Some of you may have read Geoff Nixon’s article in the Saturday Globe and Mail Toronto Section about the 33 Forest Hill bus route.  The piece gives a good overview of the issues, and that’s not just because I am quoted in it.

The list of “poor performers” comes out each year as part of the Service Plan, and this is a very dangerous document.  Why?  Because it gives politicians whose grasp of complexity may be tenuous something easy to understand:  an index, a hit list for budget cuts.  Indeed, a few members of Council complained that the TTC should just cut the routes on this list, save scads of money, and quit griping about funding problems.  Amusingly, few if any “poor performers” happened to be in their wards.

Where does this list come from?

First off, it is important to recognize that if a route doesn’t have many riders, there is no formula that is going to hide the situation on a purely financial basis.  But routes don’t just exist for financial reasons.  There are many other criteria including:

• geographic isolation (how hard or far is it to walk to another route)
• special traffic generators like schools, hospitals or industrial areas
• network role (does this route segment bridge two areas of acceptable performance, or is the route part of the 24-hour Blue Night Network)

Two sets of data are published every year by the TTC. 

One lists the alleged cost and revenue for each route along with other operating stats such as riding count, peak vehicle requirements and vehicle mileage.  [Gentle reader:  The TTC and the street railway industry before it have used miles as long as anyone can remember.]

The other lists various routes or route segments that do not make the grade at certain times of the day.  The measuring stick used here is different than in the overall list for reasons I will explain shortly.

To read the source report containing this information go here and read Appendices B and C.

Route by Route Performance Stats

The complete route list has a number of problems including:

• It is impossible to allocate fixed fare revenue on an appropriate basis to each route.  In the TTC model, passengers who ride only one vehicle for their trip (a local trip, for example) contribute less than a fare, and those who make long journeys on more than two routes allocate more than a full fare to the routes they use.  This overstates the revenue for routes serving long trips and understates the revenue for routes serving short ones.
• The costing model is a big improvement from one used in the old days, and it splits out costs that are influenced by different variables such as operators hours, vehicle miles, vehicles used, etc.  However …
• The system average cost recovery is supposed to be 80%, but the best of the routes is 29 Dufferin which gets only 75% of its cost back from the farebox.  It’s a fast downhill slope from there.  Either the costs are overstated or, more likely, the revenue is understated.
• Routes that serve many short trips (either because the route is itself short like 64 Main, or has a lot of turnover like 94 Wellesley) get lots of allocated revenue at low cost because nobody goes very far on these routes.  Routes with many long-distance riders, or with very unidirectional flows (full one way, empty the other) expend more resources per passenger and therefore come out poorly in the cost recovery stats.  They may be packed, but they lose lots of money on paper.
• Riding counts, especially on major routes like 501 Queen, are done every two or three years because the TTC can’t afford to do it more often.  At times of riding growth, the cost recovery figures don’t reflect the increased demand.
• Some routes that are logically part of another service (e.g. 503 Kingston Road is part of 502, 501 and 504) are reported separately even though their removal could require additional service on other routes to make up for any cutback.
• And finally, of course, people who don’t ride (or give up waiting, walk or take a cab) don’t show up in the stats at all.

All of this creates lots of opportunity for distortions and misunderstandings.

Poor Performing Route Segments by Time of Day

This list uses a different measure from the one described above.  The TTC calculates the number of customers who would be lost per dollar saved if the service in question were removed.  The magic line is set at .23 [don’t ask me to go into where that number comes from].  In other words, roughly, if we save $100 but lose 23 or more riders, it’s not a good idea.  Note that the “saving” is a net combination of reduced operating costs and lost fare revenue.

Let’s turn this model on its head.  We know that a bus costs around $100/hour to operate, but this is an average figure that includes peak and off peak service.  It includes fixed costs of owning and servicing a bus and its garage, and, moreover, crew costs are lower for off-peak services.  For the sake of argument, let’s use $70/hour as the cost of running a bus outside of the rush hour.  [Trust me on this.  The detailed paper on costing will come another day.] 

If each rider contributes half a fare (that’s how the TTC formula works), and that value is about 70 cents, then we must get 100 riders per bus per hour to break even.  That’s a lot of riders for a typical bus, and it will only happen if the route has a lot of turnover and/or has strong demand over its entire length in both directions.  This does not happen on many off peak routes.

If the bus gets 50 riders, it makes half of its cost in allocated revenue.  But wait!  If we cancel the service, we may lose full fares, not half fares from the model because riders may abandon the system.  Moreover, a return trip that might also have been taken will also be lost.  Therefore, the revenue impact of losing a rider can be quite substantial.  This shows the problem of looking at individual routes and trips rather than at the system as a whole.

Some of the TTC’s Poor Performers are no surprise:  Avenue Road, Bayview, Calvington, Davenport, Forest Hill, Lawrence-Donway, and so on.  But others are astounding:  Bloor West, Brimley, Coxwell, Mortimer, Wellesley (west of Yonge).  If you remove all of these routes or segments from the network where they fall below the line, the route map looks as if it were attacked by moths.  That’s what happened in the mid-1990s round of service cuts.

Big problems arise from this approach:

• Riders need a complex list of routes and timetables to know when any route they might choose to take is actually running.  The first time they discover that a route they want doesn’t run, it falls off of their mental transit map.  This is usually not a problem for a car which sits, ready to run, most of the time, in your garage.
• The savings from this type of cut are a one-time affair.  It will take at most two years’ worth of inflationary pressure for the budget crises to return, and we must look for a new round of cuts.
• Cuts to quality of service (some routes now run on headways greater than 30 minutes where only those with the patience of Job would wait for them) preserve a route on the map, but effectively eliminate it as a useful service.  The route map may look good, but don’t try to ride the system.
• Cuts to service produce crowding and delays which lead to short turns and appallingly bad service at the ends of routes.

Routes exist because they are part of a network.  Some routes will do very well by any measure, and some will always do poorly.  The question is: do people need them, are there riders, by what criteria would we eliminate some or all of the service?

Unfortunately, the TTC has so formulized this entire process that we lose sight of the riders and the need for service to be reasonably convenient and attractive.  We never see the count of people lost to transit, or the people who must now walk, hitch a ride or buy a car.  That cost is not on the TTC’s or the City’s books.

I know that this post has not fully answered the question of what we should do with the Poor Performers, but I hope that I’ve given readers some indication of the complexity of even deciding, fairly, which services are on the chopping block.  This year it may only be the so-called Nannie routes like Forest Hill [why, by the way, don’t Nannies deserve bus service but workers in an isolated industrial park do?], but next year the cuts could move up a notch or two.  We need to be sure that if we even think of cutting, we do it in the right place.

It says something about our priorities that the TTC has spent over two decades reporting on services that can be cut to save money, but the gains, such as they are, of the Ridership Growth Strategy, of actually trying to improve service are deferred over and over again.

The most bitter irony here is the huge subsidy paid to keep the Sheppard Subway operating.  I will turn to this issue in a separate post later this week.