Do We Ever Count Passengers? [Updated]

For 2006, the TTC did not produce a Service Plan because, with the shortage of buses and operators and budget, there was not much point.  Many services are awaiting implementation, but the message is “come back next year”.

One fascinating part of the annual plan was the statistics for surface routes.  I have tracked these for about 20 years, and it’s fascinating to see how often (or not) the TTC actually updates the information for each route.

  • The Carlton route shows a daily ridership of 41,200, a number that is unchanged from 2001. 
  • The two Kingston Road routes (502/503) combined show 6,100, unchanged from 2002. 
  • Queen also shows 41,200, unchanged from 2002. 
  • Spadina/Harbourfront shows 43,400, unchanged from 2003.
  • St. Clair shows 31,000, unchanged from 2002.

Where streetcar routes do show changes, there is a slight decline in ridership.  Is this a reflection of service quality?  Do what extent to the counts predate the recent growth in system ridership?

Given the TTC’s limited resources to actually monitor riding, it may be another four or five years before someone discovers what is happening on the street.

To my regular readers at the TTC:  Yes, I know that service levels are not dependent on riding counts published in the Service Plan.  However, politicians, not the brightest bunch at the best of times, have used this type of statistics to argue about resources devoted to transit services.  To other readers hoping to find information in this data, beware.

The table for 2005-2006 can be found here.

Update September 30, 2006:

As I hinted above, the TTC does not base its service planning on the counts published in that table, but on “standing counts” taken at peak points of routes from time to time.  These are done much more often than full route counts mainly thanks to staffing cutbacks that make them impractical on a regular basis for major routes.

All the same, I (and others who receive complaints about TTC service) regularly hear from people who cannot get on vehicles.  This suggests that the level of service is not keeping up with the demand.  One point about counts, of course, is that they see only the riders who are already on the vehicles, not those who were bypassed by full cars or buses.

On the subway, there is a myth that we have lots of spare capacity.  Try telling that to someone boarding near the peak point.  During the film festival (the second week of September), I was never able to board the first train arriving at Broadview westbound at about 8:45 am for several days in a row, and on one day had to let four trains pass before I could board.  Fortunately, my regular commute from Broadview to STC takes me the other way, and I get to board lightly-loaded trains eastbound while those on the westbound platform curse the service.

With respect to the accuracy of the cost/revenue data in the TTC tables, I have learned that even the revered David Gunn was fooled by this so-called analysis into thinking that widespread service cuts were the answer to our problems.  Even the pros can be fooled by something that has the appearance of a thorough analysis.  Put enough numbers on a page and people stop trying to figure out whether any of them actually make sense.

6 thoughts on “Do We Ever Count Passengers? [Updated]

  1. We are after all talking about a system where pass validation is not enforced on buses and streetcars and at downtown stations (the open gate wave at the ttc guy/gal lounging in a chair) – now we hear about fake metropasses that are only detectable when validated!

    Any news yet on metropass sales since July 1 Steve?

    Steve:  The last report on fare revenue and breakdowns covered the period to the end of June, and moreover the big jump in pass sales likely came in September (that certainly was my observation based on lineups to buy them).


  2. There may be a catch 22 here. If there is no change in services rendered, why should any additional revenue be forthcoming. Having good statistics could turn out to be vital to the future health of the TTC.


  3. While the TTC hasn’t counted the Carlton since 2001?  It has gotten a lot more crowded.  Each year has been worse and this fall has got to be the worst yet.  I live at Parliament and I often cannot get on a streetcar.

    My metropass has been fairly useless for taking the streetcar this fall and it’s not even cold yet.  The Carlton line needs more frequent service or articulated cars.  Any driver could tell the TTC the line is past capacity.  I assume the problem is that there just aren’t any more streetcars to add to the line.

    Steve:  Actually there are some spare streetcars, but no spare operators.  The TTC is in a real mess on this.


  4. Not spending $$ to get proper up-to-date statistics of ridership seems terribly short-sighted.  How can one possibly plan without knowing where best to allocate resources?

    Though the ridership of the Airport Rocket (Bus 192) is supposedly only 2700 per day with a very low 22% cost recovery the TTC has put on additional buses recently (which all seem to be full) so I guess that even those who allocate resources do not believe the figures!

    Steve:  There are major problems with the ridership counts and the allocation of revenue and costs to all routes.  These figures are meaningful within a certain context, but you have to understand why something like the 192 works out the way it does.  Here are some issues:

    Heavy riding is very unidirectional.  The reverse trip probably runs almost empty.
    Since there are no local stops, the service has no opportunity for picking up additional fares from local trips as a bus like Finch West would do.  The result is that the number of boardings (which determines the allocated revenue) is limited to the capacity of the bus.  Local routes can get more boardings/trip because of turnover along the way and this allocates more revenue to them.
    All of the 192 passengers transfer from the subway.  The revenue/passenger is adjusted to allow for routes that have a large component of one-seat rides (passengers who don’t transfer on or off and therefore should contribute a full average fare to the calculation).  Half of the revenue for the 192 passengers goes to the subway.

    There are other potential problems, but I will stop there.  As you say, the planners have the sense to look at the actual loads on the vehicles, but the politicians only see the reports showing poor cost recovery.


  5. I’m looking at the stats at the 512 St Clair and the 60 Steeles West, and the cost differences between the street car and bus made my jaw drop.

    Notice that the differences:

    Both services carry around the same number of passengers.
    The Steeles route uses 35-50% more vehicles
    The Steeles route travels double the number of kilometres
    Costs 10% less to operate

    Based on this, doesn’t it mean that more frequent and lower cost service could be provided using buses instead of streetcars?

    Steve:  Thank you for this feedback as it shows exactly the sort of faulty calculations to which this table can lead.  The starting point for the analysis is the speed of operation.  Steeles West operates an an average speed of 19.3 km/hr (8100 km / 420 hours) while the St. Clair car runs at 12.19 km/hr (3900 km / 320 hr).  The speed of the vehicle is determined by stop spacing, traffic conditions and load patterns, not to mention technology.  If St. Clair were operated with buses, they too would run at about 12 km/hr, probably slower because they don’t handle large crowds as fast as streetcars.

    We pay operators per hour, not per kilometer.  If the vehicle runs faster, the cost per distance travelled is less.  Putting it another way, if the Steeles West bus ran at the same speed as the St. Clair route, you would need about 15 more buses on the route to provide the same level of service at a substantial additional cost.

    Bus cost allocationss are made up of three components:  hours of service, distance travelled and size of fleet.  If you run more buses, but run them slower, the total kilometers does not change, but the number of hours and size of fleet both go up.  Since these account for about 3/4 of the cost of a bus route, the Steeles West route would cost  over $30,000 more per day to operate if the buses ran at the same speed as on St. Clair.

    The next point to note is that the allocated revenue per passenger on St. Clair is 68 cents ($21,100 revenue / 31,000 pax) while on Steeles West it is 76 cents ($21,000 revenue / 27,500 pax).  

    First off, the TTC takes the average fare.  This is less than the token rate because of concession fares and passholders who travel at a discount.  Next, based on riding data, the average fare is converted to a revenue per boarding.  The average rider transfers once (some twice, some not at all), and so the revenue per boarding is around half of the average fare.  Finally, on a route-by-route basis this number is adjusted to allow for routes that have a higher or lower proportion of non-transfer passengers and therefore should get more revenue per boarding.  You will notice that St. Clair doesn’t fare (pardon the pun) very well here.  The point of all this is that buses running on St. Clair would be allocated the same revenue per boarding as the streetcar, about 10% less than those riding the Steeles West bus.

    Conversion between modes is a complex business, but the short story is that the streetcar option has much more room for growth back to the former riding levels on St. Clair and other routes (killed off by service cuts) than would be possible with buses.

    Projected ridership for 2006 is 442-million with revenues of $744-million for an average fare in 2006 of $1.68.  Since half of this is 84 cents, even allowing for a bit of inflation, the revenues assigned per passenger to both routes above look low.  When you go through the entire table and add everything up, you discover that the surface routes as a whole don’t get their fair share of revenue and they all have worse economic performance, on paper, than they should.


  6. To me, these type of comparisons are imperfect – but necessary.  Costs and revenues in most service providing organizations are at best indirectly linked. In allocating costs to revenue – or in this case, revenues to costs – some assumptions need to be made.  Users of the data need to be aware of the underlying assumptions.  However, at the end of the day, someone must allocate the finite resources to the various services.

    I agree that its not the best comparison between the St. Clair and Finch W routes – because of the difference in length.  A better comparison is between that route and the Queen St streetcar route.  Here we see that weekday ridership/service hour is comparable:

    Finch W – Bus: 78.7 riders/service hour
    Queen St. (Streetcar) – 77.7 riders/service hour

    Both routes are long and have varying levels of service along the length.  It’s interesting that the Finch W uses buses – whereas Queen St uses predominately the ALRV streetcars.  Yet the ridership per service hour is about equal – using the imperfect stats provides of course.

    Steve:  You have to be careful about counting riders per hour.  A vital part of the equation is how far each rider travels and how long they are onboard (in effect, how much service they consume) as well as the degree to which a route has good bi-directional demand (well balanced, or full one way empty the other).  Routes with very different characteristics can produce the same results, superficially.

    In my experience, streetcars are actually slower loading than buses – especially when there are passengers standing in the front aisle.  Streetcar boarding is far more adversely affected when an infirm passenger or mom with a stroller boards or alights.

    Steve:  This is a design problem with our current fleet, not an inherent problem of streetcar technology.  The TTC was so concerned about people slipping past the farebox that they created bottlenecks at the front of the cars.  If we move to low floor vehicles with all door loading, this problem goes away, but is replaced with the question of fare supervision, something the TTC seems unwilling to tackle seriously.

    I just don’t see the capacity advantages of the streetcars in their current form and service design. The top streetcar route in capacity delivered per service hour (based on the imperfect statistics) is:

    509/510 – Spadina/harbourfront @ 114.2 boarding/hour

    The top bus is:

    22 Coxwell @ 109.2 boardings/hour

    The Coxwell route doesn’t have a dedicated right-of-way – nor does it have signal priority of any type.

    Steve:  You have omitted the fact that the Coxwell bus carries 7,100 passengers per day while the Spadina car carries 43,400.  If the Spadina route were converted back to buses, you would need one every 90 seconds (or better) to handle the demand on the peak part of the route.  The Coxwell bus runs every 8 minutes.  The reserved lane gives the many more riders on Spadina better priority, although as I have written before, the “priority” given at signals is a joke.  The trip time and number of vehicles on the route could be reduced if the streetcars actually had priority at intersections, but the city prefers to give priority to left turning cars.  That costs us about 3 minutes in a one-way trip from King to Bloor and three extra streetcars on that line.


Comments are closed.