The Victorian state government has got itself in a bit of a bind, entirely of its own making.
For many years it built up expectations that there'd be "more trains across Melbourne" when the Metro Tunnel opened.
"There will be a new timetable in place everywhere - on our trains, on our buses and on our trams" the premier said on October 7, 2025. Use of the term "Big Switch" further fuelled expectations for February 1, 2026.
Big Switch for some lines only
But on February 1 we ended up getting substantial timetable uplifts on just three out of sixteen lines - all going through the Metro Tunnel itself. This is very significant, creating Melbourne's first frequent corridor first to last train.
Elsewhere though it's a bit of a let-down. Three lines got minor uplifts while four will have improvements later in the year. Two of those four (Craigieburn and Upfield) will have their maximum waits cut from 40 to 20 minutes (good) but miss out on the 10 minute daytime service proposed in the 2016 Metro Tunnel Business Case (bad).
Elsewhere though it's a bit of a let-down. Three lines got minor uplifts while four will have improvements later in the year. Two of those four (Craigieburn and Upfield) will have their maximum waits cut from 40 to 20 minutes (good) but miss out on the 10 minute daytime service proposed in the 2016 Metro Tunnel Business Case (bad).
Sandringham gets very good weekday interpeak upgrades (from every 15 to every 10 min) but Sunday mornings will remain only every 40 minutes and South Yarra loses its 7 day through service to North Melbourne, Footscray and Newport that it enjoyed when the Frankston line performed this function.
Opportunities missed and opportunities future
It is robust operational practice to run trains in independent groups corresponding to City Loop portals, cross-city and Yarra (the latter apparently being the internal name for the Metro Tunnel group). It is good planning practice to introduce new timetables in stages by group (rather than all in one go like the "Big Switch" rhetoric had people believe). But it is also clear that many opportunities to upgrade by group were missed in the last decade, particularly for the northern, Clifton Hill and Burnley groups.
It is robust operational practice to run trains in independent groups corresponding to City Loop portals, cross-city and Yarra (the latter apparently being the internal name for the Metro Tunnel group). It is good planning practice to introduce new timetables in stages by group (rather than all in one go like the "Big Switch" rhetoric had people believe). But it is also clear that many opportunities to upgrade by group were missed in the last decade, particularly for the northern, Clifton Hill and Burnley groups.
For instance the government had over a decade to close 30 to 40 minute gaps on nine lines (Hurstbridge, Mernda, Craigieburn, Upfield, Belgrave, Lilydale, Alamein, Glen Waverley, Sandringham) but has hardly lifted a finger to do so.
The best time to act was then. The next best time to act is now. I discussed potential future priorities for that here. The minister herself said the 1 February changes were "only the beginning" of service uplifts, giving room for cautious optimism.
Certain objections can be dismissed as false (eg lack of trains or level crossings) while others (eg train driver numbers) can be worked around with sufficient advance planning or careful rostering (eg for shoulder peaks). For example a mere 40 trains per week could get rid of 30-40 minute evening and Sunday morning gaps for the whole week from 7am to 10pm on the Mernda line. Around 300 trips per week could cut the 30 min gaps on the Belgrave and Lilydale lines while also benefiting busy stations like Box Hill and Ringwood.
For context Metro currently run about 18000 trains per week with the "Big Switch" adding about 1200 services per week. So we're not talking about unreasonable increases here, provided that the will exists to do it.
Initiative, costs and value for money
The best time to act was then. The next best time to act is now. I discussed potential future priorities for that here. The minister herself said the 1 February changes were "only the beginning" of service uplifts, giving room for cautious optimism.
Certain objections can be dismissed as false (eg lack of trains or level crossings) while others (eg train driver numbers) can be worked around with sufficient advance planning or careful rostering (eg for shoulder peaks). For example a mere 40 trains per week could get rid of 30-40 minute evening and Sunday morning gaps for the whole week from 7am to 10pm on the Mernda line. Around 300 trips per week could cut the 30 min gaps on the Belgrave and Lilydale lines while also benefiting busy stations like Box Hill and Ringwood.
For context Metro currently run about 18000 trains per week with the "Big Switch" adding about 1200 services per week. So we're not talking about unreasonable increases here, provided that the will exists to do it.
Initiative, costs and value for money
DTP bureaucrats, unlike business entrepreneurs, operate in a context of constraint and limitation. Their budgets are fixed and they have little control over most revenue. Those of the Paul Younis mould appeared not even to care much about preventing losses (even denying the extent of bus fare evasion) or have much of a growth mindset such as could be applied to finding efficiencies, growing services, marketing or boosting patronage.
Despite this there have been some improvements to bus services and large boosts to V/Line rail services. But, Metro Tunnel's three lines excepted, service development and reform for Melbourne's busiest two modes - train and tram - has been slow.
So there's been almost a stalemate. A cold shoulder for service if you like.
Government says we're broke.
While DTP may not have been as forthcoming in championing potential reforms or efficiencies that could lead to better service as it might have been. That's despite growth in the number of executives who should be creating many times more value than they draw in pay.
It remains amazing that over 25 years on, service uplifts under Jeff Kennett (a controversial premier with a reputation for cutting public services) still compare favourably with what recent government have delivered, at least for metropolitan rail.
In an election year one solution is for sufficient pressure to be brought to bear on government so it is forced to find a way. Over the heads of the stodgier bureaucrats if necessary, not unlike what it did with the Suburban Rail Loop. Opportunities may also exist for government to use its monopsony powers to negotiate a sharper deal with would-be rail franchise operators in the soon to start MR5 deliberations.
In 2024 I looked at how much public transport cost to run. This was done by dividing what we paid operators by annual kilometres operated to get a dollars per kilometre number.
Metro train worked out at $55 per service kilometre. That's a mean number. There will be substantial variations. Understanding this is important if you want to look at what is and is not cost-effective to add. This is because a kilometre's extra service at certain times will far exceed $55 while at other times it will be much less.
There are certain understood facts about rail operations. Firstly that as it has its own infrastructure (unlike a trucking company) some costs remain fixed regardless of whether you run few or many trains. Whereas other costs do rise with the number of trains, some less than proportion, some in proportion and others disproportionally (latter most notably peak service, especially if it requires new trains, signals and tracks).
Trains need drivers and other staff. Their conditions, including working hours and pay, are set down in the MTM Enterprise Agreement 2023. This stipulates what is and is not an acceptable work roster with implications for scheduling and thus the efficiency and costs of particular timetables and service patterns.
Pages 74-85 has conditions for drivers. Notable features include minimum 12 hours between shifts, time and a half pay on Saturdays, double time on Sundays and at least 3 Sundays off per 8 weeks. There is 76 hours or ordinary full time work with a requirement to work "reasonable overtime" (paid at the appropriate rates). There is some provision for part-time drivers with the number capped at 50 new recruits under the agreement. However existing full time drivers can convert to part time or job share with mutual agreement. Part time drivers work a minimum 4 hours shift adding to 40 hours per fortnight minimum. Part time drivers cannot be rostered more than 1 shift per calendar day - ie no split shifts.
That was a lot of words. What are the cost implications if you wanted to add service at particular times? And will you get much benefit in terms of increased patronage?
Apart from the $55 per km average, I don't have the actual costs of adding Metro train service kilometres at various times. But the graph below I think is a fair stab. It's not precise enough to interpolate Y axis numbers (so don't!). But it would be fair to say that there are high and low cost types of service additions that I will discuss.
As noted above peak period service is the most expensive to add for reasons including total staff numbers, train numbers, signals, infrastructure and even tunnelling. Especially in the peak direction. Sometimes adding peak capacity is needed when trains are crowded - this is a common rationale for major projects such as the Metro Tunnel and level crossing removals. But if trains are already frequent and they are not overloaded then the gains from adding service are relatively small.
Night Network comprises one train per hour on each line for about 6 hours in the wee hours of Saturday and Sunday. However for that stations need to be staffed over hours they were previously closed. Hence running that also has a high marginal cost.
Embrace the shoulders
If the train system was an Aldi the off-peak frequencies would be in the centre aisle with all the other bargains. Off-peak weekdays is the cheapest, with Saturdays and Sundays somewhat more due to penalty rates.
Depending on line trains may run their peak of peak frequency for only 2 hours before falling off. However many drivers may still be rostered, noting that a. most drivers are full time, b. there are no split shifts and c. part-timers are rostered on for a minimum four hours.
This gives rise to a potential 'overlay' where it may be possible to run more trains in the weekday shoulder peaks for basically negligible cost in the time bands shown in grey below (again this graph is theoretical).
Despite this there have been some improvements to bus services and large boosts to V/Line rail services. But, Metro Tunnel's three lines excepted, service development and reform for Melbourne's busiest two modes - train and tram - has been slow.
So there's been almost a stalemate. A cold shoulder for service if you like.
Government says we're broke.
While DTP may not have been as forthcoming in championing potential reforms or efficiencies that could lead to better service as it might have been. That's despite growth in the number of executives who should be creating many times more value than they draw in pay.
It remains amazing that over 25 years on, service uplifts under Jeff Kennett (a controversial premier with a reputation for cutting public services) still compare favourably with what recent government have delivered, at least for metropolitan rail.
In an election year one solution is for sufficient pressure to be brought to bear on government so it is forced to find a way. Over the heads of the stodgier bureaucrats if necessary, not unlike what it did with the Suburban Rail Loop. Opportunities may also exist for government to use its monopsony powers to negotiate a sharper deal with would-be rail franchise operators in the soon to start MR5 deliberations.
In 2024 I looked at how much public transport cost to run. This was done by dividing what we paid operators by annual kilometres operated to get a dollars per kilometre number.
Metro train worked out at $55 per service kilometre. That's a mean number. There will be substantial variations. Understanding this is important if you want to look at what is and is not cost-effective to add. This is because a kilometre's extra service at certain times will far exceed $55 while at other times it will be much less.
Trains need drivers and other staff. Their conditions, including working hours and pay, are set down in the MTM Enterprise Agreement 2023. This stipulates what is and is not an acceptable work roster with implications for scheduling and thus the efficiency and costs of particular timetables and service patterns.
Pages 74-85 has conditions for drivers. Notable features include minimum 12 hours between shifts, time and a half pay on Saturdays, double time on Sundays and at least 3 Sundays off per 8 weeks. There is 76 hours or ordinary full time work with a requirement to work "reasonable overtime" (paid at the appropriate rates). There is some provision for part-time drivers with the number capped at 50 new recruits under the agreement. However existing full time drivers can convert to part time or job share with mutual agreement. Part time drivers work a minimum 4 hours shift adding to 40 hours per fortnight minimum. Part time drivers cannot be rostered more than 1 shift per calendar day - ie no split shifts.
That was a lot of words. What are the cost implications if you wanted to add service at particular times? And will you get much benefit in terms of increased patronage?
Apart from the $55 per km average, I don't have the actual costs of adding Metro train service kilometres at various times. But the graph below I think is a fair stab. It's not precise enough to interpolate Y axis numbers (so don't!). But it would be fair to say that there are high and low cost types of service additions that I will discuss.
Night Network comprises one train per hour on each line for about 6 hours in the wee hours of Saturday and Sunday. However for that stations need to be staffed over hours they were previously closed. Hence running that also has a high marginal cost.
Embrace the shoulders
If the train system was an Aldi the off-peak frequencies would be in the centre aisle with all the other bargains. Off-peak weekdays is the cheapest, with Saturdays and Sundays somewhat more due to penalty rates.
Depending on line trains may run their peak of peak frequency for only 2 hours before falling off. However many drivers may still be rostered, noting that a. most drivers are full time, b. there are no split shifts and c. part-timers are rostered on for a minimum four hours.
This gives rise to a potential 'overlay' where it may be possible to run more trains in the weekday shoulder peaks for basically negligible cost in the time bands shown in grey below (again this graph is theoretical).
Improving frequency around 10am, 3pm and 7pm weekdays would provide a gentler drop-off outside the peaks that would especially assist those with flexible working hours and potentially even relieve peak pressure. It would also make plugging the remaining hole around noon small in the number of services required.
Lines with high potential here could include Craigieburn, Mernda, Belgrave, Lilydale and a couple more like Werribee and Glen Waverley.
Sunday mornings and weekend evenings
My rough graph before had weekend service as being dearer to add than weekday service. However that doesn't mean that it should be ignored.
One advantage of very low frequencies at one particular time of the week (eg Sunday mornings) is that only a handful of extra trains need to be added to remove annoyances that prevent the network having a useful service all week.
As an example just 4 extra return trips per week per line would close 40 minute Sunday morning gaps to 20 minutes in the 7am - 10am range on lines like Mernda, Hurstbridge, Sandringham, Craigieburn and Upfield. A similar number would get Belgrave and Lilydale service from 30 to 20 minutes, with benefits for a large catchment including densely populated centres such as Box Hill.
Sunday mornings and weekend evenings
My rough graph before had weekend service as being dearer to add than weekday service. However that doesn't mean that it should be ignored.
One advantage of very low frequencies at one particular time of the week (eg Sunday mornings) is that only a handful of extra trains need to be added to remove annoyances that prevent the network having a useful service all week.
As an example just 4 extra return trips per week per line would close 40 minute Sunday morning gaps to 20 minutes in the 7am - 10am range on lines like Mernda, Hurstbridge, Sandringham, Craigieburn and Upfield. A similar number would get Belgrave and Lilydale service from 30 to 20 minutes, with benefits for a large catchment including densely populated centres such as Box Hill.
Similarly weekend evenings requires just two return trips per line to push the typical 30 minute frequency start back from 7:30pm to 9:30pm.
As both these examples are only one or two days of the week they involve a relatively small number of extra trips added per week, reducing their cost compared to if they needed to be run every day.
Both would go a long way to getting most of the network to run every 20 min or better from say 7am to 10pm. Not Sydney service standards but substantially better than current for relatively few trains and driver shifts added.
Conclusion
As both these examples are only one or two days of the week they involve a relatively small number of extra trips added per week, reducing their cost compared to if they needed to be run every day.
Both would go a long way to getting most of the network to run every 20 min or better from say 7am to 10pm. Not Sydney service standards but substantially better than current for relatively few trains and driver shifts added.
Conclusion
Embrace the shoulders. They're the best bargains around if you want to boost Metro train services. Then do Sunday mornings and early evenings to make the network even more useful.
No comments:
Post a Comment