This post is a follow-up to one I wrote last week (see the latter half of it) and also a response to a more recent post by David Walker. I certainly wouldn’t argue with David’s assertion that a Sydney-Brisbane Very Fast Train route is not likely to be remotely viable even in the medium term. Pessimistic “realism” about that route was the major premise of David’s post. The population of Brisbane/Gold Coast is only around half that of either Sydney or Melbourne and the terrain in between is much more challenging than that between Sydney and Melbourne, making construction costs significantly higher.
However, I think a Sydney-Canberra-Melbourne line is worth examining much more seriously. I certainly wouldn’t suggest (or support) construction without rigorous assessment and cost/benefit analysis by the Infrastructure Commission, but by my own rough “back of the envelope” calculations I don’t think the picture is anywhere near as negative as David Walker argues. The terrain is much more benign than Sydney-Brisbane for a start (except for the 100 kilometres or so between Campbelltown and Mittagong).
More importantly, close to 50% of Australia’s population live within 50 kilometres of the route I envisage, and the Sydney-Melbourne air route is already the fifth most heavily travelled in the world. The route I have in mind, including all intermediate stations en route (with current population of each intermediate city), is as follows:
| CITY/STATION | CURRENT POPULATION |
| Sydney | 5,000,000 |
| Mascot Airport | |
| Liverpool | 190,000 |
| Badgerys Creek Airport | |
| Campbelltown | 160,000 |
| Goulburn | 23,000 |
| Canberra | 360,000 |
| Albury/Wodonga | 90,000 |
| Shepparton | 30,000 |
| Bendigo | 150,000 |
| Tullamarine Airport | |
| Melbourne | 5,000,000 |
| TOTAL | 11,003,000 |
That population (prospective passenger) base will have increased by around 13% by 2026 when I am assuming the line would be completed (to coincide with the opening of Badgerys Creek Airport).
Contrary to David’s claim that general expert opinion dismisses the viability of VFT in Australia, the most recent report under federal Infrastructure Department auspices in 2013 found:
Our first and most important conclusion is that high speed rail has the potential to be an integral part of Australia’s future.
The evidence is clear. It would build capacity and resilience into the major east coast transport networks, alleviate pressure on the busiest intercity air services, and thereby allow growth of international and new domestic aviation operations, and help diversify infrastructure investment into more energy efficient technology. More fundamentally, high speed rail would contribute to national productivity and open up greater opportunities for regional development, help shape transport planning for cities, and improve service provision, connectivity and accessibility.
Also see this fairly short article by an Australian-German academic.
Cost estimate
The most common cost estimate for construction of a Sydney-Melbourne VFT line (including by David Walker) is around $50 billion (roughly equivalent to construction cost for the National Broadband Network). Debt servicing costs can therefore be expected to be about $1.5 billion per year on an ongoing basis once the line is complete.
Such a high speed rail link is in my view equally a “public good” whose value should be assessed on wider public interest grounds than those typically used by the Infrastructure Commission (including general environmental and greenhouse emission benefits (certainly once in operation – diverting traffic from air which almost unavoidably runs on fossil fuels to an electric-powered VFT line). I note David’s argument that greenhouse emissions will be increased during the construction phase, but that is true of ALL new development projects involving substantial construction (including a hydro power station, for example). If we accept the validity of this argument we would be obliged to stick with all our current technologies permanently irrespective of ongoing greenhouse emissions. Very convenient for the fossil fuel industry.
David also argues that air traffic won’t be reduced even once the VFT line is operating, because there is already unmet demand for air services (especially from Mascot due to the 11pm curfew) and luring some current Sydney-Melbourne passengers onto the train would merely open up slots for additional air services to other destinations that can’t currently fit at Mascot. While that might well be true, that in itself would be a strong indicator of significant additional economic activity overall generated by construction and operation of the VFT rail line. You can’t have it both ways.
Reduction of road traffic on the already heavily travelled Hume Highway would be another benefit, as would opening up opportunities for people to live in the intermediate cities and commute to Sydney or Melbourne. It’s especially relevant for Liverpool and Campbelltown near Sydney, from where many residents already commute to the CBD each day, often taking 90 minutes or more each direction. But commuting from Goulburn or Canberra to Sydney each day, or from Albury/Wodonga, Shepparton or Bendigo to Melbourne, would become an attractive option. David dismisses this advantage as “decentralisation” and asserts without analysis that urban densification of Melbourne and Sydney are preferred options. Large numbers of Sydney and Melbourne residents would strongly disagree, and the 2013 federal government report referred to earlier finds:
High speed rail has the potential to properly unite east with west, improving access to jobs, easing pressure on housing and reducing the time and cost of long commutes. It will free up much needed space for freight on road and rail network, and trigger further improvements to public transport networks by taking some of the pressure off the furthest reaching connections and freeing up resources for the remainder of the network. It would mean Newcastle, the Central Coast, the Southern Highlands and Canberra would all be within 30 minutes to an hour of the CBD, creating considerable opportunities for people to work in Sydney and live elsewhere.
Also see Meet the countryside commuters and Andrew Tallon, Urban Regeneration in the UK (Google Book extract citing lots of relevant research, especially critiques of “densification”). Densification as a compulsory imposition, irrespective of how a significant proportion of people actually want to live (as shown by their revealed choices) is a trendy concept that has irritated me for quite a few years, as illustrated by a 2005 post here at Troppo titled creative class wankery.
Assumptions
My revenue projections are based on the following:
- The VFT line would be constructed via Tullamarine Airport, which isn’t currently serviced by a rail line. It is assumed that the existing SkyBus service from Southern Cross Station to Tullamarine would be discontinued and replaced by trains travelling on the VFT line.
- The VFT line would also be constructed via Mascot (domestic and international) and Badgerys Creek airports, thereby linking and unifying air and high speed rail travel.
- Melbourne-Sydney VFT services would stop only at the airport stations and (perhaps) Canberra and would otherwise be non-stop services. That should achieve travel times of around 3 hours. There would be separate Melbourne-Canberra and Sydney-Canberra “all stations” services to provide flexible high speed travel opportunities for people travelling between intermediate centres and Sydney, Melbourne or Canberra. The entire south-eastern corner of Australia, containing half the country’s population, would be effectively integrated by such a service. You could quickly and conveniently travel anywhere within that region for (say) a weekend family gathering or a concert or theatre event in Sydney or Melbourne.
- Flexible airfare pricing makes it tricky to calculate an average for air travel on the routes with which a VFT service would compete. But a fairly random scan of Webjet suggests Sydney-Melbourne economy prices average around $150 one way, while Melbourne-Canberra and Sydney-Canberra prices are a bit higher than that despite the shorter distance. I envisage a price structure for the VFT service where Sydney-Melbourne prices would average $250 one way, Melbourne-Canberra and Sydney-Canberra around $150 and services to intermediate cities a bit less than that (say $120). A Sydney-Melbourne CBD-CBD price of $250 would actually be very attractive and competitive with air having regard to the cost of a taxi to the airport at both ends (or parking charges at one end and a taxi at the other).
- The federal government would impose a levy of $70 on each air ticket between Sydney’s Mascot Airport and Melbourne, and a levy of $50 on each air ticket between Sydney’s Mascot Airport and Canberra. The levy would have three objectives: (1) to generate revenue for the VFT line; (2) to create price signals to induce airlines to shift existing Sydney-Melbourne and Sydney-Canberra services to Badgerys Creek where feasible; (3) to open up additional landing slots at Mascot in view of the 11pm curfew.
- I assume that the airlines would be able to shift a significant number of existing flights affected by the levy to Badgerys Creek, thereby avoiding the levy. But they would still keep enough flights at Mascot to service people travelling to and from Sydney CBD and the eastern part of Sydney. My guesstimate is that 3 million passengers would still travel via Mascot with around 1 million using Badgerys creek instead.
- Given convenience and equal net travel time, the combination of avoiding expensive taxi fares and/or airport parking charges and the proposed ticket levy would also in fact mean that Sydney-Melbourne travel on the VFT service would actually be cheaper in overall terms than economy air travel. My guesstimate on those assumptions is that the VFT service would capture 50% of the Sydney-Melbourne air travel market. It is the fifth most heavily travelled air route in the world and last year carried a total of 8.65 million passengers. The revenue projections set out below assume that the VFT line would carry 4.3 million passengers Sydney-Melbourne each year.
- I also assume (without any real calculation) that the line would carry some 2 million passengers per year between intermediate cities (including Canberra) and Sydney-Melbourne at an average ticket price of $130. Given existing high air travel prices to Canberra and the additional $50 ticket levy it is reasonable to assume that the great majority of that market would be captured. I also assume that there would be a substantial daily commuter market for the VFT service, especially from Bendigo to Melbourne and Campbelltown/Liverpool to Sydney, if tickets are priced competitively.
- The federal government would also impose a “value-capture” levy once the line is complete and operational as follows: $50 per year for every household within 10 kilometres of the Sydney and Melbourne CBDs and each of the intermediate cities. My guesstimate is that this would be around 4 million households on Australia’s current population. That would generate $200 million per year. Similarly situated commercial properties (i.e. within 10 kilometres of a VFT station) would pay $400 per year. My guesstimate of the number of commercial premises subject to levy is 300,000 indicating revenue of $120 million per year from that source. In effect, both the foregoing are very modest hypothecated land taxes. Lastly, similarly situated businesses (irrespective of land ownership) would pay $200 per year. My guesstimate of numbers of affected businesses is 500,000, indicating revenue of $100 million per year from that source. Total value capture revenue would therefore be in the vicinity of $420 million per year.
VFT revenue projections
| REVENUE SOURCE | ANNUAL REVENUE ($) |
| VFT tickets (Melbourne-Sydney) | 1,075,000,000 |
| VFT tickets (Melbourne-Tullamarine)(Skybus substitution) | 150,000,000 |
| VFT tickets (intermediate cities inc. Canberra) | 260,000,000 |
| Mascot/Melbourne or Canberra air ticket levy | 220,000,000 |
| Value-capture levy (see 9. Above) | 420,000,000 |
| TOTAL | 2,125,000,000 |
Conclusion
Contrary to claims of opponents (including David Walker), it seems to me that Melbourne-Sydney Very Fast Train service might well be viable to build and operate. It will cost around $50 billion to build, implying an annual debt servicing cost of around $1.5 billion. But revenue projections, including ticket revenue and an air-ticket levy plus a “value-capture” levy on properties and businesses which benefit from the VFT service, may well achieve total gross income of more than $2.1 billion, which should be more than enough to cover debt servicing and normal operational costs, possibly even leaving a modest net profit. Moreover, the above projections have been calculated on Australia’s current population and air travel numbers. Our population will be about 13% larger by 2026 at just over 27 million, rising to 36 million by 2050. You would expect a high proportion of those additional numbers would be living in Australia’s south-eastern corner, making a VFT service progressively more and more viable/profitable as time goes on.
Ken thanks
BTW We have had to fly Canberra Sydney on short notice , in that situation around $400 to $ 500 for a return flight is a ‘good’ price.
Like most of these big passenger transport projects, I think they suffer from the same fundamental flaw – they assume that jobs (particularly white collar jobs) and the subsequent trip generation growth will continue to be in the centres of capital cities. But it doesn’t have to be this way due to technological advancement.
If we were being a bot more ‘nimble’ and ‘innovative’, we could facilitate more of the growth in jobs within regional centres.
Run a simple test in your head. If a phone company can outsource the function (job) overseas, why couldn’t government agencies and large corporates gradually relocate those functions to regional centres when : rents are lower; regional unemployment is higher (so people are prepared to do the same thing for less $); growth pressures on capital cities are reduced; regional economies diversity; and risky investments in very fast trains are not necessary in the first place.
Only problem is, as a former NT Transport Minister told be once, “you can’t put a plaque on a good idea, but you can put a plaque on a bit of infrastructure”.
Hi Jim
Regionalisation/decentralisation of industry and commerce has a fairly poor record in Australia and elsewhere. There is a huge volume of research showing the benefits of city size and densification for commercial if not residential purposes. Humans are social creatures and a high proportion of our interactions are non-verbal. People preferring to live in pleasant suburban and country environments rather than in a dense inner city “village” raises a range of very different issues.
Despite the Internet, video conferencing is a poor substitute in complex interpersonal situations (as opposed to straightforward daily dealings where it is fine). Genuinely fast broadband may help to an extent (if we ever get it) because it would facilitate genuine high definition Internet video conferencing, but there will still be major advantages to being co-located in a major CBD.
There will no doubt be some businesses which will choose to relocate to regional centres when genuine fast broadband actually exists, there is no “one size fits all”. We shouldn’t prescriptively pursue development with a view to giving businesses no real alternative but to locate in a large city, but nor should we do the opposite. VFT plus fast broadband will maximise flexibility and choice for both business/industry and people’s living choices.
Lastly I suspect VFT isn’t a “high risk” proposition for the reasons I set out at length in the primary post. Moreover I wouldn’t support VFT unless we are assured by expert cost/benefit analysis that it stacks up.
The “value capture” levy is quite regressive, and would be better done IMO as a stamp duty style tax on property transfer. That obviously defers the revenue, but with the benefit that you’re not taxing people with a named levy to pay for something they don’t use and in many cases will have spent 10 years protesting against. Remember that The Block is within 900m of Sydney Central, and while that land is worth a lot and is currently being gentrified by force, there will still be poor people in 10 years time.
It’s also worth remembering that having politicised the sh!t out of the NBN and other projects, you can expect the Liberals to either FUBAR it if it’s their project, or sabotage it if it’s an ALP one. Those “better economic managers” would no doubt turn it in half-done, very late and grossly over budget, but somehow still managing to pay bonuses to all their private sector mates that were involved. The ALP would just settle for turning it into an electoral liability despite 90% of the population wanting it.
Leaving aside your observations about the political parties, your point about my current value-capture levy being regressive is well made. It would be better to devise a progressive land tax based on Unimproved Capital Value, with rates fixed so as to generate roughty the amounts set out in my primary post. Stamp duty OTOH is generally regarded by economists as being highly inefficient and has lots of undesirable effects (e.g. substantial barrier to labour mobility).
The ACT has a land tax on rental properties that consists of fixed charge of $945 plus sliding component on value ranging from % 0.41 on lower value land, to % 0.61 on land valued above $150,000. the ACT also has a levy on commercial properties near the CBD for “City Centre Marketing and Improvements”.
Could something like that be a option?
Yes, but you need some hypothecation to the VFT matching some plausible measure of the value the VFT contributes.
Yes it would obviously be preferable for a range of reasons if there was some formula for the levy that could demonstrate a clear and evidently fair/plausible relationship between the amount of the charge and the benefit to either land value, amenity of life or business profitability (or perhaps all three). Any ideas? A variant on Gruen’s well-being index?
Nicholas why, :-)
Also, Paul Newman has a piece in The Conversation on that very subject. He concludes that the problem is finance, but I would have thought that the very smart banking set we have in Australia, and in parliament, could sort that out very quickly. Especially with interest rates and inflation so low, and all that Australian steel looking for markets. Submarines or VFT… I know which I’d rather have.
The piece, btw, is an indictment on urban and transport planning since ww2. Experts have challenged the transport paradigm for generations but vested interests and puppet pollies have always left these real problems in the too hard basket. There always seems a element of if something can go wrong it will (with a little help) and like (complementary?) enviro/ecology policy and its elements, is left for the next generation to sort through.
There are two problems with your list of intermediate cities, excluding Canberra
(1) You don’t want to stop at many of them because you will slow down your trains a lot. If stopping 6 places causes an extra hour on your trip (10 minutes per stop), then you will go from potentially just a bit slower than planes for Melbourne-to-Sydney (under the Conrad assumptions, which are faster than other commentators think) to yet another hour of waiting (i.e., going from something where I would never think of taking the plane to something where I might well take a plane).
(2) Most of the intermediate cities you mention are not especially nice places to live and are never going to be — they are inland and hence have crappy weather and are not near beaches. That’s why I prefer smaller routes that go to places that are or could potentially be nice (Geelong; The Gold Coast; The Sunshine Coast, Gosford, Newcastle etc.). Canberra is special in this regard because people will live there no matter what due to the money and jobs that are there but it’s population is also limited for exactly the same reason.
Hi Conrad
The primary post expressly indicated that Melbourne-Sydney services would run non-stop (except airports and maybe not even that) for precisely the reasons you outline. See numbered paragraph 3. Services to the intermediate centres would be by separate “all stations” trains.
I don’t agree with you on (all of) the inland intermediate cities. Bendigo is a very pleasant city, Shepparton less so. For the intermediate NSW city I listed Goulburn but I agree it’s not especially attractive. Possibly Bowral would be a better bet. It’s the natural centre of the southern highlands and lots of people live and commute to Sydney from there some of them very wealthy. It’s a beautiful area with a lovely cool climate. I recommend a visit.
Albury/Wodonga is also quite pleasant, on the Murray, gateway to wine region and snow country, and was one of the Whitlam government’s designated regional growth centres. For example, distance from Wodonga to Beechworth – 39 km; Wangaratta – 68 km; Glenrowan – 86 km; Rutherglen – 43 km; Bright (95 km (Mt Buffalo, Falls Creek etc).
I’ve never fully understood why Australians think the coast is the only desirable place to live and anywhere inland is the boondocks. It isn’t a perspective shared by most Europeans, Americans etc. I would actually regard Geelong as the boondocks, even though it’s near the coast. It isn’t a place I would choose to live.
I would, however, also like to see shorter VFT services to the sorts of places you mention, but I think we should start with Melbourne-Sydney for the reasons I outlined at length. Sydney-Newcastle would be very expensive because of the difficult terrain. Lots of tunnels would be needed.
I’d guess that Goulburn has come up because:
it’s the current junction for the Canberra branch line ,
has a lot of old under-utilised rail yards etc,
has been earmarked for expansion as a high need aged care specialty area,
has the main hospital for the region,
has a large number of schools ( many are boarding schools),
A large TAFE,
And it has the police academy.
On the otherhand somewhere around about Moss Vale – Sutton Forest would be much more attractive , closer to the physical centre of the southern highlands ( and only about 65 K’s to Wollongong, via the Illawarra Hwy).
Sorry to have missed this, but if you still want some trains that stop at smaller cities, you still need ways to schedule them with your faster services, otherwise they will be slower and I suspect the fast trains will need to stop here and there. On the line I’m used to in France, for example, there are very few services that stop at smaller cities which I assume is because of this — perhaps this is possible by using track sections where stopping trains get on-and-off the main line, which I suspect could only happen in non-peak times if peak times mean trains are going every 5-10 minutes (which I suspect would be reasonable for Melbourne-Sydney). Of course, you could build additional parallel tracks more or less in their entirety, and perhaps that wouldn’t really cost a lot more once you have actually got the rest of the stuff sorted out (I wouldn’t have a clue on that — i.e., the cost of the tracks vs. everything else).
I don’t know why Australians only like the coastal cities (and the East Coast ones for that matter — I like Adelaide as a city but it is clearly off the map for many Australians), but it seems to be the case. I personally also quite like Bendigo and some of the surrounding places. Out of curiosity I looked up the population growth in Victoria, a fast growing state here, and the growth is pretty small excluding Melbourne (which is large), and Geelong (+5000). Curiously Bendigo has +1500 for the year so would be growing quickly given it’s population and the loss to Melbourne that would happen. Of the smaller areas, it is coastal ones people are moving to that are close to Geelong or Melbourne (plus Gippsland for reasons I am unclear of) and I think the little chunk you mention (Albury/Wodonga).
Id guess that a station at ‘ Moss Vale’ etc could also function as a passing loop.
BTW
Ken how many ( fairly full ) trains an hour would be needed for it to be economic?
Out of curiosity, I also just checked the price for Marseille-Paris, which is similar to Melbourne-Sydney (although Marseille is clearly much smaller than Melbourne or Sydney). The tickets are about $100 each way, so for whatever reason the French system is set up much cheaper than your suggestion. I also seem to remember the flight prices are not dissimilar to Melbourne-Sydney (although there is nothing like Tiger with really cheap tickets — Ryan Air goes some places). TGV is also profitable, so presumably once you have the tracks in (which were probably initially subsidized), it is pretty cheap to run — another bonus of the system.
I gather the Marseille-Paris line was built in the 1970s so they have presumably paid off the capital cost some time ago. With the sort of income projections I’m talking about (augmented by levies) the Aus VFT should be able to offer flexible pricing and substantial discounts in a careful strategic way, but not standard $100 prices until the capital cost is defrayed.
Amongst other things… For decades, France made buses and planes illegal, so ensuring that the trains were well-used.
They are pretty expensive for business travel, you can by all means book cheaply in advance but for business travel you pay much more than the equivalent airfare.. !
Which I suspect would be the case here too.
Hi Patrick
I’m not writing from personal experience, just Dr Google. Something I read said that Paris/Marseille VFT tickets could be bought for as little as 25 euros if purchased 3 months or so in advance, but might be “as much as” 80 euros ($A117) if purchased shortly before the trip. I just checked airfares on the same route for 21 April (a week away) and found that the cheapest was 219 euros ($A327).
Mind you I just did a dummy online booking for a second class ticket on the Paris/Marseille VFT(TGV) for the same date (21 April) and found it cost 116 euro. First class is 162 euro (still significantly cheaper than the cheapest airfare).
Ken, did you notice the summer temperatures this year in that part of the world?
Quite seriously, inland towns in Australian are almost certainly going to routinely be very unpleasant places to be in summer in 30 to 50 years, and I reckon without some specific boondoggle reason to live there (such that sustains a place like Las Vegas) climate change will limit the growth of inland communities. (Apart from the heat, there is the matter of water supply.)
Funny how that issue isn’t acknowledged on a blog full of people who believe in climate change…
Oh – by “that part of the world” I meant Albury- Wodonga
I am trying to get my head around the economics of it all (and I support the fundamental concept).
What would be the seating capacity ? (I would imagine 700, 10 modules of 70 seats each ???)
What would be the departure schedule ?? Every 90/120/180 minutes (3 plane loads per departure ??)
What would be the operating hours (Curfew from 2100 hours (noise ?)).
We would have to import the knowledge to design, engineer, manufacture, and construct (and even operate), since our high schools are still concentrating on woodwork and home economics. And our universities are still coming to grips with really difficult science (to wit – we will need to import knowledge to design and build (and perhaps even operate !) our submarines).
But I like the proposition that Australian labour will be used in the project, and it augers well for a resurgence for Australian manufacturing.
The highest cost would be the acquisition of land (can some existing cleared routes be used with realignment here and there ?), and the construction. I see the operating costs (including maintenance) to be relatively smaller (in the overall costings).
Perhaps some of our superannuation funds can stump up some funding, with a government guarantee behind them.
And could the line be used for freight (between passenger departure times).
There are approx 3000 prime movers moving container freight each day on the Hume Highway.
Found this:
Shinkansen have comfortable seating, convenient schedules and are almost never late. Between Tokyo and Osaka 13 trains an hour run in each direction during peak. Each train has a capacity of more than 1,300 seats.
Assuming each train operates at 90% capacity, that’s 13 trains X 1300 X 0.9 = 15, 210 pax each hour each way..
And this:
Shinkansen have dramatically changed Japanese culture. They make business trips far more efficient saving travelers an estimated 400 million hours per year.
They have made remote Japanese towns and villages accessible — stimulating rural economies. They have also dramatically improved city life. In Tokyo you’re never far away from snow, beach and onsen.
Japan has a population of 126 million people crammed into quite a small land mass. I suspect the economics and logistics of that are quite different from Australia, even if we confine it to our most populous south-eastern corner. I suspect the French example might be closer to the mark for us. The Marseille-Paris line for example apparently runs 14 trains per day with 450-700 passengers per train (presumably depending on anticipated load per journey).
I imagine Australia would probably run something like 28 trains per day Melbourne-Sydney (14 in each direction) each carrying (say) 400-500 passengers. That would achieve the projected loads discussed in the primary post and deliver a sufficiently frequent service to suit the vast majority of prospective passengers. Apparently the Marseille-Paris line runs trains every 1 hr 24 minutes on average:
However the distance between Marseille and Paris is about 770 kilometres whereas Sydney-Melbourne is about 870 kilometres. It suggests the service would need to be non-stop on close to the shortest reasonably practical route and optimal engineering so the trains can sustain as close as possible to maximum speed for as long as possible. Otherwise journey times may be too long to be competitive with air travel.
Compared to Melbourne or Sydney, Marseille is a comparatively small and relative poor city (similar in size to Adelaide if you include what would be thought of as burbs in Australia like Cassis and Aix), although it is a hub you can bounce through to get to other cities in the south or to Spain (the roads are crazy around holiday time in Euroland when all the Northern Europeans drive through). This means some traffic on the route is people going elsewhere. I would expect Melbourne-Sydney would get much more traffic (indeed Marseille used to get much more traffic but Lyon seems to now be the preferred alternative for things like International travel). This means I think the travel times are a good estimate (plus 100 ks) because the burbs of Paris are like every other big city which are hard to get through, but the traffic would be much less — as a random guess I would think Melbourne-Sydney could easily have 3 times the traffic given the number of plane flights between the two (indeed, I would expect demand to increase because it would make the trip vastly easier than messing around with planes, and people are richer here than France, especially Marseille).
Ken, I was going to write a comment about your numbers being over optimistic, for example your estimate of the density of inner Sydney and Melbourne is inflated by a factor of five. Also, based on the rail share of other cities, your Melbourne to Tullamarine revenue is also off by a factor of more than five.
However, before checking more numbers, I had an idea. How about we just subsidise air travel between capital and regional cities for a trial period to see if your list of benefits eventualises? This would be far lower risk than going all in on a fifty billion non refundable project.
Would you subsidize air travel from Goulburn or Geelong to Sydney or. Melbourne etc?
For any major project, surely it’s worth spending one percent on validating it. For a $50 billion project that is a cool $500 million. So yes, you could use large helicopters for the shorter distances and planes for the longer. After all, if it’s good enough for Bronwyn…
Love your work
I did say these were guesstimates and I’m happy to be corrected. However I have now checked the figures. Inner suburbs in Sydney and Melbourne have population densities between 6,000 and 8,000 people per square kilometre. Let’s take the lower figure to be conservative. A circle with a radius of 10 kilometres around a CBD has an area of about 342 square kilometres. Thus the combined population of the designated inner areas of Sydney and Melbourne combined is about 4.1 million. The combined population of the intermediate cities on the line is around 1 million, and the population of Liverpool and Campbelltown combined is around 400,000. Total population within 10 kilometres of a VFT station is therefore at least 5.5 million people Australia’s average household size is around 2.5 people. Thus the number of households subject to the suggested $50 per year “value-capture” levy is 2.2 million. Thus the total revenue from the residential property aspect of the levy is $110 million rather than $200 million as listed in the primary post. So my estimate was a tad high but nowhere near as inaccurate as Daniel asserts
As for my passenger number guesstimates, the French high speed rail network carried 114 million passengers per year in 2o10 on 2000 kilometres of track (several lines). The most travelled line is Paris-Lyon. Greater Paris is France’s largest city with a population of 10.5 million while Lyon is the second largest with 2.1 million. These figures do not suggest that my quesstimate of 4 million passengers per year for Sydney-Melbourne (both cities of 5 million people) is wildly optimistic. In fact it may well be quite conservative on the French figures (although their VFT service is cheaper than the ticket prices I have projected/assumed because it is quite heavily government-subsidised (around 50% subsidy).
Ken, your population estimate for Melbourne has half the population of greater Melbourne living in less than ten percent of its area including hundreds of thousands on the bay and in industrial areas. I’m sure you know enough of Melbourne to know this is not plausible.
As for the Tullamarine-Melbourne guesstimate, it was based expressly on the cancellation of the existing SkyBus service from Melbourne to Tullamarine. However I did not check passenger numbers. I have now done so. SkyBus travels from Southern Cross Station to Tullamarine (as would the VFT) every 20 minutes day and night. It carries 2 million passengers per year. In the circumstances I think the assumption that all these passengers would travel on the VFT if the SkyBus was not available is a reasonable one. However the one way ticket price is $19. I had assumed more passengers and a higher ticket price (I always buy a return ticket). Two million passenger journeys @ $19 pr journey computes to $38 million revenue, much lower than my $150 million guesstimate. So that estimate WAS wrong by a factor of four, though not for the reason Daniel asserted.
In addition, I’d suspect that given “project optimism bias”, the $50Bn might be somewhat under the likely final cost.
Therefore a measly $500M could be a most prudent investment either way.
Hi Ken,
kudos for you for putting in the effort for trying to think through possible usage projections. It seems a difficult undertaking to me. I can quibble with each of the numbers (most conspicuously, you don’t seem to have any maintenance and running costs in here!). But let’s run with it anyway.
What you essentially come up with is an internal rate of return of about 3% (2.1 billion on a cost of 50 billion, which is about 4%, but the revenue is more than 10 years in the future whilst the costs are not, making the actual rate of return closer to 3%). That is simply not an attractive rate of return. No private investor would go for it, given that private costs of raising capital are much higher than 3%. Given the one-sided nature of the risks (cost-blowouts, time-delays, new technology making this obsolete, etc.) it is not a good social proposition either. We’d be much better off investing the money in India as part of a national super fund.
So I maintain that we first should ask places like Infrastructure Australia to cost-benefit such things, and that Australia at present is probably not densely populated enough to make a high-speed rail line economically sensible.
And plenty of comments above make me think the worst as to what would happen if we’d do this. Those are the comments suggesting this project would help innovation and manufacturing in this country. They make me shudder because it implies we’d insist on buying the machinery here in Australia, costing more money and creating closed workshops…..bad ideas.
I’ve given some further thought to how you would go about devising and applying a sensible metric for assessing value capture for a VFT service. Although distance from a VFT station (the metric used for my value capture levy in the primary post) is not entirely irrelevant to the added value enjoyed by a resident or property owner in a city or town serviced by the VFT, it doesn’t correlate closely to value and is highly arbitrary as the sole measure. A cleaner or factory worker living in an unrenovated old worker’s cottage in Footscray is much less likely to want or be in a position to utilise the VFT than an economist living in Port Melbourne or a lawyer living in St Kilda. On the other hand, if the cleaner owns the cottage rather than just renting it they will gain the benefit of any increased market value flowing from the energy provided by the VFT. A relatively cheap unrenovated house in Footscray might well be an attractive purchase for a yuppie economist or lawyer looking for an investment proposition or a convenient inner urban residential toehold for themselves that they can renovate.
It seems to me that a progressive value-capture levy calculated by a metric based on a combination of income level and Improved Property Value would be a fair, reasonable and practicable system. All property owners within a 50 km radius of each VFT station (“VFT catchment area”) would pay a base annual levy amount of (say) $20 per property. That catchment area has a population of approximately 11.5 million people, and so on average household occupancy rates of 2.5 persons we are talking about 4.6 million households. Everyone in the VFT catchment area gains at least some potential benefit to property value, and some potential amenity value even if they choose or cannot afford to use it at present, and so a minimal base value-capture levy is fair and reasonable.
That calculates to base levy revenue of $92 million per year. In addition, I think there should be a levy component calculated on a progressive sliding scale on all properties having an Improved Property Value of more than a designated amount (say $1.3 million). A similarly calculated base levy and progressive levy on more valuable properties would also apply to commercial properties in the catchment area of each station. After a few years of operation it should prove possible to calculate that progressive component of the levy by reference to the increase in Improved Property Value since completion of the VFT line. For example, one might calculate the value increase attributable to the VFT service by the difference between the increase in property value in the VFT catchment area and the increases in property values in cities in the rest of Australia not linked by a VFT service. That would be a levy demonstrably based on direct increased value delivered to property owners by the government’s provision of the infrastructure.
Obviously the amount needing to be raised by value-capture levy will depend on a number of factors. On the projections used in my primary post, the VFT service is projected to earn approximately $1.4 billion annually from rail ticket sales of various types (both local and interstate). It seems to me that the projections are based on reasonably realistic assumptions. However, it is conceivable that they may be on the optimistic side. The French LGV light rail system operates on a government subsidy of approximately 50% of its overall operating, maintenance and debt servicing costs. However a significant part of that subsidy is required because of a policy decision by the French government to keep ticket prices much cheaper than I am suggesting for the Sydney-Melbourne VFT service.
What is an appropriate assumption for operating and maintenance costs for the Sydney-Melbourne VFT service? For present purposes I think an assumption of approximately $500 million per year is a reasonable one. See page 6-5 of this feasibility study into a California VFT system. Given a debt servicing requirement of $1.5 billion per year, that means the service needs to generate $2 billion per year through a combination of ticket revenue, value-capture levy and non-hypothecated government subsidy. If ticket revenue is $1.4 billion as projected, then the combination of value-capture levy and government subsidy will need to raise $600 million per year.
However not all of that would be raised by the value-capture levy. Clearly there are some public good benefits flowing from the service which benefit Australians generally and don’t deliver any special benefit to residents and businesses in the VFT catchment area itself. The service would no doubt be used to a significant extent by people from all over Australia from time to time, and by overseas tourists who don’t spend any significant amount of money with residents or businesses in the VFT catchment area itself. In addition, things like benefits to safety and traffic density on the Hume Highway and reductions in greenhouse emissions (through reducing the growth in volume of aircraft emissions by reducing the growth rate of domestic air services on the busiest route in Australia) are benefits to Australians generally but not especially to people in the VFT catchment area itself.
No doubt there are methodologies that could be used to calculate the economic benefit of those factors, but it might well be appropriate to set the rates of value-capture levy to raise $300 million with the other $300 million being delivered by direct Federal government subsidy (although perhaps with some contribution to that subsidy from the New South Wales and Victorian governments). To the extent that actual ticket sales fell short of the projected $1.4 billion annually, more would need to be raised from the combination of value-capture levy and direct subsidy.
Of course none of the above includes any amount for return on capital invested, apart from whatever is notionally quantified as the value of the public good items that cannot readily be captured/internalised. I don’t think one can get around the proposition that this is a long-term nationbuilding exercise where the initial rate of return is likely to be quite small. However, as the population grows to a projected 36 million by 2050, and as the undeniable effects of global warming become larger and clearer, patronage levels and dollar rate of return would certainly increase.
Nevertheless, I am suggesting a structure where the direct subsidy amounts from government are very modest in budgetary terms, and where the value-capture levy amounts are spread quite widely and therefore are also modest in amount. I suspect that most Australians, perhaps unlike many economists, would be happy to invest in a genuine nationbuilding exercise like this even if it doesn’t make money in the first few years. The same IMO is true of the National Broadband Network. Many of its benefits lie in building and uniting a nation and community and cannot be quantified solely in dollar terms. I suspect that many of the roads and rail lines built in Australia in the 19th century would not have stacked up on an economist’s prospective cost-benefit analysis, and probably did not generate enough to justify their construction for quite a few years. I entirely accept that there is a need for a rigorous cost-benefit analysis when assessing competing infrastructure project proposals not all of which can be constructed. However it seems to me that additional considerations apply when we are talking about infrastructure that links and unites Australia or a significant part of it.
Nation building is very important, Ken, and I agree that such projects help with that. But then it needs to be argued on that basis. Ie, what other nation building things could we do for 50 billion with a rate of return between 2 and 3 % ? Why would a fast train be better than, say, having a more sensible history curriculum or having a network compulsory national community service? That is then the real debate.
You can’t first argue on the basis of dollars, getting people interested in the technical details, and then suddenly throw in your real reasons for wanting this that are entirely different. That is misleading and makes for ingenuous debates. We are not parliamentarians here!
Possibly disingenuous rather than ingenuous. But I don’t think so. You have to be able to demonstrate that a project is at least close enough to viable that it isn’t a completely idiotic idea at the outset. If that can’t be demonstrated then any such idea should never get to first base. The primary purpose of this post is to demonstrate that the Sydney-Melbourne VFT proposal reaches that initial threshold of rationality. I agree that beyond that point one needs to debate and assess the proposal against other nationbuilding ideas that will cost significant amounts of money. That debate cannot be conducted in solely economic terms, it necessarily involves non-monetary values and aspirations as well. However there needs to be a baseline set of objective facts, figures and assumptions in which to ground that debate, otherwise it will be just rancorous emotive bleating.
BTW the present rate of return to the new Chinese owners of the Darwin Port is almost exactly 3%. I doubt that they have made the investment for touchy-feely nationbuilding reasons (at least not Australian nationbuilding). Presumably they foresee a large long-term return on investment even though the short-term income is minimal. I see both high speed rail and the National Broadband Network in a similar way.
The Chinese have so much foreign reserves that they don’t know what to do with it. We’re not in that situation! But hej, if we can fool the Chinese to build the railway line, by all means….
The Chinese love engineering and technology – concrete ,steel, robots and particularly bridges.
Congrats, Ken. This has been an honest week’s work.
The Oz has a interesting piece on this:
Think the key phrase is ” use value capture to pay for big-ticket infrastructure projects ” not just a VFT.
Indeed. As I canvassed in my initial post on this topic the proposition that value-capture levies might be used for a range of projects which may be very viable and worthwhile in the long term but don’t generate enough initial return on capital to be built without an initial legup:
I would point out that major projects worldwide mostly are completed over budget and over time. The fact that the direction of the errors in project estimation are overwhelmingly in that negative direction rather than distributed about the mean points to a systematic or cognitive bias in project management. The phenomenon is termed “Project Optimism Bias”. Well explained here by HM Treasury:
https://www.gov.uk/…/attachment_data/file/191507/Optimism_bias.pdf
There are several ways to deal with optimism bias. Usually, private firms (often unconsciously) do it by requiring internal rates of return of round about 10% or more. HM Treasury in the paper above do it by making a calculation of how much the estimate needs to be increased to counteract the bias, and then the usual project evaluation criteria are applied to that increased figure.
I do suspect, as commenter Daniel has suggested elsewhere in the thread, that the same outcome of ease of travel along that corridor could be achieved more cheaply, and just as effectively by other means.
It’s not just whether the economics of a fast train would stack up, but whether it would stack up against cheaper alternatives that achieve the same or even better outcomes.
Wow. Now I know what it’s like to be in a meeting of the real characters from Utopia. And you lot disparage the NBN for being poorly thought through.
Might as well call the train line the Acela and be done with it. If this is the self-congratulatory elitist wankery you occupy yourselves with in your think tanks and discussion groups feeding policy to Labor, no wonder the Shorten era is so underwhelming.
Paul, at the risk of having my membership confirmed in Club Elitist Wankery, can you just remind us which part of the discussion here is so terrible, and why. Ken’s view is a little different from mine, but the discussion here doesn’t seem all that unreasonable.
Ken, useful discussion. Allow me a couple of notes:
* You write that …
To be fair, I’ve written a long separate post on why decentralisation is over-rated, and I linked to that post from my most recent HSR post. In fact, the decentralisation post was written largely so I could refer to it in debates where people started talking about decentralisation as a great policy aim.
* You suggest that I might be having it both ways by suggesting that a VFT won’t reduce emissions but that it won’t be economic, even though there’s obvious unmet demand for more Sydney-Melbourne corridor trips. The potential for double-counting is real here under certain constraints. But I’d point out that: a) the argument is not over whether a VFT would create economic activity, but over the size of a VFT’s net economic benefits, including externalities; and b) plenty of environmentalists seem to suggest that a VFT would be a great way to lower emissions, rather than “a great way to lower emissions slightly more slowly, in the period after 2035, than a dramatic expansion of Sydney-Melbourne corridor commercial jet flights”. The latter is more accurate.
* On value capture, note that the best study I have found seems to be the 2012 study by the University of Sydney’s David Hensher and two co-authors. Its findings suggest that we don’t really know yet how or to what extent value capture works. (There must be better studies – can someone point to them?)
* A bunch of work has already been done in the HSR Phase 2 report which gives what seem like pretty good first-cut answers to a number of questions being asked here. For a 534-page document, it’s pretty easy to work through. If we’re going to discuss whether the scheme will work or not, it seems to me that the best starting point is to take this common national reference point of the Phase 2 report and argue over where its numbers might be under or over.
It sounds like the inland freight proposal bay get the nod
Twitter discussion extract. Twitter discussion extract Everald Compton has campaigned for the Melbourne-Brisbane freight railway for around 30 years. He’s now 85 years old. A great Australian.
Second to that.
I’ve often wondered whether more value would be ‘captured’ by, instead of plonking the stations in existing reasonably-established cities such as Albury-Wodonga, putting them halfway between Albury-Wodonga and another existing city like Wangaratta that is also well serviced by the Hume Freeway. If ‘nation-building’ was the real objective, you’d probably get more bang for your buck doing this too; while people in the exisiting cities would still benefit, even if they had to sit in their car (or a bus) for half an hour on the Hume to Barnawartha/Springhurst/Chiltern, you’d be turning now-sheep farms into cities of the future on the Murray valley, and < 100kms downstream from Hume Dam. Barnawartha is a logistics hub for Woolworths if I recall, and I'm guessing they didn't use a dartboard to choose it.
And if the sheep farmer decided he/she needed a squillion to part with enough land for a new city, play him/her off with someone 10kms down the highway. Surely that must be the essence of value capture – rather than transferring the value entirely to the existing rate-payers of Albury-Wodonga?
I'm picking on Albury-Wodonga because I'm familiar with the area, but I'm sure there would be parallels say between Bendigo and Castlemaine or what have you. Maybe halfway between Albury and Wagga would make some sense too.
The U.K. Has form for doing this sort of thing with airports – East Midlands is roughly equidistant to Nottingham, Leicester and Derby from memory, but is in a paddock (or was 15 years ago when I last flew out of it).
Actually scrub Bendigo-Castlemaine -> proximity to water is a big issue if you’re building cities of the future. Besides, that track is a bit far west. You might think of somewhere east of Seymour (towards but downhill from Lake Eildon) for your last stop before Melbourne.
It would be nice to see the Kerrigan’s place at Bonnie Doon finally get some capital appreciation to match the serenity, too!