Mythbusting the costs of SRL

The Suburban Rail Loop (via PBO)

There's been a lot of political movement on SRL lately, so I thought it was worth a follow-up to my post from last year, to go through some of the myths and misconceptions that have arisen around the project. Today in Part 1 I want to explain why most of the media reporting on the PBO costings is dead wrong, and in Part 2 I'll talk about the political realities of campaiging on a project like this. 

The big headlines this week have centred around the Parliamentary Budget Office (PBO) costing of SRL, done at the behest of the Coalition. The main idea of the PBO is to independently cost policies for non-government MPs (whether the opposition, minor party or independent) who don't have access to Treasury, so that they can put more credible policy platforms together. They are genuinely independent, but they also have limited resources, so you do need to take what they say with a grain of salt - what they provide is usually a decent ballpark figure, but if they come up with something that sounds way off what you were expecting (or way off previous estimates), it's worth digging deeper into the assumptions they've made and the way they've calculated it. 

If you do dig into the numbers, you'll find something very unusual - not a mistake that the PBO's made, but a quirk of what the Coalition requested. See, this costing is not just taking a look at what it would cost to build SRL between now and 2035 (just SRL East) or 2053 (SRL East + North), it is estimating the costs to build, maintain, and operate SRL between now and 2085 - more than thirty years after construction ends. While this is not an inherently unjustifiable question to ask, no other costing that gets released about public infrastructure projects ever includes the ongoing maintenance and operational costs - so if you just take the headline figure of $200 Billion, you're not comparing apples to apples. (Not unless you also start including the costs of maintaining and operating the Monash Freeway for the next sixty-odd years, anyway). 

The PBO's costings show asset expenses for renewal as well as construction

What's more, the PBO has split this $200B figure into $125B asset costs and $75B operational costs. A naïve reading of this might suggest that the "asset" costs are the upfront costs to build it, and the "operational" costs are everything after that point - and that seems to be the shortcut most journalists and commentators are taking. But no. The asset costs include "renewal" of the infrastructure and rolling stock, which is why if you look at the graph of their costings, some of the biggest asset costs take place in the 2080s, half a century after SRL East opens. 

So let's make this abundantly clear to everyone in the state: the Suburban Rail Loop will not cost $125 Billion to build. 

How much will it actually cost to build? The PBO has not provided a figure for this, nor have they provided sufficiently detailed costings for us to figure it out for ourselves. However, they estimate that the total cumulative asset costs by 2053 (when SRL North reaches the airport) will be just under $83B. So right off the bat we can assume that at least $42B of the $125B (a third of the total) is asset renewal, not construction, because it takes place from 2053-2085, after construction has finished. Now - by 2053, SRL East will be 18 years old, so that $83B figure probably includes at least a little bit of asset renewal as well, but it's not possible to disentangle this any further based on public figures, so let's just consider that to be an upper bound. 

The nearer-term figures are much less rubbery. The PBO estimates construction and operation of SRL East to 2035 will cost $36.5B. This is about 5% higher than the government's estimate of $34.5B, and of none of these estimates are accurate down to the last cent, you expect a few percent variation depending on how you tweak some core assumptions. So basically, on the first stage, the PBO has essentially said the government got the costings right. Hardly the "blowout" some are calling it. 

Concept image of an SRL train (via Big Build)

I don't think the PBO has done anything deliberately wrong here. They've got to cost what they're asked to cost, so they had to include everything out to 2085 like the Coalition asked. I do think they've done the public a disservice by presenting the information the way they did, and not being particularly transparent about their workings out - which is an ongoing problem, unfortunately. But I'm going to paraphrase Hanlon's razor here, and attribute this to naivete and underresourcing rather than bias or malice. 

The Coalition have clearly crafted their request to get the biggest number they could, and have just shouted the headline figure as much as possible. They aren't going for nuance here, they're going for venom, and will do everything they can to make this project out to be a waste of money. More on that in Part 2. 

A political party presenting information in a way that best serves their purposes is always disappointing but never surprising. But I have to say, the bigger disappointment in all this has been the media. Some outlets seem to have been running a concerted campaign to undermine SRL for several months now, so of course a "cost blowout" narrative only furthers that agenda. But the real disappointment is seeing other, more neutral outlets parroting the same lines, when it's their job to do exactly the kind of close reading of the figures that I've done in this blog. Sadly I fear the same underresourcing that afflicts the PBO afflicts too many of our journalists, and it doesn't help when many of the "experts" they go to for comment clearly have an axe to grind, and aren't going to tell them the truth about these figures. 

It's fair enough for us to have a debate about whether SRL is the right project, whether it's being done in the right way and at the right time. But we should at least be basing that debate on the facts. And the fact is, the PBO is not projecting SRL will cost a lot more to build than the government did.