Why cutting bus fares is a bad idea

A bus at Wyndham Vale station (via Thebusofdoom)

Recommendations 45-47 of Infrastructure Victoria's 30-year Strategy deal with various changes to public transport fares. Some of these are good ideas, but others very much aren't, so let's take a closer look at the problems they're trying to fix, and why I disagree with some of their solutions. 

First: the good stuff. Recommendation 45 is to adopt permanent off-peak discounts for PT fares. Designed to encourage people to travel at off-peak times, this is fundamentally a good idea. Generally, our transport infrastructure is stretched to the limit at peak times, but has plenty of spare capacity at off-peak times; adding extra peak capacity can be a hugely expensive exercise, requiring big infrastructure projects like the Melbourne Metro Tunnel, whereas encouraging people to defer their trip into the city till after 9am doesn't cost anywhere near as much. 

Obviously not everyone can defer their travel, but any downward pressure helps. And of course the quality of service needs to be there to support this - generally, our trains in particular run at much lower frequencies outside peak times, and the longer wait times can make off-peak travel an inconvenient option, so it's great that IV recommend improving off-peak service levels elsewhere in the document. But assuming those other aspects line up, it's a good idea to use price signals as a nudge to defer travel. 

Flattening these peaks can defer big infrastructure spending (via IV)

Also in Rec 45 is the suggestion to abolish myki passes, partly because the fixed price of the passes means pass-holders don't get the advantage of off-peak discounts. This is a mistake because it ignores the bigger picture - passes may dull the effect of the off-peak discounts, but they're important for mode shift. 

If a person drives sometimes and takes PT sometimes, and they have to choose between the two for a given trip, having a pass makes the PT effectively free - which is a way stronger price signal than even the 50% discount IV are proposing. Again, service levels are generally more important than price in shifting behaviour, but to whatever extent price signals matter, free PT is way stronger than half-price PT. And this matters even if we're laser-focused on infrastructure expenditure and ignoring other things like carbon emissions; shifting car drivers onto PT defers or eliminates the need to build expensive road projects. So passes should definitely stay. 

The point they make about those on low incomes not being able to afford upfront purchases of long-term passes is a fair one. But this can, to some extent, be ameliorated by smart use of caps - eg if a 7-day pass costs as much as 5 daily tickets, the system should automatically and retroactively grant you a 7-day pass as soon as you buy 5 individual tickets in a single week. Apparently myki is capable of this but it's never been switched on; it should be. 

IV want to abolish the Free Tram Zone, and so do I (via PTV)

Recommendation 47 is to abolish the Free Tram Zone. They point out it leads to overcrowding, it's drawing people away from active transport (not cars), it's inequitable, it doesn't attract or retain tourists, and it doesn't solve the underlying problems with the fare system. I completely agree with all of this! 

Recommendation 46, though, is where the meat of their fare reform proposals are. They quite rightly note that buses (and to a lesser extent trams) are under-used compared to trains, and that it's cheaper to add more trams and buses than to add more trains. They also quite rightly note that a short bus trip to the local shops can cost the same as a long peak train into the city, and that this can act as a deterrent. They also note that "buses are overwhelmingly used by people on lower incomes", which is true but doesn't imply what they think it implies. 

So, mostly accurate observations of the problems - but their diagnosis of the underlying issues and their proposed solutions are deeply problematic, and here's my theory as to why. 

The median of the Eastern Fwy, intended for a rail corridor (via Diliff)

One of the most common types of analysis a transport economist might conduct is to compare alternatives for a government to implement on a given corridor. Let's use a real-world example: when the Eastern Freeway was built, they left provision for a railway line to Doncaster to be added in the median, but at this stage the Doncaster area is still served by rapid buses that run in general traffic (ie amongst the cars and trucks) along the Eastern and Hoddle Street. 

A transport economist might consider whether it would be best for the government to build that railway down the median; build a dedicated busway; use that median space to add extra general traffic lanes; and so on. They will weigh up construction costs, operation costs, economic value derived from quicker journeys, cost recovery from selling tickets, etc. This will involve modelling how many people will use the new infrastructure, and the level of price elasticity - that is, how much more or less people will generally be willing to pay for a better or worse service - and so they might assume that people will be willing to pay more for a faster train than they would for a slower bus. In these scenarios it often makes sense to lower prices slightly on an "inferior" product, in order to attract more people onto it and sell more tickets. 

This kind of analysis generally makes sense if you're advising a government making an infrastructure decision - sometimes building a railway genuinely would be excessive and it'd be a better use of taxpayer dollars to build a nice busway (1). But it seems as though IV have shoehorned this government-level analysis into an individual-level situation, and it doesn't really work. Why not? Because individual travellers don't get to decide where the bus goes! 

Melbourne's buses mostly run perpendicular to railways, as they should (via PTV)

These analyses work well to compare direct competitors, but in Melbourne the buses generally aren't direct competitors with the trains - they complement and feed the trains! Most routes either run perpendicular to the trains, or otherwise go places that the trains can't go - feeding people from housing into the train lines, or feeding train passengers to suburban shops or universities beyond walking distance from the rail network. So it's not often that a passenger even could chose a bus as a direct substitute for a train. Which is broadly the way it should be - it's one thing to have different modes competing with each other in a business case before a decision is made, but quite another to have them competing out in the real world, where they're duplicating costs and cannibalising revenue from each other. 

That said, if they want to use buses to take pressure off the trains, there's nothing really stopping them from selectively adding radial bus routes that do this, where it makes sense to - but the key to getting people to actually use them would be providing a fast and frequent service, probably with dedicated lanes and traffic light priority. Cheap fares won't be sufficient to get people onto a crap bus, and won't be necessary to get people onto a good bus. 

It's worth noting that in one of the key scenarios where we need change, we have definitive proof that cheap bus fares won't shift behaviour. Too many people drive their cars to the local train station, and we want to get more of them onto the bus; but if Person A drives to the station then takes a train into the city, and Person B takes the bus to the station and then takes the train, they'll both pay exactly the same fare - whatever ticket covers train travel in that zone will also cover the connecting bus. If the bus being free isn't enough to entice people onto it, price clearly ain't the problem. 

Tarneit's carpark is at capacity, despite buses being effectively free

What do we actually need to do, to encourage more people onto buses? Most of what we need has nothing to do with fares, and to IV's credit, has actually been proposed elsewhere in the strategy: provide better services that can provide a quick and convenient alternative to driving; reallocate road space to facilitate this; price congestion; price parking at stations. Do these things, and watch bus patronage soar. 

But one of the problems they observe does genuinely reflect an underlying problem with the fare structure, and that's how flat it is. As they note, taking a short bus ride to the local shops can cost as much as a train from Pakenham to the city in the morning peak - which is genuinely insane, and is a result of successive governments fiddling with the fare system to win votes from suburban commuters. But this isn't a problem with the buses themselves, it's a question of distance - taking a short train trip from Cardinia Road to Pakenham to visit the shops should be cheaper, too. So differentiating by mode isn't the solution - making fares more proportional to distance is. 

A rough idea of what a six-zone fare structure in Melbourne could look like

Some have proposed having a purely distance-based system, eg a certain number of cents per km travelled, but this gets complicated and messy and unfair very quickly. We should stick with a zone-based system, we just need a lot more zones - with sufficient overlaps so that people travelling short distances pay a fair amount and aren't penalised by living near a zone boundary. It might be something I can flesh out in a dedicated post later, but just to give you an idea of what I mean - a city the size of Melbourne needs at least 5 or 6 zones, and arguably more. (London is about two thirds the size of Melbourne (2) and it has 9 zones.)

Infrastructure Victoria have made insightful observations about some of the problems with Melbourne's fare system (3). Some of these they have correctly diagnosed and suggested good solutions for - having permanent off-peak discounts and abolishing the Free Tram Zone are definitely the right moves. But they've fundamentally misdiagnosed the causes of low bus patronage and have therefore given the wrong solution to the problem; real fare reform would remain mode-agnostic but be much more proportional to distance. And given that many of these problems have arisen because of political tinkering with the fare system, an independent body like Infrastructure Victoria is exactly the right body to start that conversation with the government. 

1. To be clear, this was just the most illuminating example I had to hand, I'm not actually advocating for a busway instead of a railway for Doncaster. The busway the government is planning to build on the Eastern is an attempt to greenwash the North East Link project, not a genuine attempt to provide good PT to Doncaster. The buses actually flow fairly quickly in general traffic on the Eastern, but grind to a halt on Hoddle St; the proposed busway ends at Hoddle St, so it won't do anything to fix this. 

2. Chesham-Upminster is ~60km as the crow flies, while Sunbury-Pakenham is ~90km. 

3. I say Melbourne's, rather than Victoria's, because there are also huge problems with V/Line's fare system that they don't even mention. A topic for a future post, perhaps.

More discussion on IV's 30-year strategy here