Does Transit Work? A Conservative Reassessment--Part 2

Market Share Related to Walking Distance

If we turn to the past, say to the height of the streetcar era in the 1920s, we see that transit played an enormous role in the life of American cities. Not even the most ardent transit opponent questions that historical fact. But even then, transit was not used for all types of trips.

Before the automobile age, the great competitor of transit was walking. This goes back to the very beginnings of public transit:

In the 1920s, walking was the primary mode for a very important type of trip: shopping. In most residential areas, almost every street had a small store, usually part of someone’s house, that sold basic items such as bread, milk and tobacco. Also in walking distance, usually on a major commercial street, were meat markets, bakeries, drug stores, hardware stores, shops that met people’s everyday needs. Only major shopping trips, such as "going downtown" to a department store for clothing or furniture, were made by transit.

In addition, a great number of what are "trips" today were then needs met by home delivery. Instead of going to the grocery store, people phoned in an order and the groceries were delivered to their house, often by a boy on a bicycle. Most homes had the services of a milkman, bread man, ice man and "egg lady." Farmers came around in wagons or trucks in the summer, selling vegetables and fruits door-to-door. There were tens of thousands of "drummers" on the roads, selling brushes – remember the Fuller Brush Man? – vacuum cleaners, and a wide variety of other household goods house-to-house. And, of course, when you got sick, the doctor came to you, you didn’t go to the doctor.

The point is that these types of trips, especially most shopping trips, were never transit competitive, not even in transit’s heyday. Yet today, they make up the single largest category of trips. A 1983 study found that 35.6% of total trips nationally were for shopping, medical or dental visits or other errands. In comparison, only 22.8% of total trips were work related, including commuting.³¹ A regional study (of California) done in 1980 put home-based shopping trips alone at 26.1%, again the largest source of home-based trips. 86.4% of those trips were done by automobile. Interestingly, the past still showed its hand: the next most common mode for home-based shopping trips was walking, at 8.3%. This study gave transit 3.7% of shopping trips.³² Other studies generally agree: the 1983 report cited previously gave transit a 1.1% share of "Family and personal business," down slightly from 1.6% in 1977 (in 1983, 87.9% of such trips were by car).³³

Two Chicago studies make it clear that when the trip purpose is shopping, not even rail transit makes a difference – which again is historically consistent, because in the 1920s, our baseline, most public transit was by rail, on streetcars. The first Chicago study, done in 1970, notes the rising importance of shopping: "Among the nonwork trip purposes, shopping trips exhibited the greatest change in relative importance between 1956 and 1970. In 1970, trips with purpose to shop constituted 12.6 percent of all trips, while in 1956 they amounted to only 5.5 percent of the total…. Accordingly, more shopping trips are being made, and many of the shopping trips that were formerly accomplished by walking (which would not be included in the survey) are being made by auto."³⁴ The latter point is doubly significant: the shift of shopping trips from walking to automobile was still underway in the 1950s and 1960s, and surveys tended not to count such trips when made on foot, thereby ignoring a large and important share of trips now made by car that were not taken from public transit.

Of the shopping trips made in 1970 in Chicago, 96.1% were made by car, in a city with extensive rail transit. In fact, only .3% were made by commuter (suburban) rail and .8% by rapid transit. Bus actually did better, at 2.6%.³⁵ The second study, with data from 1994, confirms the general picture. Looking specifically at CTA’s Orange Line, it found that only 2.2% of trips were for shopping.³⁶

The point is conclusive: shopping and many other types of personal trips are not transit competitive trips today, because they never were. Their nature as non-transit competitive trips goes far, given their high share of total trips, to explain the magic 1% (or 2% or 5%) of total trips made on transit. Further, if transit is today a "failure" for not carrying more trips of this nature, then it was just as much a failure at its peak in the 1920s – when everyone agrees it was a success!

What, then, are transit competitive trip purposes? The most important is trips to and from work, commuting trips. This, too, has a long history, reaching all the way back to transit’s glory days. A study of transit in Boston in 1930 states:

The trend toward commuting as transit’s main function has strengthened with time and is evident in any transit operation in the nation, sufficiently so that one example is enough. The 1994 study of ridership on Chicago’s Orange Line found that 60.6% of the riders were riding to or from work. The next largest category, school trips, was a distant 13.6%.³⁸

In fact, the question should be put the other way: is commuting the only type of trip that we may classify as transit competitive? Our answer is no, and again it comes from history. When electric railways – streetcars, interurbans, and that favorite New England hybrid, the country trolley – first came on the scene in the 1890s, one trip purpose they competed for with gusto and success was entertainment trips. Many an amusement park began as an enterprise of the local trolley or interurban line, which rang up fares by the thousand as it hauled people from the sweltering city to its roller coasters, ponds and picnic groves. "Trippers" brought their own picnics and simply got off the car at a country stop, found a pleasant meadow to lunch in and, not infrequently, walked part of the way back along the line for a constitutional. Streetcars hauled vast throngs to ball games and other public entertainments: the local baseball team was not the "Brooklyn Trolley Dodgers" for nothing. In fact, just taking an open-air car out into the country on a warm summer evening was entertainment, and many people rode for the sheer joy of it. Some companies had "party trolleys" with Victrolas, and early in the century, an enterprising pair of newlyweds made their honeymoon journey from Delaware to Maine entirely by trolley car. They even wrote a book about it!

Is it fair to argue that today’s transit should be able to compete in the entertainment market? Most transit professionals would probably argue not. But we disagree. The same standard that leads us to argue against expecting transit to carry many shoppers – history – argues for measuring its ability in the recreation market. And there is evidence that it can again compete here, from the fast-growing popularity of "heritage" trolley lines, which people ride both for transportation and for fun, to the throngs carried to and from the Atlanta Olympics on MARTA, 25.3 million people in 17 days.³⁹ In Washington, D.C., 70% of all trips to the new MCI sports center are on transit.

Thus we have our definition of transit competitive trips: availability of transit X quality of transit X trip purpose (commuting to work or entertainment). Does transit work? The answer depends on how many transit competitive trips it carries. To see what that answer is, let’s turn to three case studies.

Our three case studies are Chicago’s Metra commuter rail system and the Light Rail systems in San Diego and St. Louis. We chose them because, first, each represents high quality transit. They are not merely rail systems, but well run rail systems. Each should be able to compete effectively for transit competitive trips.

Second, these systems represent the future. Virtually all new rail transit systems in America will be either commuter rail or Light Rail. At costs sometimes exceeding $100 million per mile, Heavy Rail "metros" have essentially priced themselves out of the game. The most recent attempt to build a Heavy Rail system from scratch, Los Angeles’s Red Line, quickly degenerated into a technical, financial and political fiasco. While analysis in terms of transit competitive trips can certainly be applied to Heavy Rail systems, doing so would be little more than an academic exercise, because no more are likely to be built.⁴⁰

Finally, we chose these three systems because they give us different perspectives in time. Metra, which was established as a successor to private railroad commuter service in 1983, is a relatively new commuter rail system, but not so new that its performance is mostly speculation. San Diego offers the oldest "new" Light Rail system, that is, Light Rail in its "second coming" after the decline and near-disappearance of the streetcar. And St. Louis offers the "latest thing" in Light Rail, a system very similar to what any city now considering Light Rail would get (or at least should get). Its one line represents a classic "starter line," and it was built for the eminently reasonable price of $20 million per mile. St. Louis has also pioneered the use of volunteer labor on a modern Light Rail line (as distinguished from Heritage Trolleys), which we believe is a highly important precedent.

Before we look at the first of our cases, we need to add a word about method of analysis. Our goal is to determine how effectively these three high quality systems compete for transit competitive trips, i.e., what percentage of such trips they carry. In a world of ideal statistics, that would be relatively easy. First, we would determine the size of the population served by these three rail systems, based on the "catchment area:" the number of people who can walk to a train station or trolley stop, plus the number who can drive and find parking, plus the number of bus seats serving the train stops (minus some number reflecting reluctance to take a bus, even to a train). Then, other statistics would tell us how many commuter or entertainment trips that population generates. We would count the riders on the trains or Light Rail cars, compare that number with the total, and have the percentage of transit competitive trips carried by transit.

Unfortunately, in the real world, it’s not that simple. Not surprisingly, since we are introducing a new measuring standard in "transit competitive trips," we find the statistics have not been compiled that way. No one has attempted to count transit competitive trips. We hope someone will do so. Many a useful and publishable paper is to be found in doing so. But so far, as best as we can determine, it hasn’t been done.

That leaves us rummaging about in the data as we find it. It is not a satisfactory situation, but it is not a hopeless one either. There are indicators, enough that it is possible to put together a useful picture. Let’s take a look at our three cases and see what we find.

Metra is the nation’s second largest commuter rail system, with twelve lines, 546 route miles, 1.6 billion passenger miles in 1996 and 240 stations serving more that 100 communities.⁴¹ By commuter rail standards, Metra easily meets the availability test for transit competitive trips (Heavy Rail, Light Rail and bus systems generally have better availability than commuter rail, because they are more concentrated in the city’s dense core area). As already noted, Metra also meets the quality test: its 97% on-time performance is the best in the country.⁴²

So how does it do? In short, vastly better than the 1% or 2% or 5% figures usually cited. Chicago’s Central Business District (CBD) remains the economic heart of the region. "More jobs are concentrated in downtown Chicago than anywhere in the region, and more than half a million people commute to those jobs every day.… More than 1.1 million trips are made each day to Chicago’s central area…. Between 50 and 60 percent of trips made to the CBD are on transit (emphasis added)."⁴³ Further, "Commuter rail has become the most prevalent form of transit for CBD work trips, followed by bus and rapid transit."⁴⁴ Specifically, Metra carried 21% of CBD commuting trips in 1990.⁴⁵

A look at transit trends in Chicago makes the success of Metra and the high quality service it represents even more clear. Between 1980 and 1990, "transit’s CBD market share for work trips dropped from 67% to 55%."⁴⁶ The principal drop was in bus riders; rapid transit fell slightly; and

In other words, people were showing preference for one rail system over another, in part on the basis of quality. And Metra’s success continues to build: the most recent ridership statistics show Metra ridership up from 73.4 million in 1996 to 75.2 million in 1997.⁴⁸

The importance of transit to Chicago is further illustrated by considering what would happen if the 50 to 60% of trips to the CBD made on transit shifted to automobiles.

One statistic brings transit’s role home to anyone who knows the city of Chicago. "The Dan Ryan/Kennedy Expressways carry 200,000 vehicles per day, while parallel CTA/Metra (rail) lines carry 182,000 riders."⁵⁰ If all 182,000 rail riders drove instead, those Expressways would simply stop, gridlocked.

Metra’s phenomenal success, as the largest and still-rising mode in a transit service that carries 50-60% of all CBD trips, suggests a question: just who is riding those trains? The answer helps dispel some of the false images of transit we criticized in our earlier study. Metra’s riders are overwhelmingly white (86%), well educated (66.7% college graduates), professional/technical (47.9%) or managerial/business owner (27%), and very, very well off. In 1995, 24.9% of Metra’s riders had household incomes of $100,000 or more. Only 5.6% had incomes under $25,000. And -- perhaps the definitive indicator of Metra’s ability to compete for transit competitive trips -- 85.6% of Metra’s riders had an automobile available for the trip, if they wanted to drive.⁵¹

Consistent with our definition of transit competitive trips, in 1996 92.7% of Metra’s riders were on work or business related trips. What about entertainment/ recreational travel? On the surface, the numbers are not encouraging. In 1996, only 2.7% of Metra trips fell into this category, down from 4.4% in 1991 and up just slightly from 2% in 1985.

However, 20% of one-way ticket users were on a social or recreational trip, and 43% of all Metra riders indicated they had used Metra for recreation at least once in the previous six months.⁵² Metra itself sees a major future for this market. Its Marketing Plan for 1997 states, "The recreational market serves as the main focus of our secondary market."⁵³ "Recreational travel promotions include discount coupons and weekend Family Fares extended to include weekdays during the summer. Popular suburban attractions include the Ravinia Festival, Arlington International Racecourse, and riverboat casinos."⁵⁴ It will be interesting to follow Metra’s progress in gaining recreational riders.

Metra is a testimonial to what high quality transit can do in terms of competing for transit competitive trips. To its credit, it is not resting on its laurels. In 1991, Metra adopted a program called "Future Agenda for Suburban Transportation," or FAST. FAST will add service to 50 more municipalities and 100 stations, plus 280 route miles, expanding transit availability – the sine qua non of competitive success. It will also add up to 36,000 new parking spaces, beyond the 18,000 added since 1988. Parking is a major component of transit availability. And FAST will add a whole new dimension of quality service: speed. In addition to its 97% on time service, Metra will attempt to cover 50 miles in 45 minutes. In rush periods, that will make Metra faster than driving. Metra understands that for its upscale clientele, time is money.⁵⁵

Speaking of money, how much does Metra cost? Metra recovered 58% of its operating expenses from the farebox in 1997 – the highest percentage in the country for a commuter rail service – plus 5% of passenger revenues for capital financing. And it did so with fares that were less than 5% above 1983 rates.⁵⁶

In sum, from every perspective, Metra does it right – and it works!

While the San Diego Trolley and Chicago’s Metra commuter rail system both run on rails, the similarity largely ends there. The two cities are very different. San Diego’s CBD is much less important to the region than Chicago’s. While Chicago’s commuting market includes a large proportion of wealthy executives and managers, San Diego’s is tilted toward immigrant laborers. San Diego is also very much part of California’s well-known "car culture." To a visitor from an eastern city like Chicago, San Diego appears to be an endless string of suburbs, connected by roads.

The two rail systems are as different as the two cities. Metra is a long-distance, high speed commuter rail line. San Diego’s Trolley is just that: a streetcar system, though one with a great deal of private right-of-way. Stops are frequent, runs are comparatively short and speeds are much lower: while Metra can talk of "50 miles in 45 minutes," the average speed of the San Diego Trolley is 21.9 miles per hour. San Diego recently began service on its own commuter rail system, the Coaster, but this study will focus on the Trolley.

The San Diego Trolley has much to tell other cities which are considering Light Rail. The San Diego Trolley is the first Light Rail system built in the United States since World War II. Its initial line, running from downtown south to the Mexican border opposite Tijuana, opened on July 26, 1981. The system has steadily expanded since, but its history now reaches back almost 20 years. That is enough time to look at initial promises with some perspective and see whether they have been borne out.

But first, we need to see how the Trolley meets our basic criteria: availability and quality of service. Availability presents a mixed picture. In downtown San Diego, it is excellent. Not only is the Trolley in easy walking distance of the main business district, it also directly serves the downtown entertainment and convention centers, the waterfront and the Santa Fe railroad station, which is the end of the line for the Coaster commuter train (in fact, the Trolley has two direct interchange stations with the Coaster). The only important downtown destination not served is the airport, a line to which is planned.

The Trolley’s availability outside the downtown is mixed. The Orange (east) Line serves many suburbs, the Blue (south) Line fewer. The most recent line extension intersects the critical I-15 corridor, but that corridor is merely intersected, not served directly along its length. Nor does I-5 have a parallel Trolley line except for a 3.5 mile segment from downtown San Diego to Old Town. On the whole, the Trolley serves residents south and east of the city better than those to the north, where the more affluent suburbs generally lie; this is evident in the demographics of the ridership.

In terms of quality, the Trolley rates high overall. The equipment is simple but reliable and comfortable. Track quality is excellent. Stations are basic but adequate, and the "honor fare" system is easy to use. However, two factors detract from quality service. There is a perception of danger from crime. The actual crime statistics are not bad, but the mix of races and social classes on board some lines causes apprehension. This is a growing problem on transit nationally, reflecting decreasing order in the society as a whole; we will come back to this problem in the conclusion to our study.⁵⁷

The other detractor is the time between trolleys, which generally run on a fifteen minute interval; the Blue Line has a 7.5 minute interval during rush hours between Old Town and the Mexican border. In the palmy days of the streetcar, many a system advertised, "Always a car in sight." New Orleans’s St. Charles Avenue line, the last survivor from those happy times, still follows that attractive practice. When a rider has to wait up to a quarter of an hour before the wire sings and a car appears, he starts thinking of getting around some other way. Originally, much of the San Diego system was single track, which limited frequency. Now that it is all double track, frequency should increase.

There is a third way in which the San Diego Trolley does not quite measure up to Metra, though it is in no way San Diego’s fault. The available statistics do not allow us to measure its success in attracting transit competitive trips quite so well. But we can find some useful indicators, and as with Metra, they point to success far beyond one percent.

A first indicator is predictive success. If we look at the original ridership predictions, then compare them with actual ridership, what do we find? Many transit critics have said that original predictions have virtually always been high, as a tactic to "sell" Light Rail. Then, the real numbers come in much lower.

According to a 1982 study, by 1995 the original Blue (south) Line of the San Diego Trolley should have been carrying from 28,000 to 30,000 daily riders.⁵⁸ In fact, by 1991 – four years early – it was carrying approximately 32,000 riders per day. Ridership was up to 34,000 per day by mid-1995 and to 51,135 per day in May 1998 (including 5,080 on the Old Town segment and 5,350 per day on the Mission Valley line). In 1991, ridership for the second line to open, the Orange (east) Line, also "exceeded expectations; ridership on that line was 24,560 per day in May 1998."⁵⁹

A 1992 study offers a number of other indicators as to the Trolley’s success in capturing transit competitive trips. "The Trolley has taken single-occupant vehicles off the road, while increasing transit ridership in its corridors."⁶⁰ In 1985, 29.6% of Trolley passengers previously drove alone; by 1990, that figure was up to 36.9%.⁶¹ "41 percent of trolley passengers (36.1% in 1995) ride by choice, compared to only 26% (24.9% in 1995) for all transit users…. the data seem to indicate that the boost in choice riders for the region depends heavily on LRT (light rail) service."⁶² The study "indicates that a higher proportion of trolley riders earn $30,000 or more than riders of the system as a whole. Taken together, these data seems to indicate that the trolley attracts middle- and upper-middle-income workers, even though they could drive to work."⁶³

The 1992 study also has something to say about how the Trolley works in cost terms. Over its first ten years, as ridership grew continuously, the real cost per passenger dropped from $.91 to $.56, the cost per train mile fell from $6.82 to $5.60 and the cost per car mile went down from $3.47 to $2.23 (all figures in 1982 dollars).⁶⁴ And there is icing on that cake: "The closest the trolley came to breaking even overall was in FY 89 when the recovery ratio reached 95.31 percent…. In FY 89, 90 and 91, the South Line actually ran a profit."⁶⁵ In contrast, in FY 89 the San Diego bus system recovered 42.9% of its costs from the farebox.

What has happened since 1991? A recent report by the San Diego Metropolitan Transit Development Board shows an interesting picture. It compares the vital statistics of transit in San Diego – bus and trolley – in 1976, 1986 and 1996. The contrast between 1986 and 1996 is the most instructive, as Light Rail was in operation in both years. Between 1986 and 1996, service miles grew from 13,687,286 to 21,216,572 (23,990,000 in 1998). That growth included almost a doubling of the Light Rail system. This substantial addition of high quality rail transit worked its usual magic: ridership grew from 35,192,140 to 62,168,114 (72,744,800 in 1998) - substantially more, as a percentage, than the growth in service miles. The growth in total system ridership shows that rail riders are not merely people who previously took the bus; in fact, in 1990, just under 25% of Trolley riders formerly took the bus.⁶⁶ Finally, with all this growth, costs fell. Between 1986 and 1996, the farebox recovery ratio for the whole system – Light Rail and bus – rose from 44.15% to 52.5% (51.9% in 1998), and the subsidy per passenger declined from $.76 to $.69 ($.71 in 1998).⁶⁷

How does the San Diego Trolley perform in competing for the two classes of trips that transit should be able to compete effectively for, trips to work and recreational trips? 44.8% of riders are using the Trolley to go to and from work. This compares to 40.6% of riders of the system as a whole, indicating the Trolley competes for commuters slightly better than the bus system.⁶⁸ We must say we would expect this difference to be higher.

The explanation is not inadequate parking. 47.9% of Trolley patrons walk to the Trolley stop, which is consistent with our concept of availability. 13.9% drive to the Trolley, compared with only 2.6% who drive to take a San Diego Transit bus.⁶⁹ This is consistent with rail transit’s superior ability to draw riders from choice.

But in fact, not many people drive to the Trolley. With an average weekday ridership of 75,692 (in May 1998), the San Diego Trolley offers only 5800 parking spaces. And, on average, only 25% of these spaces are filled.

A possible explanation is reflected in another key measurement of the San Diego’s Trolley’s ability to compete: the percentage of patrons who are riders from choice. In 1995, only 36.1% of the Trolley’s riders had a car available.⁷⁰ That is less than half the figure for Chicago’s Metra. We suggest the reason is the demographics of the area served. As noted previously, the San Diego Trolley offers little service to the wealthier suburbs, which generally lie north of the city. Those suburbs are more likely to generate park-and-ride business, as Metra's experience demonstrates.

If we look at our final indicator and ask how San Diego’s Light Rail system does in competing for recreational trips, we find a surprise: it does remarkably well. In 1995, almost 18% of all trips on the Trolley were for recreation or entertainment.⁷¹ A more recent measurement comes from Super Bowl Week in 1998, when nearly one million people used transit to get to and from Super Bowl events.

Shades of 1910! This picture virtually duplicates the scene at any ball game, amusement park or church revival meeting at the height of the streetcar era. The fact that it can be duplicated in 1998 is clear evidence that rail transit can compete for entertainment trips. It is doing so not only in San Diego, but in Cleveland, in Atlanta – where MARTA’s rail system was the transportation backbone of the Olympics – and wherever else rail serves major recreational and entertainment locations.

In sum, while the data is less solid for the San Diego Trolley than for Metra, it is clear that in San Diego Light Rail is competing for transit competitive trips more effectively than are buses. More than a third of the Trolley’s riders have a car available but choose to take the Trolley instead. As Trolley service is expanded in the northern suburbs, that figure should rise. And the Trolley’s recent gains in ridership – in 1997, it carried 19.5 million people, the highest number in its history – show that it is competing ever more effectively within those portions of the metropolitan area that it serves today.

Go to part 3

Some of these pages may include links to documents in the Adobe PDF format. Please download the Adobe PDF reader if you have not already done so.