OTTAWA - Summary
Canada’s National Capital region houses about 1.1 million people. OC Transpo currently serves nearly 60 percent of this population, and carries nearly 85 percent of the region’s transit ridership. The transitway network operated by OC Transpo has a system length of 46.3 km (28.7 mi). About half of the region’s 400,000 weekday passengers travel over some part of the transitway system, although reported “transitway” ridership includes a significant share of passengers who do not travel over transitway segments away from the Ottawa CBD.
The Urban Transit Area (UTA) population served by OC Transpo increased by more than 28 percent from 1982 to 1997. Taking vehicle size into account, the undertaking increased annual service (“standard equivalent km”) by 17 percent through 1992, essentially matching the UTA population increase (18 percent). However, ridership fell after 1984, and this overall trend was not reversed until 1998.
The 1984-1998 ridership decline is unfortunate, but less troubling than several negative trends evident during the same period: decreasing labor productivity and efficiency of labor utilization, increasing maintenance costs and apparent incidence of road failures, and decreasing fuel economy.
Operating cost per revenue hour increased at a “real” rate of about 2.5 percent per year during 1982-2002. In other words, unit operating cost escalated at an average annual rate 2.5 percent greater than the economy-wide rate of price escalation in Canada. The authors estimate that the net effect of real increases in pay rates, work rule changes favorable to labor and real increases in fringe benefits was an average annual (real) unit operating cost increase of about one percent. Therefore, roughly 40 percent of the cost impacts estimated above may be attributed to “increased labor costs.” The remainder - that is, the majority of the real unit cost escalation experienced during 1982-2002 - must be attributed to other factors.
The overall cost impact was substantial. The difference between inflation-adjusted operating cost per revenue service hour at 1982 and 2002 is (2002 CAD) $39.05. If this difference did not exist - that is, had OC Transpo managed to avoid real increases in unit operating cost - total operating cost for 2002 would have fallen $65 million (2002 USD 24 million) below the amount actually reported. Cumulative total operating cost for 1982-2002 would have fallen $1,360 million (2002 USD 865 million) below the “actual” sum.
The various negative trends suggest “inherent” or “structural” inefficiencies associated with Ottawa’s transitway program. Additional research is clearly indicated to determine underlying causes and identify potential counter-strategies.
Returning to ridership indicators: the number of weekday peak-period passengers (a.m. and p.m.) carried by OC Transpo almost doubled from 1972 to 1982. It then increased to 1985 (by roughly 20 percent), then remained relatively static (at slightly below 190,000) to 1989. A change in methodology might account for some of the decline implied by statistics post-1990, but a clear downward trend was apparent by 1993. This indicator increased by more than 23 percent (155,000-190,000) from 1997 to 2002.
The number of weekday midday passengers carried by OC Transpo more than trebled from 1972 to 1982. It then increased to 1985 (by roughly ten percent) and then remained relatively static (at about 95,000). The change described above might account for some of the decline implied post-1990. This indicator fluctuated around 80,000 per weekday to 1999, then increased by nearly 28 percent (to 102,000) at 2002.
The increase in the number of weekday evening passengers 1972-1982 was even more dramatic than the midday increase. This indicator fell from 1983 to 1985 (by nearly 23 percent, to 27,000) then remained relatively static. A clear upward trend emerged following the “strike year” (1996). This indicator increased by nearly 35 percent (to 40,000) from 1997 to 2002.
Saturday ridership nearly doubled from 1972 to 1982, fell from 1983 to 1986, then remained relatively static (at about 145,000) through 1989. A very sharp drop of 33 percent during a single year (97,000, at 1990) does not coincide with implementation of APC. Following a period of volatility, this indicator increased by 34 percent (to nearly 148,000) from 1997 to 2002.
Sunday ridership also nearly doubled from 1972 to 1982, but subsequent trends were different. Excepting the peaks during 1984-1986 and 1990-1991 (which might reflect changes in motor fuel prices), Sunday ridership remained quite stable (within the range 40,000-44,000) to 1992. Following implementation of APC and the 1996 strike, reported Sunday ridership increased by nearly 45 percent (to 87,000) from 1997 to 2002.
The average distance traveled per revenue passenger increased as successive transitway segments were opened. For this reason, changes in annual travel (passenger-km) did not mirror the trend in revenue riders. Annual passenger-km increased to 1984, remained relatively static to 1990 (at about 800 million pass-km per year), then declined to 1996 (by about 20 percent, to less than 650 million pass-km). Thereafter, annual passenger-km increased by nearly 34 percent, to nearly 870 million at 2002.
Changes in employment levels and distribution explain part of the 1984-1998 ridership decline reported by OC Transpo. UTA employment increased by more than 38 percent from 1981 to 1997 - significantly outstripping the rate of population increase. (This trend also suggests an increase in the number of employees who travelled between residences outside the UTA - e.g. in Quebec - and jobs within.) During the same period, central area employment grew by 21 percent. In other words, employment throughout the UTA grew by more than 100,000 but less than 14 percent of this growth, fewer than 15,000 new jobs, occurred within the central area. In addition, available data suggest that central area employment declined slightly (by less than 2 percent) from 1991 to 1996. As noted above, weekday peak period riders and annual passenger-km declined during this period.
Increases in real fares might explain part of the 1984-1998 ridership decline, but the relationship is not clear. Real fare per passenger increased by 16 percent from 1982 to 1987, but fare per km remained stable because the average travel distance per passenger grew longer. During this interval, weekday ridership increased by 12 percent, Saturday ridership declined slightly and Sunday ridership grew by eight percent. The trends from 1987 to 1992 are intriguing: continued increase in real fare per passenger, significant increase in real fare per km, and declines in weekday and Saturday ridership - but not Sunday ridership. Moreover, at 1992, weekday ridership had declined from 1984 but was still slightly above the 1982 level - despite fares per passenger 31 percent higher and fare per km was 14 percent higher than at 1982. Sunday ridership increased by 10 percent during 1982-1992 but Saturday ridership fell by nearly 25 percent. Increases in real fares might explain part of the decline in ridership per capita from the early 1980s but other factors (e.g. changes in employment locations) were evident; additional research is indicated.
An extensive body of research makes clear that consumers assign considerable disutility to transfers between vehicles. However, the actual influence of this preference for “one-seat” service is less clear. Kain (1992) suggested that most of the ridership decline during 1984-1988 was the result of increased transfer activity; this was disputed by the (then-) General Manager of OC Transpo, John Bonsall. No obvious correlation between the number of transfers per revenue passenger and various indicators of ridership per capita is visible above (“Ridership Trends”). The OC Transpo transfer rate increased by 20 percent during 1972-1982 while per-capita ridership increased dramatically; the authors note that high transfer rates are characteristic of Canadian and U.S. cities (e.g. Montréal, New York, San Francisco, Toronto) where transit attracts “European levels” of ridership. The OC Transpo transfer rate continued to increase through 1987 and stabilized thereafter, while per-capita ridership indicators continued to change significantly. The authors do not doubt the existence of a “transfer penalty” effect per se, but believe that its influence is often exaggerated in theory.
The authors again note, with interest, that the rate of OC Transpo customer contacts - and the inferred rate of customer complaints - increased sharply during 1985-1989. This coincided with a sharp increase in “road calls” and the initial period of ridership decline from 1984. Additional research is clearly indicated to quantify the degree of change in customer satisfaction and determine the influence on ridership.
In conclusion, the authors hope that this paper will encourage further analysis of bus rapid transit operating cost issues in the “real world.” The number of urban corridors in Canada and the U.S. with traffic sufficient to justify the investment for rail transit is small compared to the number where traffic levels justify relatively lower investments for improved bus service. The need for further analysis is clear.