SOME EFFECTS OF CONVERSION FROM STREETCAR TO SUBWAY TRANSIT LINES IN TORONTO
Donald N. Dewees
Research Paper No. 76
Centre for Urban and Community Studies and Department of Political Economy University of Toronto
January 1976
SO.ME EFFECTS OF CONVERSION FROM STREETCAR TO SUBWAY TRANSIT LINES IN TORONTO
I. Introduction
The construction of urban rapid transit lines has always been expensive, and the cost of this construction has been rising more rapidly than the general rate of inflation. Current additions to the Toronto subway system cost over35 million dollars per mile to construct . .!/ What benefits justify these enormous expenditures? While rapid transit means different things to different people, three major arguments have been used in favour of most of the subway construction that has been undertaken in North American during the last quarter century. The first is to improve transportation service for some particular group of travellers. A second frequently offered reason is that improvement of the transportation service will attract the inevitable growth in travel to public transportation. This is good presumably because of the undesirable side effects of the alternative forms of transportation, prin- cipally the automobile. A third motivation, closely related to the second, is to reduce congestion on city streets by "getting motorists out of their cars and into public transportation". Because of concern about air pollution, noise, and highway congestion, it is recognized that substantial social benefits could be achieved if the total number of motorists were actually reduced. It is often the hope that such a reduction can be achieved by improving public transportation performance with a subway or surface rapid transit line. - 2 -
Despite the tremendous expense of constructing an urban subway system, the existing empirical literature is not decisive in determining the extent to which these expectations may be fulfilled. The first effect, improving transportation service for particular travellers, is quantified in some studies. This is usually described by the change in average speed or the reduction in travel time for a particular trip or trips. Rarely is it mentioned that there are inevitably some persons for whom transportation service is made worse by the conversion from bus or streetcar service to a rapid transit line. Since this is not mentioned, it is not quantified.
The effect on modal split and on total motoring in the city are also not carefully studied. In many cases there will be a survey shortly after a new service begins to determine what mode of transportation was used by the patrons of the new service before that service was instituted. These surveys are used to assert the number of motorists who have been removed from the highways by the transportation improvement.~ One would expect however, that major changes in travel patterns would occur over a period of years, rather than instantaneously upon opening of the facility. Yet, when a long time horizon is examined, other changes take place in the city, including expressway construction, suburban development and sub stantial changes in travel patterns generally. These combine with generally rising incomes to mask the direct effect of improving the performance of one transportation facility.
The growth and change of cities over time raises a conceptual problem with examining the effect of a subway on transit riding or motoring: whether one is interested in the change over time as the city grows, or in - 3 -
the difference between these travel patterns now and what they would have been at this date if the subway had not been constructed? In a dynamic city, these two questions are radically different. It is non sense to ask how Torontonians would travel today if there were no subway system, since in the absence of a subway system that city could not possibly have grown to its present density, nor could current commutation patterns involving long work trips have been achieved. It is difficult to make meaningful comparisons between Toronto in 1975 and
Toronto in 1950, yet it is equally difficult to guess how Toronto would look today if no subways had been constructed.
This study cannot solve the problems of evaluating the long term effect of subways on travel behaviour. It represents only an attempt to answer some basic questions about these impacts. Three questions will be considered here. The first is the measurement of service quality changes for travellers along Bloor Street, before and after opening of the Bloor subway in the middle 1960's. The second is an examination of the effects on Bloor Street vehicular traffic of replacing the streetcars with subway service. The information derived here will be useful both for estimating the effect on auto congestion, and the effect on overall modal split. The third issue is an examination of some cordon count data which bear upon the modal split for commuters to the downtown area of the city and to a larger central area. These data cover more than 20 years, and give some indication of the long run trends that have taken place while the subway system has been expanded. Although it is not possible to draw cause and effect conclusions from these data, they do suggest the limits of - 4 -
possible effects from the subway expansion.
II. Comparison of Subway and Streetcar Service
It seems clear that replacing a streetcar line with a grade separated rail rapid transit such as a subway should improve service.
Operating speeds may be twice as great for the subway as for streetcars.
Before one can evaluate the side effects of the conversion from one service to the other however, it is necessary to have a quantitative description of the change in service quality. A simple specification of headways and operating speeds is not sufficient. The time saving from higher operating speeds is greater for long trips than for short trips, so that different travellers are affected differently. Furthermore the distance between stops is almost invariably increased. This increases the amount of walking to reach the transit facility, and to some extent there fore degrades the service. It seems likely that travellers who are mid way between subway stations and travelling only short distances may find service is worse with the subway than it was with a streetcar which would have stopped closer to their origin and destination.
This section will develop a model to describe transportation costs in a single variable called "weighted time". The cost model will then be used to evaluate travel costs for various trips by streetcar and by subway along Bloor Street. Conclusions will be drawn regarding the change in service quality for different locations and different trips.
A. Weighted Time as a Measure of Transportation Cost
The primary factors considered by travellers who are evaluating - 5 - alternative transportation modes are the time and price involved in each. We can ignore the price element when comparing streetcars and a subway in Toronto, since a flat fare is applied equally to both modes. Thus the difference between the two can be captured entirely by differences in time required for a trip.
A number of empirical studies give some guidance on the disutility that passengers place upon travel time, waiting time and walking time, in dollars per hour or as a percentage of the hourly wage rate.JI If we know the value or disutility that the traveller attaches to an hour spent in each of these activities, then we can compute the total cost of any trip between two points by multiplying the amount of waiting, travel and walking time by the hourly value in dollars of time spent in each of these activi ties. This would yield a total dollar cost for the trip. It may be argued that travel time is valued at approximately one third the hourly wage rate of the traveller.if Thus we can use travel time as a numeraire rather than dollars, and adjust waiting and walking time by their relative weights to convert them to equivalent travel time hours.
In this study two bases for time calculation will be used. In the first, we will use a naive assumption that all time is valued equally. The total time cost for a trip is simply travel time plus waiting time plus walking time in this calculation. The alternative assumption is based upon several empirical studies of the value of time. It is reasonable to assume from these studies that waiting time for a bus or a subway train is valued at approximately 1.5 times travel time. Walking time to or from a station, is valued at three times the rate of travel time valuation. Thus when we - 6 -
compute weighted travel time, this will involve travel time plus 1.5
times waiting time plus three times walking time. Both sets of results
are presented so that the importance of the weighting of times can be observed in the final result.
B. Description of Streetcar and Subway Services
Bloor Street in the West and Danforth Avenue in the East form a major straight artery running through the city of Toronto, crossing
Yonge Street in the middle of the city. Streetcar service originally ran from Jane Street, about five miles west of Yonge to Lattrell Avenue, almost five miles east of Yonge Street. On February 26, 1966, subway service replaced the streetcars between Keele Street, about four miles west of Yonge and Woodbine Avenue, almost four miles east of Yonge. In
1968, the subway was opened an additional three miles in each direction to Islington and Warden.
Table 1 shows the service levels offered by the streetcars in 1961 and by the subway in 1971, along Bloor Street and Danforth Avenues. The average vehicular speed on the subway is more than twice that of the street car both during peak and off peak periods. Service frequency however is somewhat less for the subway than for the streetcar. The much larger subway trains can carry a greater passenger. volume with less frequent service than is necessary for the streetcars. In fact, during rush hour, the Bloor streetcars were operated in two-unit trains. It should be noted that a significant simplification is involved in representing subway operation with a single average speed for the Bloor line. Speed is deter mined primarily by station spacing, and the stations near the outer ends - 7 -
of the line are further apart than in the centre city. Thus the actual
subway speed is somewhat less than that shown in the centre city, and
somewhat greater at the end of the line. These differences however are
not great, and do not seriously affect the results that follow.
TABLE I
Streetcar and Subway Scheduled Service
Streetcar Subway (1961) (1971)
Speed
Peak 9.96 MPH 22.5 MPH
Off-Peak 10.57 MPH 23.l MPH
Waiting Time (Minutes-Seconds) (= 1/2 headway)
AM Peak 1-15 1-42
PM Peak 1-05 1-38
AM Off-Peak 1-30 2-32
PM Off-Peak 1-00 2-32
C. Service Comparisons Using Weighted Time
Two different comparisons are used here to evaluate the relative performance of the streetcars and subway along Bloor Street: total travel time to Yonge Street, and trip length for equal performance of the street- car and subway. Total time to Yonge Street is the time required to travel - 8 -
from various points along Bloor Street to the intersection of Yonge and
Bloor. This includes the time required for walking to the subway station
or streetcar stop if any, the time waiting for a vehicle and the travel
time to Yonge Street. Because the station spacing is shorter and the walking
distance correspondingly shorter, the streetcars tend to have an advantage
for short trips, while the subway has an advantage for longer trips. Thus
we have also calculated the length of trip for which a traveller would be
indifferent between using the streetcar and the subway. All calculations
are made for both peak and off peak periods, with unweighted and weighted
waiting and walking time.
Walking time can be a large component of the total time required for
a trip, because of the slow speed of walking compared to either travelling
on the streetcar or subway. Thus the tradeoff between streetcar and subway
depends importantly upon where the trip begins and ends. If the trip begins
and ends exactly at a streetcar stop, and if a subway station is located at
those two stops after the line is constructed, then walking time is minimal.
For such a traveller station spacing is irrelevant; he considers only waiting
and travel time. On the other hand, if the traveller begins his trip midway
between two subway stations, and his destination is at a station, he may
face a walk of up to .3 miles. Since streetcar stops were generally spaced
7 or 8 per mile, the longest walk to a streetcar would have been .07 miles.
Thus the traveller who is midway between stations cares seriously about
station spacing. Thus all calculations are made both for travellers at a
station and travellers who would be half way between stations at one end of their trip. We have not considered the still more difficult case of a trip with both ends between stations. - 9 -
Table 2 shows total time to Yonge Street during the peak period
assuming all times have equal value. The first half of the table is
for a traveller beginning and ending his trip at a streetcar stop or a
subway station. For all distances shown in the table, the subway has
a noticeable time advantage, ranging from under one minute between Yonge
and Castle Frank to a 14 minute advantage between Jane and Yonge Streets.
That difference in total time approaches the subway total time, so that
the subway trip requires 45% less time than the streetcar trip used to.
These comparisons, like all others in this section, allow for the
time required to enter and leave the subway station. On the Bloor line,
it takes an average of 66 1/2 seconds to walk from the nearest curb out
side the station to the nearest point on an east-west platform. The
shortest recorded time is at Islington with 50 seconds, and the longest
time is between the Bloor line platform at Yonge and the street with approximately 100 seconds. A summary of these times is presented in appendix A.
The second half of Table 2 shows the total time to Yonge Street for travellers whose origin is midway between stops or stations. In this case, the streetcar is actually faster between Bathurst and Yonge and between
Castle Frank and Yonge, with the Castle Frank advantage amounting to
2 1/2 minutes. Thus travellers making these short trips were somewhat better off with streetcar service than with subway service, when walking time is included. After the trip length reaches two miles or more, the subway regains its advantage, as shown by the 3 1/2 minute advantage from
Dufferin street and the 2 minute advantage from Pape Street. Near Jane or Main Streets, the subway advantage is on the order of 8 or 9 minutes, - 10 -
TABLE 2
Total Time to Yonge Street: Peak Period ~ (All times equal value)
Yonge Castle Station Jane Keele Duffer in Bathurst Frank Pape Coxwell Main
Miles 5.0 3.8 2.6 1.5 1. 0 2.1 3.2 4.4
a) Origin and Destination at Station
Streetcar 31-16 24-05 16-50 10-12 7-11 13-49 20-27 27-40 Subway 17-11 14-01 10-47 7-53 6-33 9-29 12-25 15-37
b) Origin and Destination Between Stations
Streetcar 32-41 25-20 18-14 11-36* 8-35* 15-13 21-51 29-04 Subway 23-35 17-37 14-41 11-47 ll-09 13-05 15-37 21-19
~ All times in minutes and seconds.
* Streetcar is faster. - 11 -
even for travellers midway between stations.
Table 3 repeats the calculations of Table 2 using off peak schedules
rather than peak period schedules. Travel, waiting and walking time are
still valued equally in Table 3. In the off peak, streetcar service is
somewhat better relative to the subway than in the peak period, so that
even for the traveller at a station there is a slight advantage in taking
the streetcar at Bathurst and Castle Frank. The disadvantage of the street
car at Jane and Main is somewhat less than previously because running speed
has improved. Still the subway shows an advantage for most travellers
whose origin and destination are at the station. As in Table 2, for travellers
midway between stations, at Bathurst or Castle Frank, the streetcar holds
a measurable advantage over the subway. Beyond a two mile distance however,
the subway regains its advantage which grows to the end of the line.
Table 4 shows the total time to Yonge Street in the peak period where waiting time is valued at 1.5 times travel time, and walking time is valued
at three times travel time. The modal comparison for travellers whose origin and destination are at a station is little changed from that of
Table 2 except that all times are increased because of waiting time valuation
and subway times are further increased because of the importance of the walk into and out of the station. Here the streetcar shows a 2 to 4 minute advantage for trips to Bathurst or Castle Frank, and a tiny advantage even to Pape. The subway time saving is not significant until trips become longer than 3 miles or so.
The radical change brought about by the heavier weighting of waiting and walking time shows up in Part b of Table 4 where origins and destinations
are midway between stations. Here the streetcar is always faster than the - 12 -
TABLE 3
Total Time to Yonge Street: Off-Peak~ (All times equal value)
Castle Station Jane Keele Duffer in Bathurst Frank Pape Coxwell Main Miles 5.3 3.8 2.6 1.5 I 1.0 2.1 3.2 4.4
a) Origin and Destination at Station I I
Streetcar 29-37 22-49 16-00 9-45 6-56* 13-04 19-25 26-14 Subway 17-45 14-37 11-3.0 9-39 I 7-21 10-12 13-04 16-11
b) Origin and Destination Between Stations
Streetcar 31-01 24-13 17-24 11-09* 8-25* 14-28 20-49 27-38 Subway 24-09 18-13 15-24 12-33 11-57 13-48 16-16 21-53
~ Times in minutes and seconds.
* Streetcar is faster. - 13 -
subway, even for trips as far as Jane or Main Streets. Travellers
between Bathurst and Yonge or Castle Frank and Yonge actually record
a 10 minute equivalent travel time loss from using the subway as com
pared to using the streetcar if they are midway between stations. At no point is the subway faster, and its disadvantage ranges from 1 1/2 minutes at Keele to 14 minutes at Castle Frank. Even at Main Street and
Jane Streets 4 1/2 to 5 miles from Yonge Street, the streetcar has a five minute advantage over the subway.
Table 5 repeats the calculations of Table 4 for the off peak period.
Here again the streetcar is faster for travellers originating as far as
Dufferin or Pape Streets, over 2 miles from Yonge. For travellers between stations, the streetcar once again is always preferable to the subway, with savings of five minutes to 16 minutes.
Another way to compare the performance of the stre~tcar and subway is to calculate the trip length at which travellers are indifferent between the two modes. Table 6 shows these trip lengths for indifference assuming all times have equal values. Travellers whose origins are midway between stations when the stations are .29 miles apart would prefer the streetcar for trips of less than 1.26 miles and the subway for trips longer than that distance during the peak period. In the off peak, the break-even point would be 1.63 miles. The first four lines of the Table show that as station spacing increases, the trip length for indifference also increases to 3.68 miles in the peak period and 4.26 miles off peak when stations are 1.1 miles apart. This increase in the trip length of indifference because of the longer walk required to reach more widely spaced subway - 14 -
TABLE 4
Total Time to Yonge Street: Peak Period~ (Waiting= 1.5 x Travel; Walking = 3 x Travel)
Yonge Castle Station Jane Keele Duffer in Bathurst Frank Pape Coxwell Main
Miles 5.0 3.8 2.6 1.5 1.0 2.1 3.2 4.4
a) Origin and Destination at Station
Streetcar 32-27 25-16 18-00 11-22* 8-21* 14-59 21-47 28-50 Subway 23-17 20-07 16-53 13-59 12-39 15-35 18-31 21-44
b) Origin and Destination Between Stations
Streetcar 36-39* 29-28* 22-12* 15-34* 12-33* 19-11* 25-59* 33-32* Subway 42-29 30-55 28-35 25-41 26-27 26-23 28-06 38-49
-a/ T'imes in' minu' t es and second s.
* Streetcar is faster. - 15 -
TABLE 5
Total Time to Yonge Street: Off-Peak~ (Waiting = 1.5 x Travel; Walking = 3 x Travel)
Yonge Castle Station Jane Keele Duff er in Bathurst Frank Pape Coxwell Main
Miles 5.0 3.8 2.6 1.5 1.0 2.1 3.2 4.4
a) Origin and Destination at Station
Streetcar 30-52 24-04 17-15* 10-30* 8-11* 14-19* 20-40 27-29 Subway 24-43 21-35 18-28 15-37 14-19 17-10 20-02 23-09
b) Origin and Destination Between Stations
Streetcar 35-04* 27-16* 21-27* 15-12* 12-23* 18-31* 24-52* 31-41* Subway 43-55 32-23 30-10 27-19 28-17 27-58 29-38 40-15
-a/ T'imes in' minu' t es a nd second s. * Subway is faster. - 16 -
TABLE 6
Trip Length for Indifference Between Streetcar and Subway (All times equal value)
Distances in miles
Trip origin midway between stations
Station Spacing Peak Off-Peak 67 second station entry except as noted
0. 29 miles 1.26 1.63 (Chester-Pape)
O. 50 miles 1.89 2.31
0.60 miles (Lansdowne-Dundas W.) 2.19 2.64
1.10 miles (Royal York - Old Mill) 3.68 4.26
0.50 miles (43 second station entry) 1.65 2.06
(100 second station entry) 2.22 2.67
Trip origin and destinaticn at stations
Station spacing irrelevant 81 1.14 - 17 -
stations, while streetcar station spacing is assumed to be constant.
The first four lines of the Table assume that 67 seconds are required to enter or exit from a subway station. Tl-ie i1ext two lines, under the heading 0.5 miles station spacing, show how the indifferent trip length varies according to station entry and exit time. If the entry time is only 43 seconds, the indifferent trip length drops from 1.89 to 1.65 miles in the peak period. When 100 seconds are required per station entry (as at Yonge Street) the trip length of indifference rises to 2.22 miles. Thus walking into and out of the station is a significant factor in comparing streetcar and subway performance even when walking time is valued the same as travel time.
The last line on the Table shows the indifferent trip length when origins and destinations are at the station. The short distances of .81 miles on the peak period and 1.14 miles on the off peak reflect the relatively small penalty imposed by walking into and out of the subway station. Since few transit trips are less than one mile in length, this means that most travellers whose origins and destinations were at a subway station would find service better with the subway than it had been with the streetcar. This would include all passengers transferring to the Bloor line from other bus or streetcar lines. Only one set of calculations is shown for trips originating and terminating at stations because station spacing is irrelevant to the choice among modes here.
Table 7 repeats the calculations of Table 6 using weighted travel time. Here the low speed of walking combines with the high assumed value of walking time to greatly increase the indifferent trip length. Even - 18 -
with station spacing at 0.29 miles, the streetcar is preferred at trips
less than 3.62 miles long in the peak period and 4.45 miles in the off
peak period. When station spacing rises to 1.1 miles, the streetcar is
preferred for trips up to 10.78 miles in length miles in the peak period
and 12.68 miles in the off peak period. These calculations show that many
travellers would have been better off using a streetcar than the subway
for the trip lengths shown here.
The findings of this section can be briefly summarized. Travellers
whose origin and destination are at a subway station will find that the
subway offers better service than the streetcars for all but the very
shortest trips: under 1 mile if times are valued equally, or 2 to 3 miles
if waiting and walking time are more highly valued. Travellers originating
between stations however, will find that their evaluation of the two modes
depends strongly upon the length of their trip and the station spacing. Even
if times are valued equally, a trip of 3 or 4 miles may be made more
advantageously by streetcar than by subway. For travellers between stations,
when waiting and walking time are valued more highly than travel time, the
streetcar may offer better service for trips of up to 5,6 or even 10 to 12 miles. Since many transit trips are shorter than these distances, there may be a significant number of urban transit riders who find service worse
rather than better when surface lines are replaced by even a high quality
subway line.
This suggests that the replacement of a streetcar service with subway
service may cause some previous transit users to switch to another mode,
either walking or using the automobile. It would be possible to estimate - 19 -
TABLE 8
Trip Length for Indifference Between Streetcar and Subway (Wait= 1.5 x travel; Walk= 3 x travel)
Trip origin and destination midway between stations
Station Spacing Peak Off-Peak
67 second station entry except as noted
O. 29 miles 3.62 4.45
0.50 miles 5.50 6.50
0.60 miles 6.39 7.47
1.10 miles 10.78 12.68
0.50 miles (43 second entry) 4.80 5.74
(100 second entry) 6.50 7.59
Trip origin and destination at stations Station spacing irrelevant 2.28 2.99 - 20 -
whether the number of passengers for whom service deteriorated was
significant or not by extending this study to a mapping of areas near
the Bloor Street transit lines to determine within which areas service
had been improved and within which areas it had degraded under the
various assumptions examined here. When one moved far from Bloor Street,
all passengers should be using north-south bus or streetcar service to
reach Bloor Street, and therefore would be arriving at stations. Thus
those for whom service is worse are those between stations and not too
far from Bloor Street. Some calculation of the boundaries of the areas
for which service was worse and examination of transit travel patterns could give an indication of the relative importance of the service degra dation.
Two important limitations on these results should be noted. First, calculations are based upon scheduled operations. It is common, however, for surface transit vehicles operating in heavy traffic and with signals to become bunched together, particularly during rush hour. To the extent that this occurred for the Bloor streetcars, the average waiting time would have been greater than that shown in the schedule. If service is scheduled every two minutes, but vehicles arrive in groups of two or three, then the average time between service is not two minutes but four or six. The bunching ten dency is far greater for surface vehicles than it is for subway trains.
Thus the performance of the streetcars is probably somewhat overstated.
A second ommitted factor is the importance of variations in time. We have worked entirely with average values here. However it seems clear that people attach some value to reliability of transportation service independent - 21 -
of the average speed of service. Because streetcars are subject to
traffic interference, the variance in their speed would be greater
than the variance in the subway system speed. To the extent that this
variance is important to travellers, it is a negative factor for the
streetcars that is not considered in this evaluation.
III. Traffic Effects of Replacing Streetcars With a Subway
It has been suggested that replacement of surface bus or street-
car operations with grade-separated rail rapid transit such as a subway may actually increase some traffic levels. The primary mechanism for this
increase would be the improved traffic flow on a street that was now
free of surface transit vehicles. While this question is far more narrow
than considering whether the improvement of transit services affects
total motoring, it is not without interest. Data are available to test
the impact of replacing streetcars with a subway on Bloor and Danforth
Streets in Toronto. The Bloor-Danforth subway was opened on February 26th,
1966 between Keele Street and Woodbine Avenue. On the same data, street
car service along this route was terminated. The subway was later extended
to Islington in the West and Warden in the East, with service beginning
in May 1968.
Table 8 shows traffic volumes on Bloor Street and Danforth Avenue at selected locations shortly before and shortly after the opening of the first section of the subway in 1966. These data show substantial increases in traffic flow at most locations along both streets, in both the heavy and light directions during both morning and evening rush hour. The general increases in traffic volume extend even beyond the Woodbine station on TABLE 8
TRAFFIC VOLUMES ON BLOOR STREET AND DANFORTH AVENUE BEFORE AND AFTER SUBWAY OPERATION
-LOCATION AM PEAK HOUR PM PEAK HOUR Westbound Eastbound Westbound Eastbound BEFORE AFTER BEFORE AFTER BEFORE AFTER BEFORE AFTER
Danforth Avenue at: Jackman Avenue 882 1183 405 615 587 724 977 1260 Logan Avenue 962 1084 386 644 606 695 836 1214. Carlaw Avenue 963 1225 395 642 460 762 874 1134 Pape Avenue 911 1219 434 474 524 744 795 1173 Jones Avenue 897 1518 537 435 479 807 1198 873 Donlands Avenue 809 1148 422 627 423 713 922 1087 N Greenwood Avenue 778 1129 429 628 539 728 849 1074 N Monarch Park Avenue 930 1122 371 385 641 605 717 1144 Coxwell Avenue 967 977 340 638 527 594 820 1163 Glebemount Avenue 938 1146 415 449 503 623 931 1054 Woodbine Avenue 1159 1198 499 488 688 711 1145 1112 Cedarvale Avenue 1063 1473 509 492 635 907 925 1247 Gledhill Avenue - Patricia Dr. 1241 1228 458 539 729 624 1085 1085 Main Street 1180 1352 559 544 765 798 1127 1048 Bloor Street West at: Spadina Avenue 433 566 871 976 631 891 739 760 Bathurst Street 287 377 432 762 481 767 361 449 Palmerston Avenue 318 315 452 724 462 699 393 431 Manning Avenue 259 360 523 740 603 900 369 405 Keele Street - Parkside Drive I 434 469 908 814 882 1020 653 804 I :
Source: Data from Metro Toronto Department of Roads and Traffic - 23 -
Danforth, to the last three locations in the Danforth portion of the
table. While traffic increases are less beyond the subway terminus than along the subway line, they are not negligible here. This undoubtedly reflects an increase in vehicles that will utilize the now more rapidly flowing portion of Danforth where streetcars no longer operate.
The data in Table 8 are summarized briefly in Table 9. The total number of vehicles travelling inbound in the morning rush hour on Bloor and Danforth increased 26% and 24% respectively. In the evening rush hour the outbound peak traffic flow increased 40% on Bloor and 19% on Danforth.
Thus peak direction traffic for both rush hours appears to be up approximately
24%. Traffic moving against the rush hour also increased in all cases, although by somewhat less than the increase in the heavy direction. The off peak increase appears to be about 22% when averaged over both rush hours and both roads.
Care must be taken in interpreting the data presented in Tables 8 and
9. They do clearly show an increase in traffic on Bloor Street and on
Danforth Avenue when the streetcars were replaced by the subway. They may not however be interpreted as increases in total motoring for the city. It is quite possible that some of the additional vehicles have been diverted from other parallel streets. To the extent that these added vehicles have come from other routes, they represent improvements in traffic flow, but not net additions to total motoring. Only those drivers who had previously not driven at all could be counted as net additions. These data are in adequate to determine that number. - 24 -
TABLE 9
Average Traffic Volume Changes (Percent of Pre-Subway Levels)
Bloor Danforth Total
Morning Inbound (peak) 26% 24% +25%
Outbound (Reverse) 21% 24% +23%
.Evening Outbound (peak) 40% 19% +23%
Inbound (Reverse) 10% 24% +21% - 25 -
The data in Table 8 were collected shortly before and shortly after
initiation of the subway service and termination of streetcar operations.
Because the subway was constructed a hundred yards or so north of Bloor
Street and Danforth Avenue, rather than directly under the streets, traffic on both roads was little affected by the construction itself. Particularly
in the months immediately before opening of the subway, traffic levels on Bloor and Danforth should have been unrelated to the construction activity. Thus the change in traffic volumes here should reflect the re moval of streetcars and the improvement of transit service and little else.
It has also been ascertained that no important changes were made in
Bloor Street or Danforth Avenue or immediately adjacent highways at any time close to opening of the subway. Table 10 lists major highway changes
.that might conceivably have affected traffic on Bloor Streets and Danforth
Avenue. The only relevant expressway construction was opening of the
Gardiner expressway which was completed by 1964, two years before the sub way opened. Thus it seems most unlikely that any of the traffic changes shown in Tables 8 and 9 could result from changes in the highway network it self. These changes must then fairly be interpreted as resulting from the removal of streetcars and operation of the subway service.
IV. Modal Split to the Toronto Central Business District
Between 1950 and 1973, over 20 miles of subway line were constructed radiating North, East, and West from the central business district of
Toronto. During the same period a six lane expressway was constructed serving the downtown area from the West, and another six lane expressway - 26 -
TABLE 10
Road Improvements Possibly Affecting Bloor-Danforth Traffic
1961-1971
Year Road Improvement
1961 Bloor Street widened: Royal York Road to Dundas West
1961 Danforth Ave. resurfaced: Jones Ave. to Woodbine Ave. and Main St. to Dawes Rd.
1962 June 29 - Gardiner expressway opened Westbound: Jamison to Spadina Aug. 1 - Gardiner expressway opened Eastbound: Jamison to Spadina Dec. 3 - Gardiner expressway opened Westbound: Spadina to York
1964 Sept. - Gardiner expressway opened Eastbound: Spadina to York - Gardiner expressway opened : York to Don River
1966 Dundas Street reconstructed: Bloor to Dupont Keele Street widened: Bloor to Annette
1967 Warden Avenue widened: Danforth to Eglinton Victoria Park Avenue widened: Danforth to Dentan
1969 Danforth Avenue widened: Sibley to Massy
1971 Kipling Avenue widened: Bloor to Rathburn - 27 -
constructed serving downtown from the North-East. See Figure 1. The population of Metropolitan Toronto increased over 70% from 1.3 million in 1954 to 2.2 million in 1973. Real income per capita more than doubled during the same period.
The increase in passenger carrying capacity represented by these facilities can be estimated rather crudely. If each expressway lane carries approximately 1500 vehicles per hour then the total of six additional inbound lanes would have a capacity of 9000 vehicles per hour or 18,000 vehicles for the two hour morning rush. With an occupancy rate of 1.5 persons per vehicle this would allow 27,000 additional persons to enter the centre city per rush hour. The Yonge Street and University subway lines could bring over 40,000 more persons into the downtown area than previous streetcar service could carry. The Yonge line combined with the
Bloor-Danforth subway could bring more than 50,000 additional persons into the midtown area around Yonge and Bloor during the two hour morning rush.
Finally, the lakeshore line of the GO-transit commuter system carries between five and ten thousand commuters into the downtown area every morning, although these are not counted in the available travel data that follow. It therefore appears that transit capacity to midtown has in creased by over 50,000 passengers per morning rush, and by about 40,000 into the downtown area. Highway capacity to downtown has increased by approximately 27,000 passengers per rush hour.
If the primary constraint on commuter access to the core area were transportation system capacity, then one should expect transit ridership to the core area to have grown far more than automobile ridership during the period 1950 to 1973. In fact, just the reverse has occurred. Table 11 shows that the number of automobiles and automobile passengers entering - 28 -
Figure 1
Transportation Facilities and Traffic Cordons in Toronto
York Mills Ave. I I
.µ ~ U) ~I QJ Iii p.. i:::' °''0 QJ:>. I ><' r-l r-l ~I i::: --~-~---~ Eglint.Q..!L.AY.~. 9 I I /ai \ / ~ ~1 h QJ \ Intermediate i::: 1 / ·.-! \ Cordon ""' \ ------·- - - - ·- - -·-·r, ~ 0 Bloor __t) __ '..') (;. -~/"- th\-~&---'"!"t·--~· ~ ~ 0 f f Danforth Ave. ~ St~}l - 0 0 CJ CJ fl ,, ?"\ "' ~ 0 I \ •, Downtown } Cordon G'~q. - -1· -; l \ ; () f I -- -- .J.-0<::><- b I / . -- -<;-\': L--?,u; .. / -- - '-P<-,,. I ') I I i '•, "'&..,. I '. I I / -~ ~ "'o/q ~--··----· -·-·••• ·~~--·-*--r" l . I .-/ •.. _ :y I···------!------i '·· ...... , __ ·...... - - ---'.. I _- .,,..,...( .. -""- ___, ~ -·------· ''{ (------\ i \~ ///__ /ji /
~-...-/ \. /
-t+----9-----Y----9- Subway
Expressway - 29 -
downtown increased 50% between 1950 and 1953, while transit riding to downtown decreased by 7,000. Between 1953 and 1973, the number of auto mobiles entering the downtown area increased by another 40%, while automobile ridership increased 20%. During the same twenty year period from 1953 to 1973 transit riding to downtown Toronto decreased by 10,000 passengers per day from 81,000 to 71,000. Most of this drop in transit patronage took place between 1953 and 1961, despite the opening and operation of the Yonge Street subway line.
The direct effect of particular transportation investments can be seen only dimly in these data. Transit ridership increased by 4,000 passengers per day in 1955 as compared to 1953, presumably as the result of opening the Yonge Street subway line in 1954. Still, transit riding declined steadily for five years thereafter. It is difficult to discern any effect of the University subway line, and the opening of the Bloor line could only be observed by the high ridership in 1967 which is not repeated again until 1970. Between 1969 and 1972 transit ridership increased without the assistance of any new transit projects.
Motoring seems to have risen following the completion of the Gardiner expressway to York Street in 1963, with two years of significantly higher motoring in 1964 and 1965 than had previously occurred. Otherwise there is no discernible relationship between expressway projects and motoring to the centre city.
The picture is slightly different when the intermediate cordon across
Allendale, North Toronto and Don Valley is considered rather than the limited downtown Toronto cordon. Completion of expressway projects between 1961 and TABLE II Morning Rush Hour Traffic To Toronto 1950 - 1973 (7 - 9 A.M., thousands of autos or persons) . a/ Entering- Downtown- Toronto Across the AllendaleJNorth Toronto and Don Valley Cordons Persons Persons Autos by Auto by Public Transit TOTAL I Autos by Auto by Public Transit TOTAL
1950 11.1 19.5 87.6 107 51 13.4 22.6 85.8 108 52 16.4 30.6 81.6 112 46.5 80 114 194 53 17.3 29.l 81.2 110 52.9 89 112 201 54 16.5 28.3 75.1 103 55 18.4 31.2 85.6 117 54.7 91.3 114 195 56 17.8 30.0 82.9 113 57 18.5 30.8 83.8 115 61.9 97.3 113 211 58 20.2 33.8 81.0 113 63.6 99.3 108 206 59 19.2 32.5 80.9 106 65.2 99.7 108 208 60 19.8 32.l 74.3 105 66.1 103 106 209 61 20.6 33.4 71. 0 104 71.3 109 104 212 w 62 18.1 29.2 75.7 105 74.4 114 104 218 0 63 20.1 31. 9 71. 9 104 74.7 112 104 217 64 23.0 36.7 71.8 109 76.4 112 108 221 65 23.2 36.0 69.6 106 68.8 103 103 206 66 21. 6 32.4 71. 7 104 70.4 106 112 218 67 22.5 33.8 75.5 109 67.9 102 115 217 68 21. 3 32.0 59.7 91. 7 . 81. 5 107 114 222 69 23.7 35.6 70.4 106 70 25.0 37.4 74.7 112 73.3 110 121 231 71 22.9 31.9 73.0 105 72 24.5 34.2 76.5 111 73 25.2 ' 35.1 71. 9 107 69.6 94.5 122 217
~ Dundas to Front Streets between Jarvis and Simcoe. Source: Metro Toronto Key Facts. - 31 -
1965 is accompanied by a significant growth in motoring that continues
through 1964. The drop in motoring in 1965 may result from a change in the cordon boundary definition. Transit ridership however, has grown by almost 10% in the period 1953 to 1973, when the intermediate cordon count is considered. While there is no discernible effect from the Yonge Street subway line construction in 1954, subway ridership leaps by almost 10,000 in 1966 when the Bloor line is first opened. By 1970, when that line has been extended to its present termini, ridership has grown by another 8,000 passengers per day. Thus it would appear that the number of commuters using transit to the limited downtown area has not been substantially increased by the transit investments during this 20 year period. It does appear that transit ridership to the midtown area of Bloor and Yonge has increased significantly, perhaps in part due to the replacement of the streetcars by a subway line. The increase however, is far below that which would be per mitted by the increase in capacity brought about by new transit projects.
The modest increase in motoring between 1959 and 1973 to the down town area is a small fraction of the total that would have been permitted by the completion of two major expressways. Between 1960 and 1969 the number of automobiles entering the intermediate cordon has increased only modestly, again far too little to account for the large increase in highway capacity.
If one considers the data in Table 11 together with the population, vehicle ownership, and transportation investment data of Table 12, one can only conclude that some important information and variables must be missing.
The number of persons going downtown in the 1970s is slightly less than the number going downtown in the mid 1950's, despite vast increases in the capacity and performance of the transportation system. The total number of - 32 -
persons entering the intermediate cordon area has increased only 10 per
cent between the middle 1950's and the 1970's. The effect of all this
transportation invest.iuent therefore, \•las clearly not to permit a major
expansion of downtown employment, since that employment appears not to
have increased according to these figures.
What then were the effects of these projects? The above data are not
sufficient to answer that question, but two hypotheses seem likely. First,
given the rapid expansion of suburban employment and residential facilities during this quarter century and the expansion of the highway network serving
suburban areas, it is quite possible that the facilities considered here have merely allowed the core area to maintain its original employment activity. Perhaps without the subways and expressways the downtown area would have contracted in employment and activity. Second, the additional facilities must allow persons to cormnute downtown and to midtown from substantially greater distances than before. With transit travel time cut almost in half for long trips, and the average motoring speed approximately doubled, work trips could double in length without increasing total commuting time. It would be interesting to examine origin-destination surveys made at the ends of this period to determine whether the change in travel patterns is more dramatic than the small changes in total persons destined for the central area.
Several research questions are suggested by the data and discussion above. First, it would be useful to calculate changes in performance of various transport links to the downtown area so that the relative importance of the transportation projects discussed here could be assessed. Second, it TABLE 12
Population, Vehicle Registrations and Transportation Facility Openings I b/ Popula tion.:Y Re~istered Vehicles ------~---- (millions) All Vehicles Automobiles Persons/ (OOO's) (OOO's) Automobile Openings of Transportation Facilities 1950 5.1 1951 Ontario 4.8 1952 data 4.7 1953 4.4 1954 1.28 351 290 4.4 Yonge Subway: Eglinton to Union Station 1955 1.32 397 333 3.9 - 1956 1. 36 429 364 3.7 - 1957 1.44 453 378 3.8 i (Humber River to westena or C.N.E.) - 1958 1. 51 474 397 3.8 Gardiner expressway - 1959 1. 55 498 421 3.7 T Hereafter referred to as G.E. - 1960 1.59 547 448 3.5 - 1961 1.62 559 458 3.5 I Don Valley Parkway (Bloor to Eglinton). Herearter D.V.P. 1962 1. 65 574 467 3.5 G.E. (Jameson to Spadina) Dec. - west from Yor'K to Spaa:ina.- 1963 1.69 594 480 3.5 I University subway, D. V. P. (Eglinton to Lawrence) G.E. (East - I Spadina - York) - 1964 1. 74 622 504 3.5 ! G.E. (York to Don River) D.V.P. (Bloor to Gara:iner) - 1965 1.80 633 521 3.5 - 3.3 Bloor-Danforth subway (Keele to Wooanine) D.V.P.(Lawrence ~ 0 401) 1966 1.88 695 571 - 1967 1.93 692 568 3.4 D.V.P. (401 to Shephard) Bloor-Danforth subway: Islington to Warden - 1968 1. 98 785 706 2.8 - 1969 1. 99 825 742 2.7 ' - 1970 2.01 802 711 2.8
1971 2.09 844 749 2.8 .
712 3.0 ! 1972 2.14 803 - 1973 2.20 889 788 2.8 - b/ w !!_/ 1954-67 Metro Toronto Key Facts. - Sources: w 1968-73 Metro Police Dept. 1954-67 Metro Toronto Key Facts. Annual Statistical Report. 1968-73 Statistics Canada 53-219, "The Motor Vehicle". - 34 -
would be useful to compile data on residence location and work place
density over the entire metropolitan area for the same 25 year period
so that potential travel demands could be estimated. Finally, it would
be interesting to calculate changes in transportation access among many
residential and work place zones within the metropolitan area and changes
in this access over time, to be compared with the housing and work place
data.
V. Conclusions
The first issue addressed in this study was the change in transit
performance from replacing the Bloor streetcar service with a subway. The
faster scheduled speed of the subway means that travellers whose origin
~nd destination are at subway stations will find that the subway offers
better service for all but the shortest trips, with time saving increasing with trip length. The service is worse only for trips under one mile in
length if times are valued equally, or two to three miles if waiting and
walking time are more highly valued. Travellers originating between stations,
however, will find that their evaluation of the two modes depends strongly upon the length of their trip and the station spacing which is much greater
for a subway than for a bus or streetcar line. Even if times are valued equally, a trip of three or four miles may be made more rapidly by street car than by subway. For travellers originating between stations, when waiting and walking time are valued more highly than travel time, the
streetcar may offer better service for trips of up to 5, 6, or even 10 to
12 miles. Thus a significant number of transit riders may find service worse
rather than better when surface lines are replaced by a high quality subway - 35 -
line.
The second issue addressed was the effect on Bloor and Danforth vehicle traffic levels of replacing the Bloor-Danforth streetcars with the subway. It appears that the number of rush hour vehicles increased by 24% because of better traffic flow after the streetcars were removed.
This is at least some evidence against the hypothesis that rapid transit can effectively reduce motoring.
Finally, this study examines travel patterns to downtown Toronto during a 20 year period from before any subway operation to after com pletion of the Bloor-Danforth lines. During this period two major down town oriented expressways were also built and tremendous downtown office construction was completed. Between 1954 and 1973 transit patronage to downtowTh shows mixed trends with some increases and some decreases.
During the same period, auto travel to downtown increased by 15 to 25 percent. Thus while improved transit facilities may have helped to retain transit patronage, there is little evidence that the transit share of the traffic was actually increased.
These results should not be interpreted as showing that replacing surface bus or streetcar lines with subway rapid transit is not beneficial or not warranted. Clearly many people in Metropolitan Toronto have benefited greatly from the excellent subway service. The results do suggest, however, that the benefits must be carefully calculated and limited. It takes more than just good rapid transit to significantly reduce automobile travel. - 36 -
APPENDIX A
Subway Station Access Times (Walking Seconds)
Time from nearest Additional Time curb to platform to Bloor Street
Station In out
Islington 50 49 35
Lansdowne 63 64
Ossington 68 54 31
Spadina 75 81 33 E/W St. George 68 66 42 N/S St. George 43 42 42
Bay 46 100 65 95 (Cumberland) 'Cumberland) (Bloor) (Bellair) 0 (Bloor) 38 (Bellair) - N/S Yonge 76 74 zero E/W Yonge 99 74 zero
Pape 61 51
Coxwell 59 69 52
Average time to curb 66.5 seconds, based on single average time for each station. - 37 -
Footnotes
* Associate Professor of Economics. The author is grateful for the support provided for this project by a grant from the Connaught Fund of the University of Toronto to the Centre for Urban and Community studies. Dawn Flood provided efficient assistance with the research. Helpful comments on an earlier version of this paper were made by Dave Nowlan and Richard Bird.
1. The Spadina subway extension will cost $220 million for 6.17 miles, or $35.6 million per mile. "Spadina Subway Progress Report No. 3", Toronto Transit Commission, November 1975.
2. "It was estimated that, had GO Transit (Toronto commuter rail road system) not been operating, around 450 additional vehicles would have been travelling westbound along this section of the Q.E.W. (a parallel expressway) between 5:00 and 6:00 P.M. This would, have represented an increase over the measured volume of around 9% ..•••• " GO Transit Evaluation and Alternatives for Expansion, Ontario Department of Highways, January 1969, p. 6.
3. For a summary of these studies see D. Dewees, "The Impact of Urban Transportation Investment on Land Value", University of Toronto - York University Joint Program in Transportation, Research Report 11, April, 1973, and Stanford Research Institute, The Value of Time for Passenger Cars, Menlo Park, California, 1967.
4. Ibid.
5. Canadian personal income per capita in 1961 constant dollars was $1,339 in 1954 and $2,887 in 1973.