alon knows more than me, but they keeps saying tunnels are expensive but stations are where the cost blowups are coming from.
Yes some soils can cost 30% more to build a tunnel in than others, but once your soil survey is done along the line you can accurately predict the costs and order the right TBM(s) to do the job. Spain in fact credits part of their low cost tunneling costs to deciding to order an overpowered TBM just in case they encountered more difficult rock than expected.
When the new Eastern Span of the Bay Bridge (between San Francisco and Oakland) was running hugely late in schedule and hugely overbudget, then California Governor Schwartzenegger proposed starting over with a simple, cost effective design. Brown and Brown (mayor Willie Brown/SF and mayor Jerry Brown/Oakland) refused, saying that their legacy would not be a “freeway on stilts” and it needed to be a “signature” design. So, a 2500% cost overrun and $7 billion for 2.2 miles is the result.
I’ve posted this before but it bears repeating, especially in this context:
That is a feature not a bug. Emasculate the oversight while singing the glories of the super efficient private sector means the private sector gets very efficient at extracting more money. it works in other areas too.
Matthew Hutton: “I’d have thought the bigger issue was the public sector not being prepared to hire a ~5-10 person permanent design team where they are given a competitive salary and conditions compared to the private sector.”
Further to my earlier comment, your sensible solution is confounded by the revolving doors problem. The only way to have such a team, and critically to keep it stable and intact with inherited domain knowledge etc, is to have laws preventing such public servants from going to work for the companies they previously supervised or audited etc, at least for a period (most likely 5 years to be effective). There are those who claim that you have to pay big bucks and give the freedom to move back & forth between public and private, to attract good talent. I reckon that is a steaming pile. Plenty of good and talented people are happy to work for the public, and build a satisfying career. And wouldn’t worry about such a restriction. Indeed I believe the system would be auto-selecting the right kind of people instead of the current arrangement where it is perceived as a mere stepping stone to big bucks in the consultancy game.
In fact it is the politicians who won’t do it because they want the right for themselves to immediately jump from their elected position–and often ministerial level–to a senior job, usually on the non-exec board or as high-paid consultant, exactly in the area they oversaw as a politician. We’ve had increasingly outrageous flouting of the two-year rule. This conservative government has take it to new lows, simply acting as if the laws don’t exist. Last election cycle we had the retiring minister of defence go to work for the biggest defence lobbying outfit in Australia, within days of leaving office.They were involved in the biggest defence contract in Australia’s history–the $50bn French submarines–which he had negotiated while in office. He claimed there would be Chinese walls.
There was another guy who actually went on private salary days before he retired from being a politician; he should have been prosecuted but of course nothing was done. Another awful case was a minister of trade accepting $800,000 pa consultancy for Landbridge, a subsidiary of Shandong Landbridge Group which is controlled by PLA interests, who then bought the Port of Darwin as a link in their maritime Silk Road. This is the only large scale commercial shipping port in the whole of northern Australia–for about 3,000km east and even longer (4,000km?) west, yet it escaped federal Foreign Investment Review Board review, ie. it didn’t even go to review! Six years later the same government that didn’t blink over it then, is in the process of reversing it (to upset the Chinese even more). The ex-minister continues to blatantly state there was no conflict of interest.
Another well known case was of Bob Carr after retiring as one of the longest-serving Premiers of NSW went to work for Macquarie who had financed and built most of the road tunnels under Sydney that were promoted during Carr’s term of office (and to the exclusion of public transit; indeed some of the PPPs included notorious clauses forbidding the state from running transit that would compete directly with the toll roads, or mandated closure of nearby roads that would effectively force-feed drivers into the tunnels).
The reason politicians don’t want this rule, and blatantly broach it, is obvious: their worth to the private sector diminishes fairly quickly with time out of office.
I suppose this type of insidious corruption is universal and has been with us forever but it is my contention that since the Reagan-Thatcher years, in the Anglosphere at least, selfless public service has gone out the window (“only for losers”) and likewise concepts of the national interest.
Can you discuss a potential pitfall of “in house expertise”?
The city of Nuremberg has an “U-Bahn Bauamt” which is a permanent part of the civil service. Without specifically U-Bahn construction they have nothing to justify their existence and as such there is a certain push to not be mode neutral in expansion discussions. This leads among other things to studying potential U-Bahn extensions that are far from the required benefit cost quotient above 1.0 while proposed tram extensions like the “Old town tram” with a benefit cost quotient above 3 are no where near construction.
Granted, this is less of a problem in a place like Madrid where the light rail system is fairly small compared to the subway, but shouldn’t such in house expertise be as mode neutral as feasible?
Is it not more of a problem that they have a team solely dedicated to U-bahn design instead of overall transit design in general and picking the best technology that suits the desired service? Off course if their sole purpose is to design U-bahns, they would always suggest one even where it does not make sense.
I mean definitely you’d want the team to do all public transport at the very least. Really perhaps including roads and cycle lanes.
We have DOTs all over that know how to design roads, but don’t care to do good transit or cycle lanes. (they will do cycle lanes, but often they are only useful for exercise)
The way I see it, they could, but half the department would still be out of a job or have a very light workload, to the point where people just ask “why not privatize the damn thing?”.
Building a U-Bahn needs a lot of people. Moving to tramways (as an example), while still keeping the architects, planners, and rail systems experts employed, may mean letting go of some of the civil engineers and geology experts at least.
I recall being told on here that it was better off being done at national level, makes me suspect that’s better because at a national level there’s more than enough projects to keep everyone occupied, spanning the gamut from HSR to tramways and roads.
Yes, and this leads to questions like, why is the Bauamt specific to the U-Bahn rather than to public transport in general?
https://www.nuernberg.de/internet/u_bahnbau/u_bahn_wir.html they don’t say…
1) In case you didn’t see it, Caltrain has put out an estimate of doing one grade separation, at Broadway in Burlingame.
Click to access $!2314a+-+Presentation+-+Grade+Separation+Program+Update+-+added+1-4-2021.pdf
2) I am intrigued by the idea of a planned project being forcibly stopped and redone if it doesn’t get at least three bids. Would that work in the US? Would that work in the New York 2nd Ave Subway? Would it work for California HSR? Tudor Perini talks a lot about being the only bidder on many projects.
If I remember correctly, there were 3+ bidders for all 4 construction packages executed by California HSR so far. CP 5 could be the first package with only 2 bidders invited to the final bid?
Interesting studies. However I feel like this has all been noted before. Is there any count of how many studies over the years have found much the same thing, with no affect on policy?
Yes, I’m feeling cynical today.
A couple of points about the Australian experience:
1. high construction costs are a recent phenomenon. It’s really on the last ten years or so that it’s been an issue. But the rate of cost inflation is alarming, and doesn’t show signs of abating. Sydney is the only city with enough transit projects to really do an adequate comparison, but it is indicative of the nationwide trend (nominal $AU prices, although general inflation has been pretty low):
1999: Airport line, 9km, 5 stations (all underground): $900m, $100m per km
2008: Epping-Chatswood line, 13km, 5 stations (all underground): $2.3b, $176m per km
2019: North-West metro, 23km, 8 stations (2/3 underground): $7.3b, $317m per km*
2024 (under construction): City & South-West metro, 17km, 7 stations (all underground): $15.5b, $911m per km (a lot of it is through the CBD, so a premium applies here)*
2027 (planned): WSA metro, 22km, 6 stations (mainly above ground through unpopulated land): est. $11b, $500m per km*
2030 (planned): West metro, 23km, 9 stations (all underground): est. $27b, $1.17b per km*
* includes stabling, rolling stock and possibly also an operations/maintenance contract
2. It’s hard to explain why project costs should be skyrocketing, especially since in most regards project management has become considerably better. The first two projects were scandal-prone, chaotic affairs overseen by a dysfunctional government. But they were completed for prices that are unimaginable bargains today. Meanwhile, the latest estimates for the West metro, breaking the $1b per km barrier, don’t seem to have anyone in the media or the public batting an eyelid.
3. In general, though, the rushing critique is an important one. The thing is, new railway lines are basically seen as something one political party gifts to the electorate as a reward/bribe for being elected. With the result that there is little coordinated or advanced planning, and projects are sprung on the public virtually without warning. But the government then wants to see tangible results within their expected term of office (usually they envisage being in power for two terms, so eight years), and so the whole process is rushed through with no regard for cost overruns. The “suburban rail loop” in Melbourne is a very good example of this practice: it was promised a few weeks before the election with only very sketchy planning since the party proposing it didn’t want to spoil the surprise, and then a dubious CBA had to be retconned to justify it. At the same time, other worthy projects lie in purgatory for other, subject to endless studies that only serve as stalling tactics (in Melbourne, the long-planned Rowville and Doncaster lines).
Materials are probably a factor, although I couldn’t say how significant. But even simple raw materials like steel, lumber, and concrete have skyrocketed in costs the past few years. Some are pandemic related, others due to China soaking up a lot, or trade disputes, and some due to environmental regulation.
There might be an argument on whether this is a trend you see in many countries, not just what this comment section likes to rag on about.
I doubt materials are much of a factor, and the Grattan report (despite being economically very liberal) also downplays labour as a significant factor. It really comes down to lack of long-term planning, project management and the interface between the government and contractors.
Thanks to Alon for covering this and df1982, michaelrjames for providing local input.
I’m also baffled by Sydney’s costs and the rate of increase, especially with a right-leaning Government finally willing to invest in transit (though no real concept of walking and cycling and the absolute silver bullet they present for our city and the budget).
When I engage experienced engineers and consultants on the cost issue, they tend to have few answers, but when pushed, recommend changes that Alon and others have shown are the problem, not the solution. I find the de-skilling of the public sector to be a convincing argument as many have said including Senator Faruqi.
This is highly speculative, but I honestly feel that the Government has embraced spending to such a degree in polarity to predecessors who built nothing, and competes so aggressively with Melbourne on which is the bigger/better city, that the higher number is better, not worse. There’s no way this can be the cause of the increase, but the incentive to mitigate or address the problem seems very low. No data to back this observation up.
Alon: “no way for me to confirm my suspicion that Australia resembles Canada and Singapore in only having had a cost explosion in the last 20 years.”
That has been my distinct impression however without a firm basis in data, and patchy impressions due to my long absence. But in that period the country has gone from being low-cost to high-cost in most things with the biggest change happening in the first decade of the new millennium. And not just in Sydney and Melbourne but all over, including in remote areas. I suppose it was a Sydney phenomenon that spread nationwide. Most of our bad habits are British imports but in the same period it has become co-equal with American influences. Yes, we’re f’d.
An interesting recommendation of the Grattan Institute is that governments should have more spine to enforce the contracts they sign. But it is a big ask in a democracy and esp. Australia with short terms and effectively even-shorter terms due to election dates being at parliamentary discretion. I described this phenomenon on your blog with respect to the 12.8km Sydney CBD-Randwick tramway in which the main contractor waited until quite late–when it was within a year of being completed and thus essentially irreversible–to hit them with another billion dollars of charges on the back of opaque issues. It happens that the state government had an imminent election (but then they are always imminent) and so of course, after a bit of son-et-lumiere it caved and reached a $576 million settlement. It went from about $1.5bn to end up costing at least $2.7bn. It was a PPP.
Related to some other comments under this post about construction cost in Hong Kong:
According to analysis by a Chinese netizen at https://www.zhihu.com/question/31031089 ,
Apparently almost half of the Hong Kong High Speed Rail project cost (84.4 billion HKD) is the cost of building the station (30.5 Billion CNY).
On the other hand the post claim the, while in China regular stations cost (structure cost only) are counted toward high speed rail project cost, major hub stations that are typically funded independently are being excluded, and that mean for example the Guangzhou – Shenzhen – Hong Kong High Speed Line project cost figure excluded the cost for the two largest stations on the line, Guangzhou South and Shenzhen North, with the two stations combined costing an estimated number of ~20 Billion CNY. All accessory works around all stations, like traffic rerouting and improvement works that are consequence to the construction of High Speed Rail station, are also being excluded from the total project cost in China, while in Hong Kong such figure is being included into the project cost (costing at least 11.8 Billion HKD)
And the user also pointed out that the High Speed Rail project in Hong Kong started construction before design and surveying works are done, which resulted in a number of ad hoc changes during construction and increased the cost (Is this like CAHSR?)
But as another user pointed out, the selling of property development right on the station have been more than enough to cover the expensive station construction cost.
Yeah, this is also what I saw in Beijing. That said, Beijing South was a 7 billion yuan project in the mid-2000s, which is on the one hand really expensive (it’s $3.2 billion in 2021 dollars), and on the other hand a 2.5th-order cost for the Beijing-Shanghai HSR line.