From the Red Line
Published in

From the Red Line

The writing on the wall

If you didn’t see the frequency-ridership spiral, you better see it now.

It appears that the total termination of night bus services has triggered a variety of reactions. Bus enthusiasts react with shock and horror as they are wont to do with any service cancellation, whilst partygoers interviewed by TODAY newspaper basically replied with “wait, we had those?” and night shift workers said that their employers typically arranged transport for them anyway.

Amusing as it might sound, the reasons for failure were presumably the writing on the wall that should have been seen long ago. Last time I discussed how the situation could be improved, but in the light of what appears to be a full retreat from the nightlife sphere, we need to look at how we even got here in the first place.

Counting the casualties

Newspaper excerpt, emphasis mine:

TODAY also spoke to several young people, aged 21 to 28, and found that most were not aware that these Night Rider services were available before the pandemic.

Ms Sarah Lee, 25, an undergraduate, said: “I took the bus once in 2018 because I was headed home after some drinks and decided to try it. I didn’t really take it again because I’d rather just pool and take a taxi with my friends. It was just slightly more expensive to do so but at least can get home directly.”

This smells like bad publicity. Yes, on one hand it might be possible that two years of suspension means that everyone forgot about these bus services. But the proof is in the pudding.

The question we have to ask ourselves is, where was the notice of resumption when nightlife businesses were allowed to reopen on April 19? Sure, there may be some hesitation especially with capacity limits still being applied for a while and fear of a policy reversal. But that hasn’t stopped the partygoers, and if there were any significant increases in the accident rate, we would have seen them by now.

(edit 28/6: realized I forgot to finish the previous sentence)

When the Causeway reopened earlier that month, Causeway Link were very quick to announce their cross-border shuttle service, and numerous public assurances had to be made by political office holders on both sides on the progress of resumption of public bus and later KTM services due to the need to get through the red tape on both sides.

We haven’t seen the same thing here with night bus services. There was no rush to reinstate the night buses especially when it could be done with the reinstatement of 401 to East Coast Park. Okay, maybe they took their time with 401. But on-demand alternatives sprang up in the meantime, which might have helped build the case for bringing back 401. It doesn’t help that a driver driving night buses is not one who could be reused for the AM peak by extending his shift, since he would need to go home and rest during the 8–9am hot period when people are trying to get to work.

But what about on-demand night services? That was tried too, and they thought it didn’t work out for now, perhaps since passengers went to other scheduled alternatives like 2N and 4N. Like East Coast Park, though, it’s always possible for private bus operators to fill the vacuum. Heck, I’d even make a wager that the fares charged by private bus operators might end up similar to what the existing night bus network charged.

This makes me think, though, that URA might be somewhat justified if they want to think that self-driving pods will save us all. After all, there are several on-demand bus “coordinators” such as SWAT Mobility and RushOwl out there already. The LTA’s concern about high software development and system maintenance cost might be assuaged if they just assume a regulatory role and let the private sector take the lead.

Death of a thousand cuts

If you were to attempt to chart out the history of night bus services as Land Transport Guru did, you would find that a lot of the cutbacks actually took place in the 2000s, and the story about to be told here is pretty much that of the frequency-ridership spiral, as reduced ridership forms a feedback loop. Less passengers means service cuts, resulting in even lesser passengers and even bigger service cuts.

Remember that night buses started as a private-sector initiative before the days of bus contracting. TIBS were very ambitious with their plans at the turn of the millennium, opening most of their (currently-surviving) night bus network in 2000. The last NightRider buses at the time left at 4.30am, facilitating connections to the first feeder buses of the day in the estates. Furthermore, later in 2000, all services were converted to bidirectional operation so they could be of use to more people.

Around the same time, SBS also started up their night bus network — but these were more of 24-hour operations of daytime routes. From a service perspective this may be the better option, since people can just remember one route and they do not need to learn of a separate night bus network. However, in 2006 SBS also reoriented their night bus network in a similar fashion to what TIBS did.

2001 saw aggressive TIBS expansion, but this folded as quickly as it began with the 13 additional routes shuttered within months of their launch. SMRT also began subcontracting out night bus services to TIBS to follow their MRT lines when the MRT was closed (also a position I advocated for earlier). This was later merged with TIBS’ own night services in 2004; the same year that TIBS stopped bidirectional operation of night services on weekends, and completely stopped operating on weekdays.

The last major cut between then and now was in 2014, when SMRT Buses, as successor to TIBS, pushed all the last NightRider departures to 2am, following SBST who did something similar in 2007–2008. 2am is very early considering that there are night outlets in the Clarke Quay area that can remain open until 4am or maybe even later, which means either missing half the fun — literally, you’d have to leave 2–3 hours earlier — or forget the bus and just call a car. The cost of the car can be handwaved as the cost of those extra hours at the nightlife outlet.

Then of course there were the severe frequency cutbacks, where SBST routes could be as bad as every 40 minutes. SMRT routes weren’t as bad, but still not great at 20–30 minutes. All these pre-date the BCM. So I guess what was left to do was to get rid of the ineffective night bus system as it stands, which is what has been done.

Three dimensions

What have we learnt from this case? The success of a bus route lies in how well it manages to strike a balance between the three dimensions of coverage, capacity, and frequency. Yes, this means go big or go home. We have achieved coverage through a comprehensive bus network, but what’s the point when everything is at such low frequencies? Forcing transfers may well result in the same passenger experience as waiting 20 minutes as a bus stop for a one seat ride even if that bus is a double decker.

But conversely, too much coverage on an individual route can be an issue as well, as passengers are forced to ride with the bus as it snakes through town with several detours. This adds journey time. As an example of what not to do, here’s an example feeder bus route in the Klang Valley. Good luck understanding the diagram (I didn’t).

Yikes (source: Moovit)

You can see the same problem with night buses. With no feeder bus network to rely on, the night bus is forced to make large detours through the town.


So yes, whilst a private-hire car might be more expensive, the experience on the ground shows that people are willing to pay to get home and in their beds by 5am instead of 6am. That probably means the night bus doesn’t seem that attractive, especially after you consider the low frequency as mentioned above, and high fares at $4.50 a pop. If there are four in a group, that’s $17 — quite a way to the cost of a cab.

It doesn’t really help that taxis are just that cheap on an exchange rate basis in Singapore compared to other major world cities. This means that the “Uber effect” on public transportation usage, as mentioned by Alon Levy, may be more pronounced here. I haven’t adjusted for purchasing power, but my hunch is that even if I did, taxis would still be relatively more affordable.

Sure, there may be an issue with surge prices on private-hire cars as well, but I daresay that might be a side effect of the past few years of being spoiled by VC-sponsored promo codes and that prices are just returning to a more normal level as the VCs reduce their own spending. There might be some political pressure to restore some level of night service if surge prices get out of hand, but I won’t count on it especially as this is also an issue in the day with adequate public transport options.

Prime directives

After the whole 506/66 issue, MP of the affected area Gerald Giam raised a parliamentary question on bus service frequency, especially when that round of cuts apparently put some more pressure on 228 than it was able to handle. It’s a very interesting question that received a non-answer, although the question itself merits examining, as another example of why frequency is so important.

MP Giam’s question raised some very interesting metrics and he asked for the actual compliance with these metrics. I shall break down his question, which is: what percentage of

  • all public bus services and
  • feeder bus services,

operated with headways of

  • no more than five minutes during morning and evening peak hours
  • and no more than 10 minutes during off-peak hours, respectively.

Sure, some of these may be wishful thinking, but the rest are reasonable questions. The arbitrary distinction between feeder and trunk services means that there can be trunk services playing short-distance feeder roles, so it’s also important to hold them to higher standards. And indeed the LTA does, with 10 minutes as the baseline for half of trunk services alongside eight minutes on feeders.

This appears to be quite acceptable, with eight minutes being the average transfer waiting time according to Moovit Insights. However, bus-bus transfers may not be so fast especially if at least one service is in that half where 15-minute service is considered to be acceptable. Should the LTA, then tighten its policies to go to 10 minutes as a baseline, during peak hours?

Is that even acceptable? Is that even possible? Are we even hitting the 15-minute requirement today? What happens when we can’t? This is what is happening with night services, after all, with the restricted operating hours and low frequency.

A strategic retreat

The previous talk about superblocks in the Long Term Plan got me looking more into Barcelona in general. Interestingly, a side effect of the road closures created by superblock implementation is that there’s also lesser space for buses to go, because buses are vehicles and they need to use roads. What did Barcelona do? Well, they significantly cut back their bus network. If the TMB’s literature is to be believed, the resulting bus network in Barcelona maintains frequencies of 5 to 8 minutes throughout daytime hours.

We’re building superblocks too, in Tiong Bahru for one. But what matters are results; are those results repeatable here? We shall see — again going back to the point on whether AVs and road transport vehicles should be allowed inside the superblock. If good urbanism results in an increase in walkability or active mobility especially in the estates, it may be able to reduce the need for buses to take detours, which could promise faster travel and even some cleanup of excessively duplicating bus services. Yes, coverage by route-km will be decreased, but better results may be obtained if it’s possible to walk out to more frequent bus services.

Likewise, the bus network does not run in a vacuum. New rail developments are promised nearly every year up until 2032 and maybe further on with the remaining phases of the CRL. To preserve bus frequency elsewhere, they could be more aggressive with railstitution and redirection of resources — especially when the Bus Contracting Model means any extra resources such as drivers needed by a package has to come from the same package.

Whilst existing rail-related service changes may take place mainly in the frequency and capacity domains (by using single deckers and running them less often), this risks kickstarting the frequency-ridership spiral. People don’t want to take the bus when they have to wait a long time for the bus. Might it be better to just take the hit at one go instead of slowly sliding down the spiral?

Even at night when there is no rail service, night buses already fail the frequency test. Poor frequency and excessive routing increasing travel times indicate an overspecialization in coverage, which still has not achieved balance and thus alternative modes remain appealing. It might get even worse in the day.

When bus services fail these basic tests, how can we save it? And at what point do we cede the ground to other transport modes, either rail (in the day) or the private sector? After all, others may be better suited to offer more choices to serve these groups through ride-hailing and maybe even on-demand buses. Or perhaps someone could even operate Japanese-style capsule hotels near popular nightspots, so people can sleep there while waiting for daytime public transport to start.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store


Sometimes I am who I am, but sometimes I am not who I am not.