Updated March 25, 2016 at 1:45 pm: Charts and commentary have been added at the end of this article regarding monthly ridership statistics published by the TTC through the CEO’s Report.
In a previous article, I wrote about a TTC management report giving a “Ridership Update” for early 2016. This was triggered by concerns that overall ridership had stagnated, and the obvious question management might hear: “what are you doing about this”.
During the debate at the March 23, 2016 Board meeting, it became clear that to some extent, management was making up the story as they went along. The report includes data for the first two months of 2016, and February was particularly bad with a drop against both the budget target and against results in 2015.
However, once the debate was well underway, management reported that March to date was almost on budget, and was up 2.5% over 2015. Why was this information not in the report, or at least in a supplementary paper published before the meeting got underway? The entire tone of the debate would have changed with the sense that, maybe, things were turning around and ridership was growing again.
The situation was further complicated by repeated claims from management and from TTC Chair Josh Colle that ridership was not on a decline, but that it simply did not achieve its budget target. This quite flatly is not true as anyone capable of reading TTC reports can see.
In the February 2016 CEO’s report, a chart showed the data for 2015 with comparative numbers for 2014. The 2015 counts run consistently behind 2014 actuals from June onward, and all of 2015 is below the budget target.
Moreover, the actual full year count of riders for 2015 was down from 2014 as shown both in notes within the CEO’s Report and in the ridership and fare tracking table available on the City of Toronto’s Open Data website. (Thanks to @nfitz1 for pointing out that the spreadsheet had been updated with 2015 data.)
The reported total ridership for 2015 was 534.0 million while for 2014 it was 534.8m. That’s not just a failure to hit the budget target, that is an actual, albeit slight, decline. (The 2015 value does not include approximately 4m Pan Am Games free rides so that year-to-year comparisons are on the same basis.)
That tracking table provides an interesting long-term view into the evolution of TTC ridership and fare payment methods.
Total system ridership bottomed out in 1996 at the end of the early 90s recession, a period when through a combination of falling employment, rising fares and service cuts, the TTC managed to lose a substantial portion of its ridership. (This fall also provided a window during which there appeared to be “surplus” capacity in parts of the system allowing fantasies about subway extensions to take root, but that’s another topic.) Since 1996, with a brief drop in 2002-3, ridership has climbed consistently until 2015.
There is a challenge to management and politicians in an ever-growing era – the moment growth slows or reverses, they have no sense of what to do. At the TTC, this first happened in the early 1980s when, for the first time in decades, ridership fell slightly. The result was a new policy of “tailoring service to meet demand” that translated into “let’s pack those vehicles as tightly as possible”. Only with the Ridership Growth Strategy of 2003 did the TTC return to Service Standards that embraced the idea that less crowded service would attract more riders, a policy undone by the Ford/Stintz era at the TTC and still only partly reversed under Tory/Colle because the TTC had no vehicles with which to provide more peak capacity.
TTC management attempted to back-pedal from the appearance of imminent cutbacks saying, “no we never said that” and attempting to undo the “sky is falling” impression that the original report triggered. In a scrum after the meeting, CEO Andy Byford talked about a possible delay or deferral of service improvements, but strongly implied that cuts were not on the table. This does not accord with the content of his own report which suggests that actions under consideration include:
• A freeze on further service additions until it can be determined if the year-to-date ridership results are only temporary or more indicative of a lasting trend;
• Expense reductions that would be achieved from across-the-board service reductions in 2016 and 2017.
The next round of major improvements would normally come with the fall schedules in September or October 2016. These are set in place by early summer because of the lead time needed to organize crewing and other factors. Therefore, the go/nogo decision is only months away. The TTC Board has asked management for a review of what options and tactics are available to increase ridership, in effect, an updated Ridership Growth Strategy in which the TTC would understand in detail the behaviour of its market and present various options to attract more riders. The vital distinction in an RGS is to ask “what could we do”, not to start from the premise that “we can’t afford to do anything”.
The chart above shows that both weekday and weekend ridership have grown, but the ratio between them has also been shifting. Although weekdays still account for the lion’s share of trips, there has been a continuous decline in the ratio from 5:1 in the mid-80s to under 4:1 today. As the next chart shows, what is happening is that weekend riding is growing faster than weekday riding.
The chart below uses 1996 as a reference year because that was the low point for ridership. Since that time, and especially since 2003, weekend ridership has grown faster than weekday ridership, although it showed more of a downturn in 2015. This group of trips (together with off-peak riding on weekdays) does not get enough attention in ridership reviews that focus on peak period commuters. They are important, but they are not the only TTC riders.
The situation is also intriguing when we look at ridership by mode. In 2015, subway ridership continued to rise while surface modes suffered a downturn.
Looking at this over the entire period of the data with 1986 as the base year shows how ridership on various parts of the system has evolved, relatively speaking.
- The SRT opened in March 1985 and data for that year represent only 9 months’ operation. 1986 is used as the base year so that all modes are 12-month values in that year.
- SRT ridership bounces around from year to year, dropped in the early 1990s, rose on and off to 2010, and has dropped back almost to 1986 levels since. There is a problem with these data because periods when the line was shut down for major repairs do not count toward “SRT” riding.
- The bus counts stayed flat in the early 1990s in part because the area buses served continued to see population growth that offset the general economic decline in the older parts of the city.
- Streetcar riding was most affected by the decline on downtown employment, both in office and manufacturing, and it started a long climb upwards in the late 90s.
When 1996, the bottom of overall system ridership, is used as the base year, another important pattern is evident. Growth on the streetcar and subway networks has been stronger than on the bus network since the 90s. This is especially a challenge for the streetcars because the fleet, up until the arrival of the new Bombardier Flexitys, has stayed the same size for two decades. All of the growth has been accommodated without any increase in the vehicles available to carry riders. (Note that buses substituting for streetcars or for the SRT count as bus ridership, a statistical anomaly the TTC would do well to correct as it creates the appearance of lost ridership where, on a route by route basis, none may actually exist.)
The TTC’s tracking data also includes a breakdown of fares by class of user and by type within each class.
Over the long haul, the type of fare where we see the greatest change is for adults. Seniors and students still account for fewer fares in 2015 than they did in 1985 (71.0m in 2015 vs 76.8m in 1985). Riding by children is also fairly flat although there is a small jump shown for 2015. Note that this is an estimated figure based on the effect of free children’s fares starting with previous years’ data and assumed elasticity of demand.
Since 1996, the biggest jump in adult fare trips has been with the Metropass. The elimination of tickets saw a migration to tokens, but the Metropass growth was also strong suggesting that some riders moved over to that fare medium as a convenience.
A major problem for the TTC in “counting” fares is that over half of all rides are now taken using a medium which is a bulk purchase (some form of pass), and these sales must be convered back to an estimated number of “rides”. The Metropass average today is about 73 trips per pass, but the actual range of usage varies widely. It is not impossible to take over 100 trips per month with a pass, and some riders take less than 50. When sales fall, this is most likely to be among riders for whom the pass price is least attractive due to low usage, but the TTC counts the lost riding at the average value. This likely overstates the actual number of trips lost when a passholder reverts to tokens.
The tiny usage of Presto shows just how massive a jump in system demand will occur as riders convert to that mode over the next year. What might be minor annoyances, numerically speaking, from bad implementation decisions will become major headaches when usage grows from only 13.3m in 2015.
Senior and student fares have been shifting to passes from tickets over the past decades, and the trend implies that over half of the trips in this fare class will be by passes in only a few years.
The history of Metropass sales was shown in the 2016 budget by the following chart. Total pass sales declined, but this was concentrated in the full-price pass while other categories grew. One aspect of the Presto conversion that has not yet been decided is the form that passes, or their equivalent, will take including:
- Whether the monthly discount plan will be eliminated.
- Whether a “pass” will be sold as such (unlimited riding in exchange for up front payment) or as an automatic cap on fares charged to any rider based on actual usage (the GO Transit model).
Clearly, passes are a popular method of fare payment as shown by the fact that over half of all trips are carried on some form of pass today, and this has been the case for almost a decade. The TTC ignores the importance of passes as a fare medium at their peril.
Actual trip counts for passholders are dominated by the Adult Metropass, although the Post-Secondary pass has grown since its introduction in fall 2010.
Updated March 25, 2016: Commentary on Monthly Ridership Statistics
In addition to the annual summaries used to produce the charts in the original part of this article, the TTC provides ridership stats on a monthly basis through the CEO’s Report.
There are a few caveats required before diving into the charts:
- Because there are fewer than 12 regular Board meetings per year, there are also fewer than 12 published CEO’s reports. Therefore, some months’ data never appears in a public report. This is particularly true during an election year when there is a long break in the meeting schedule.
- Although the TTC sometimes refers to ridership per month, they actually are using 12 reporting periods per year with a standard pattern of 4-4-5 weeks to produce four, 13-week quarters regardless of the actual lie of the calendar. This leaves 1 or 2 days over (depending on whether it is a leap year), and I believe that these are included in period 12.
- Year-to-year comparisons may vary depending on whether holidays such as Christmas fall on weekdays or weekends, and whether Easter falls in March or April.
The raw monthly numbers are shown in the chart below. Where no data is available, there is no data point shown, but the monthly spacing is retained for visual continuity.
In this format, the distinction between the three sets of numbers is difficult to see in many cases because values are close together. The purpose of this chart is to show the regular peaks corresponding to 5-week periods. Even though it is a five-week period, the values are low for period 12 in 2013 and 2014 because of the harsh winter. Even at the end of 2015, period 12 does not show the same spike visible for period 12 in 2012.
The same data plotted in bar chart form look like this:
Even here, the distinctions are difficult to see and the trends are simply invisible.
However, if we look at the variation between current counts and those from the previous year, as well as the budgeted values, some patterns become more obvious.
The values versus the previous year stay positive until mid-2015. This corresponds to the comparison chart published by the TTC. After May 2015, the change goes negative except for a spike up in December (period 12). This is not surprising given how bad the weather in previous Decembers was compared to 2015, but one cannot assume that this is a lasting trend. Indeed, by January (period 1 of 2016), the change had fallen below zero slightly.
The numbers compared to budget are also important. Starting quite abruptly in 2013, the TTC never has as many riders as they budget for, and the percentage drop declines visibly through 2015 and into the first month of 2016. This begs the question of whether there is consistent overestimation of ridership in the budget, a mechanism as discussed earlier, which can provide headroom for improvements both in projected revenue and in budgeted service mileage.
Finally, we come to the moving total ridership year-to-year.
The blue line above shows the moving annual total (the sum of of the previous 12 periods’ worth of ridership), while the green line shows the values reported for the previous year. (Note that previous year values will not necessarily line up with “current” a year ago because different calendar boundaries have been used by the TTC.)
The yellow line shows the percentage change from “previous” to “current”, and this line shows a steady drop from 2012 through to the end of 2015. In other words, although ridership has been growing, the rate of increase in the year-over-year numbers has been falling for a considerably period. This has only been noticed recently when the monthly figures started to drop below the previous year’s data.
The values in this chart smooth out any short-term variations because they show 12-month totals, not individual monthly data. Only long-standing effects will show up. What is particularly troubling is the fairly sudden collapse of the space between current and previous starting in mid-2015, the point where monthly change became negative. The question, still unknown, is whether coming months will bring increases so that current values again track higher than the previous year.
The decline in the rate of ridership growth goes back several years, and it cannot be traced to a short-term effect, but rather must be viewed as a longer-term problem with the attractiveness of TTC service to riders. The problem has been hidden under overall growth.