Skip to content
Quick Start for:
Texas Performance Review 
Capital Metro 
Chapter 5
 
Transit Operations


Daily bus service is, of course, the cornerstone of Capital Metro's mission.

In fiscal 1997, riders took about 30.6 million trips on Capital Metro.1 Most of these trips were on regular bus routes and the University of Texas (UT) shuttle, but many people also rode the `Dillo around downtown Austin or caught an express bus to and from the suburbs. Other riders rode to work in Capital Metro van-pools, and still others took advantage of the authority's special door-to-door services for people with impaired mobility.

The total operating cost connected with providing these services was $70.3 million in fiscal 1998. In that year, Capital Metro took in $7.9 million in fares from its various services, or a little more than 10 percent of its costs. The Operations Division, which manages the day-to-day affairs of Capital Metro's transportation services, is the authority's largest unit, accounting for $53.4 million or three-quarters of its total operating costs. Salaries and benefits represented $30.8 million of that cost, with another $13.4 million going towards contracted transportation services.2

Capital Metro contracts with StarTran for all of its operations staff, including more than 400 full- and part-time bus drivers and 95 mechanics. This contracting arrangement is necessary for Capital Metro to comply with a federal law that requires transit companies receiving federal transit funds to protect the collective bargaining rights of their employees. State law, in turn, prohibits Capital Metro from engaging in collective bargaining; therefore, its unionized employees work for StarTran and are represented by the Amalgamated Transit Union and the International Union of Electronic, Electrical, Salaried Machine and Furniture Workers, both affiliates of the AFL-CIO.

As part of this review, TPR compared Capital Metro's performance with 18 other transit authorities, based on 1995 statistical reports all transit systems must submit to the Federal Transit Administration. This comparison indicates that Capital Metro is a relatively high-volume carrier in terms of services, hours of operation, and ridership. The authority has fewer collisions than are average for its peers. However, Capital Metro's operating costs are relatively high, even when its high-volume service delivery is taken into account.

Capital Metro also reported more roadcalls for repairs than its peers. To investigate this higher-than-average number of roadcalls, TPR examined Capital Metro's maintenance activities and found that its number of miles between roadcalls, a standard measure of maintenance effectiveness, was 25 percent below its peer authorities in 1995 (4,301 miles versus 5,741 miles). However, by fiscal 1997, Capital Metro's Maintenance Department had brought this figure up to 7,425 miles between roadcalls, an improvement of more than 70 percent. The department attributes this largely to a more aggressive approach to preventive maintenance, developed as part of its 1996 business plan. While all Capital Metro departments were recently instructed, as part of the 1999 budget process, to develop departmental business plans, TPR found that Maintenance was the only department that has already implemented a business plan, as well as specific performance measures (including the number of miles between roadcalls).

In examining Capital Metro's higher-than-average ridership figures, TPR found considerable evidence to question their validity. In a selected sample of buses, ridership reported electronically from the authority's fareboxes was as much as 28 percent higher than the number of passengers counted by ride checkers on the buses.3 Capital Metro's management has been aware of this discrepancy since September 1997, yet has taken no action to improve the reliability of this information, even though ridership is the single most important measure of a transit authority's performance.

TPR also found long-standing and unexplained discrepancies in Capital Metro's accounting for its fare revenues. TPR noted significant variances between the amount of revenue reported electronically by the farebox and the actual money count performed by authority employees. For one recent five-month period, these two tabulations of fare revenue were not reconciled at all.

TPR also discovered that Capital Metro supports a number of unproductive, little-used routes in an effort to provide service throughout its entire area. The authority has no strategic plan or cost-effectiveness criteria to help it decide where and when to expand or cut service. Service decisions instead are based on increases or decreases in ridership and revenue miles and hours (the number of miles involved in active service routes and the total number of hours in which the authority's buses are available to pick up passengers, respectively), with little attention to cost-effectiveness. In fact, TPR found that some services cost Capital Metro as much as $27 per passenger per trip, with each passenger paying only 50 cents of that amount.4

To address such problems, TPR recommends that Capital Metro take immediate steps to improve the reliability of its ridership numbers and to better control and account for fare revenue. If and when the authority purchases new fareboxes for its buses, it should favor proven, reliable technology, not simply the latest expensive, high-tech farebox on the market.

TPR also recommends that Capital Metro develop a strategic plan and the criteria and goals needed to improve the overall cost-effectiveness of its bus services, and that the authority immediately cut certain unusually high-cost services to show that it is serious about becoming a more cost-conscious operation.

UT Shuttle

Since 1988, Capital Metro has provided shuttle service for the University of Texas through contracts with various private transportation companies. Since 1991, Capital Metro's contractor has been Dial Activated Vehicle (DAVE) Transportation Service, Inc. As part of the contract, Capital Metro provides DAVE with a fleet of 88 buses, major bus repairs, fuel, and a wide variety of "fixed costs." The service is a net money-loser for Capital Metro; UT will pay Capital Metro $4,252,398 for the service in fiscal 1998, while Capital Metro will pay DAVE $6,310,097 in the same year.5 When other costs are considered, such as indirect administrative costs and depreciation on vehicles, Capital Metro's taxpayers are subsidizing UT's shuttle program by nearly $2.4 million in fiscal 1998. TPR recommends that Capital Metro's next contract with UT, to be negotiated in 2003, should fully recover the program's costs.

TPR also learned that irregularities in the bidding process leading up to Capital Metro's present contract with DAVE may have contributed to the odd fact that DAVE was selected despite the existence of a substantially lower proposal from another company. TPR recommends that Capital Metro ensure that its next contract solicitation for these services, due in 1999, selects the best vendor at the lowest cost possible.

TPR also found that Capital Metro's habit of aggregating data concerning the shuttle program into systemwide performance statistics exaggerates the authority's efficiency and makes it difficult for the board to make informed decisions concerning the future of this service. TPR recommends that Capital Metro begin collecting and analyzing UT shuttle ridership data and other information separately, and that systemwide performance data be reported both with and without the UT shuttle data.

Alternative Transportation Operations

Capital Metro's Alternative Transportation Operations (ATO) Section manages several programs including a van-pool program that provides authority vehicles to groups of commuters for a modest per-passenger fee. TPR found that ATO is not taking advantage of volume purchasing discounts; doing so could save the authority several thousand dollars annually.

In all, the recommendations in this chapter would save Capital Metro more than $1.3 million annually or more than $6.7 million over the next five years; actual savings should be considerably greater, but many of these cannot be estimated.

PROPOSAL 22

Improve the reliability of ridership data and the control and accounting of fare revenue.


Background

Capital Metro's fareboxes, electronic machines mounted in each bus, perform two key functions. The first is to collect, count, and record fares from the passengers, whether coins, dollar bills, or tickets. The second is to record the number and type of passengers boarding each bus, based on information keyed in by the driver. These farebox data on passengers form the primary source for Capital Metro's overall ridership figures.

The drivers are responsible for keying in data on all passengers on all regular routes directly operated by Capital Metro. Certain services, such as the UT shuttle, are contracted to a private provider with special provisions for the contractor to estimate ridership. Since UT shuttles do not collect fares, their ridership is estimated through random samples taken during 10 to 12 days each semester. Ridership on other contracted services (such as TeleRide) is based on data recorded manually by vehicle operators. These counts then are used to establish averages from which estimates of annual ridership are developed.

Capital Metro estimates that its total ridership, as measured in passenger boardings, was more than 30.6 million in fiscal 1997 (Exhibit 12). Capital Metro's fixed route bus service accounted for almost 19 million boardings or 62 percent of the total. The second-largest segment, with 7.4 million boardings or almost a fourth of all ridership, is the UT shuttle operation.6

EXHIBIT 12
Capital Metro Ridership
Fiscal 1997


Type of Service Number of Passenger
Boardings
Percent of Total Passenger Boardings
Fixed-route bus service 18,982,029 61.9%
UT shuttle service 7,442,125 24.3
Downtown Dillo 1,028,449 3.4
Park-n-ride service 496,199 1.6
Special transit service 483,492 1.6
Other 2,192,978 7.2
Total 30,625,272 100%

Source: Capital Metro, Planning Department.

In fiscal 1998, Capital Metro expects to collect about $3.5 million from the sale of tickets, passes, and money collected in fixed-route bus fareboxes (this figure does not include fares associated with contracted transportation services such as the UT shuttle).7 Tickets are pre-purchased individual bus rides, while passes provide monthly unlimited rides.

The standard fare for adults on Capital Metro fixed-route buses is 50 cents, while express routes are $1. However, many passengers ride for free, such as senior citizens, people with disabilities, preschoolers, UT students, police officers, and Capital Metro employees, retirees, and family members. Other trips are free because the route itself is free, such as the Downtown Dillo service; or because the entire Capital Metro system is free at a particular time, such as on Ozone Action days. Passengers boarding with tickets or passes ride for half-price or less of regular fares. Most regular school-age students ride for half-price as well.

Ridership data unreliable

Capital Metro's ridership appears to compare extremely favorably with other transit systems. TPR's review of 1995 data from 18 similar transit authorities found that Capital Metro's overall ridership was more than 76 percent higher than the average of its peers. Even after UT shuttle ridership is excluded, Capital Metro's reported ridership was almost 30 percent higher than the average of the other authorities in 1995.8

TPR found, however, that Capital Metro's systemwide ridership numbers are questionable at best. The authority's ridership numbers are based largely on data electronically keyed in on each bus by the driver. These data may be corrupted in many ways, including both equipment defects and human error; yet TPR found that Capital Metro lacks many of the checks that other transit systems routinely use to independently validate the reliability of their farebox ridership data. More detailed sample ridership data is collected by the Planning Department, but the department samples only about three-quarters of 1 percent of the system's riders each year, a sample too limited to allow for meaningful estimates for the system as a whole.

Ridership is one of the most critical statistics used to determine transit system effectiveness. Yet TPR found that the authority's Planning staff has no faith whatsoever in the reliability of ridership data coming from bus fareboxes.9 And the public, rightly or wrongly, continues to believe that Capital Metro runs too many empty buses, despite the extremely high reported ridership numbers.10

Revenue discrepancies

TPR's initial interviews with Capital Metro staff indicated that discrepancies between farebox reports of revenue collections and actual cash counts are significant, persistent, and unexplained. TPR sought an explanation for these discrepancies from the authority's internal auditor (before this position was eliminated), Finance manager, Revenue Collections leader, Accounting supervisor, and fixed-route superintendent.11 None of these managers could explain why farebox counts so often fail to match the actual cash collected. When questioned by TPR, the fixed route superintendent was surprised that the amounts do not match. The former internal auditor indicated that the discrepancies typically differed by at least 25 percent, but could not provide a reason.

Fareboxes and fares

As riders board the bus, each one puts coins, dollar bills, or a ticket into the farebox or shows the driver a bus pass or appropriate identification card. (Capital Metro issues identification cards allowing the bearer to ride free, such as senior citizens and persons with impaired mobility.)

Capital Metro drivers must keep track of about 20 different types of passengers and fares, including students, people with disabilities, Capital Metro employees, and a number of other categories (Exhibit 10). Each farebox is equipped with a numeric keypad facing the driver. The keypad features written descriptions of various types of passengers next to each of ten keys (Exhibit 13). The driver is supposed to key the correct passenger type for each rider.

EXHIBIT 13
Fares and Types of Passengers*
Fiscal 1997


Key Fare Collected Type of Passenger Number of Passenger Boardings Percent of Total Passenger Boardings
0 Free UT student on non-UT routes 982,430 4.8%
1 Free Transfers, Downtown Dillo, Oops Tickets,
Ozone Action days
6,410,490 31.2
2 Show pass Student pass 590,632 2.9
3 Free Wheelchair passengers 73,830 0.4
4 Free Mobility impaired 1,668,944 8.1
5 25 cents Student fare 893,106 4.4
6 Free Capital Metro employee, retiree, or family
member; police officers, preschoolers
909,408 4.4
7 Show pass Adult pass 3,491,687 17.0
8 Free Senior citizens 507,967 2.5
9 N/A Special projects 1,353,047 6.6
No Key 50 cents or ticket Adults paying with money or prepurchased
tickets
3,625,136 17.7

*Passengers on fixed routes, park-and-ride, and the Downtown Dillo.
Source: Capital Metro.

Keypad categories


Key 0 on the driver's numeric keypad is entered for UT students who ride Capital Metro buses other than the UT shuttles. UT students can ride free on all Capital Metro routes, all the time. In fiscal 1997, 4.8 percent of Capital Metro's passenger boardings, or some 982,000, were by UT students riding regular buses.

Key 1 is used to record a number of different types of messengers who ride free, including transfers from another bus; people riding on designated Ozone Action Days; riders on the Downtown Dillo; and passengers who have received "Oops tickets," which are free tickets provided to passengers if Capital Metro makes a mistake such as passing by a waiting passenger, or to the passengers of a bus that breaks down or is otherwise detained. In fiscal 1997, 31.2 percent of Capital Metro's passenger boardings (6.4 million rides) were captured on Key 1. Of these, more than a million boardings were on the free Downtown Dillo service. TPR estimates that Oops tickets accounted for about 20,000 passenger boardings, based on an average of 55 Oops tickets submitted a day during a sample time period. The remaining 6.4 million then either were transfer or Ozone Day passengers. Capital Metro cannot calculate passenger boardings for each of these groups since they are all lumped together on one key.

Key 2 on the keypad is entered for riders with student passes, either regular passes or express passes (good for express routes to and from outlying areas of the city). Students with valid student identification cards or report cards can purchase express passes for $8.50 or regular student passes for $5. The driver is supposed to check that the current month and year appears on the pass; no money is received. Key 2 also records students from the Austin Independent School District's (AISD's) magnet schools who show an "Apple Route Rider" card to ride for free. In fiscal 1997, nearly 591,000 passenger boardings, or 2.9 percent of total ridership, were by students using passes; Apple Route Riders are not counted in this figure.

Key 3 is entered for wheelchair passengers. Wheelchair passengers are supposed to carry a mobility-impairment card, although it is obvious to the driver that the passenger is so impaired even if he or she doesn't have the card. No money is received, as wheelchair passengers ride for free. In fiscal 1997, about 74,000 boardings, or about 0.4 percent of the total passenger boardings, were by wheelchair passengers.

Key 4 is entered for other passengers with mobility impairments or those who present a Medicare health insurance card. Passengers who show an orange (permanently disabled) or red (temporarily disabled) mobility-impairment card can ride for free. Anyone who has a note from a doctor demonstrating either permanent or temporary impairment can obtain a mobility-impairment card from Capital Metro's Customer Service staff. In fiscal 1997, 8.1 percent or nearly 1.7 million of Capital Metro's passenger boardings were made by persons with mobility impairments.

Key 5 is entered for non-UT students who show a student identification card (either their own or Capital Metro's student ID card) and pay a half-price cash fare of 25 cents. In fiscal 1997, more than 893,000 passenger boardings (4.4 percent of the total) were by students paying half-price fares. Note that key 5 is for students paying with money, while key 2 is used for students with bus passes.

Key 6 is entered for Capital Metro employees (active and retired), police officers, and preschoolers five years old or under. These passengers all ride for free. Key 6 also records family members of current and retired Capital Metro employees; one family member of a current or retired employee can ride for free with the authority's Family/Retiree Pass. In fiscal 1997, more than 909,000 passenger boardings, or 4.4 percent of the total, were in this category.

Key 7 is entered for adult passengers with express or regular passes. These passes can be purchased for $17 for an express pass or $10 for a regular pass. Passengers show their pass to board the bus but no money is received by the bus operator. In fiscal 1997, nearly 3.5 million passenger boardings, or 17 percent of total ridership, were by adults using a monthly pass to board the bus.

Key 8 is entered for senior citizens. Anyone 65 years of age or older can present a driver's license, Texas ID card, or Capital Metro-issued senior citizen card to ride for free. In fiscal 1997, about 508,000 passenger boardings (2.5 percent of total ridership) were by senior citizens.

Key 9, for Special Projects, counts the number of riders for special events involving Capital Metro buses, such as the rodeo. (Fares may or may not be collected depending on the event.) In fiscal 1997, Capital Metro counted more than 1.3 million boardings, or 6.6 percent of its total ridership, under the Special Projects category.

The driver does not press a key when a regular adult rider places 50 cents or a ticket into the farebox; the farebox automatically counts these riders based on their fare. In fiscal 1997, Capital Metro recorded more than 3.6 million passenger boardings of this type, representing 17.7 percent of the authority's ridership.

Drivers are required to set the farebox appropriately at the beginning of each route or special run to designate the type of route. For example, the driver may set the farebox for a regular route initially; it would then read a standard fare as 50 cents. If the driver shifts to an express route, he or she should reset the farebox to read a standard fare as $1. The farebox equipment also records the time of day at which each passenger type is keyed in, giving the authority information on when passengers ride as well as how many.

Farebox collections and probe reports

Once a bus operator's route has ended, an attendant drives the bus through a car wash and then delivers it to the "probe shack," a small facility in Capital Metro's bus parking lot. There, the farebox is "read" by an electronic probe connected to a computer that reads the box's contents while it is still in the bus.

The attendant is supposed to read all the buses' fareboxes and generate a daily report that shows the total revenue collected and the number of different types of riders (Keys 1, 2, etc.). However, a Capital Metro service island attendant told TPR that sometimes the fareboxes are dumped without any probe, while on other days the reverse occurs-the farebox is probed but the money remains in the bus.12 Yet this information is used by the Finance Department to track revenue and the Planning Department to track ridership.

Once the data have been downloaded to the computer, the attendant removes the farebox from the bus and places it in one of four large revenue vaults maintained next to the probe shack. The revenue vault automatically opens the farebox (which cannot be opened by the attendant) and empties the money into the vault. Money from all fareboxes is combined in these vaults.

To better understand the farebox process, TPR staff members selected two buses at random, inserted coins into their fareboxes, and compared the probe reports to the actual coins collected. The revenue amount on the probe reports matched the actual revenue TPR staff inserted into the fareboxes, but the number and type of riders entered on the farebox keypads failed to match the number and type indicated on the probe report. TPR staff found that the farebox keys sometimes stuck, counting more passengers than the number entered.

Ridership data collection

The Planning Department is responsible for collecting and analyzing farebox ridership data. A technical support specialist from Planning visits the probe shack every weekday morning and downloads the data from the previous day onto a floppy disk, because the probe shack's computer terminal is not connected to any of the authority's other computer systems.

The technical support specialist then loads the data into Planning's computer. Using Oracle software, she scours the data for errors, makes corrections, and runs a summary report that shows overall totals for the different types of passengers according to the keys on the farebox. The technical support specialist reports finding hundreds of errors in the farebox dataevery day and that, on average, at least 85 percent of the report must be modified.13 Common errors vary from duplicate information to no information at all for a route, or numbers that are clearly too small or too large to be accurate.

The specialist compares these data to historical data for the same routes to determine if figures are clearly outside of the norm. Many corrections are based on historical information for that route. Some variations are due to external factors, such as increases in ridership on Ozone Action days, when all bus services are free. The Planning staff attempts to keep up with all these factors, but has little or no contact with the drivers to determine why certain numbers look wrong or to tell them how to enter data accurately. TPR doubts that accurate corrections can be made without better coordination with the drivers or some other means for determining whether suspect numbers are actually wrong or simply due to fluctuations in daily ridership.

Once these corrections are made, the data are compiled into a summary ridership report that is shared with a number of different departments. The report is routinely forwarded to the System Performance Analysis Department, for instance, for use in compiling the monthly general manager's report to the board. These data also are available to the Planning staff for ongoing analysis.

The revenue room

The revenue room, part of the Treasury function within Capital Metro's Finance Department, is located in a building near the probe shack and revenue vaults. In the mornings, employees from the revenue room use a forklift to lift each of the four receiver vaults and take them to the revenue room, where they are opened and their contents counted.

Two Treasury clerks perform the count from 7:30 a.m. to 11:30 a.m. four days a week (fares collected on Friday, Saturday, and Sunday are counted on Monday). The coins are fed into a coin counter that sorts and counts automatically. Bills are counted manually and banded together in stacks of $100. The money then is bagged and sealed with plastic ties and dated. The clerks prepare a detailed daily summary showing the amount of coins, bills, tickets, and Oops tickets counted. On average, the authority collects about $8,000 in cash daily. The revenue room employees also prepare a four-part deposit slip; two copies are sent to the bank (inserted in the money bags), another is filed by Finance, and the fourth is sent to Accounts Receivable.

Fareboxes for Capital Metro's sedans and vans are locked in the vehicles and removed on Monday and Wednesday nights. The service island attendant supervisor is the only person with the keys needed to remove the boxes from the vehicles; these fareboxes are not probed. Instead, an attendant locks them in a cart and rolls them to the revenue room each Monday and Wednesday night, where the money is counted on the following Tuesday and Thursday mornings.

Counted money is not stored in the revenue room's vault but instead remains in the open revenue room until the armored car arrives to deliver the money to Capital Metro's depository bank, Bank One.

The two employees counting the money wear smocks over their clothes as they do so. Yet revenue room procedures require personnel "to wear pocketless coveralls during all phases of the daily count. Street clothes should not be worn under the coveralls."14 This policy, intended to provide a measure of physical security, apparently is not enforced. TPR watched the revenue room employees count money three times during the review and none of them ever wore pocketless coveralls. Smocks cannot prevent pilferage, since the worker can still reach a pocket.

Two security cameras are mounted in the revenue room, but TPR found that no one monitors them. Before the elimination of the Internal Audit Department, the Internal Audit director reportedly monitored the revenue room counts or would videotape them for later study. Even then, these tapings do not appear to have occurred with any regularity; TPR asked for all videotapes of cash counting for fiscal 1997 and Capital Metro could only locate two. The cash counts presently are conducted without supervision or monitoring of any kind.

Capital Metro's security coordinator monitors the activities of the contractor that provides security services at the authority's main headquarters, and the Austin Police Department officers who work for Capital Metro part-time by responding to security-related incidents and patrolling service areas. Other responsibilities of the security coordinator include the development and evaluation of security processes and procedures, response to security-related incidents, and surveillance of Capital Metro facilities and properties with video and other monitoring equipment.

The coordinator does not monitor the Revenue Room, although the task certainly appears to be related to her other activities.

Total breakdown in reconciliations

According to department procedures, one of the Treasury clerks should reconcile the daily probe reports with the daily cash-count summaries each week, and send a report of this reconciliation, the "probe/revenue analysis report," to the Treasury clerks' supervisor, the manager of Finance, and various other Operations managers, including the fixed-route superintendent and the Vehicle Maintenance supervisor. However, when examining this process in March 1998, TPR found that the Treasury clerks had not performed a revenue reconciliation since September 1997.

The last "daily" probe report for several months was run on November 4, 1997, apparently due to a simple lack of the computer paper needed to print the report. Yet the paper was not ordered until January 8, 1998-more than two months later-and not received until February 23, 1998. Finance personnel simply told TPR that they could not get more paper for several months, but no one at Capital Metro could explain why it took so long to order the paper. TPR observed the probe report printer and there is nothing special about it or the paper it uses, so Capital Metro's "explanation" appears questionable.

TPR followed up on this situation in May 1998 and found that Finance had resumed generating reconciliation reports on April 11, after TPR's initial questions about this issue. Still, it is disturbing to realize that Capital Metro's fare revenue was not reconciled for more than five months due to a lack of commonly available computer paper.

Variances

When they are prepared, the daily probe report and revenue report often do not reconcile properly. TPR examined the last probe/revenue analysis report of 1997, for the period of August 6, 1997 through September 26, 1997. The average variance between the two counts for the entire six-week period was 12 percent. (TPR's analysis excluded Ozone Action days, on which no revenue is collected.)

Each probe/revenue analysis documents daily and weekly variances between the probe report and the revenue report. These variances change wildly and disturbingly from day to day; for example, on August 11, 1997, the probe report indicated that $11,262 should have been collected, but the cash count produced only $8,337, a 26 percent variance. Other days covered in the report showed variances ranging from as little as 2 percent to as much as 76 percent. It should be noted that this sample included several days on which probe reports were obviously wrong (such as a report for one day showing one cent in fare collections).

The fact that probe reports and revenue reports do not reconcile is well-known among Capital Metro employees. In fact, the Capital Metro reconciliation report reviewed by TPR has notes throughout such as "Probe figures have been way off again." Little has been done to correct this problem. The Treasury clerk who prepares the reconciliation reports told TPR that she receives few questions about reconciliation reports. It seems clear that no one is seriously attempting to determine why the reports do not reconcile, or questioning why reconciliation reports should show any variances at all.

In May 1998, TPR obtained a probe/revenue analysis report for one week in May. As of May, reports once again were being prepared daily and reconciliation reports prepared and distributed weekly. Unfortunately, the probe and revenue reports still do not match.

For example, on May 15, 1998, the probe report showed that $8,286 should have been collected, while the revenue report showed $6,715 in actual collections, a 19 percent variance. It appears that some $1,571 in fares went missing that day; yet TPR still finds that little or no effort is being expended to find out why these variances occur, who is responsible for them, or what needs to be done to correct the situation. Whether the problem lies with fareboxes, bus driver error, unmonitored counting staff, or some other cause entirely, Capital Metro must find answers and must control its cash more effectively.

Reasons for discrepancies?

Efforts to understand the reconciliation problem are made more difficult by the fact that money and tickets from all fareboxes are combined together before counting, thus making it impossible to examine the results of a single farebox. The Finance manager indicated that it is possible to single out individual fareboxes for counting and comparison to the probe report, but no one has ever done so before.

When TPR tested the fareboxes, the farebox keys sometimes "stuck," counting more passengers than the number entered by the bus operator. However, the fixed-route superintendent in charge of regular bus operations said that this is not a significant problem.

All drivers are supposed to complete a "defect card" if they have a problem with their fareboxes. Capital Metro indicated that the most common problem is a jammed bill and coin counter. If this occurs, the operator can put the farebox into "coin override" mode and the money will drop in. The driver then must push a key on the keypad to account for the money and rider, while under normal circumstances the farebox would automatically count the rider with no action from the operator. No specific key is assigned for the use of coin override mode, so the driver must pick a key at random, thus registering an invalid ridership figure.

Bus riders told TPR that some bus operators do not enter a key on the farebox each time a passenger boards, but instead enter several at once, after a group of passengers has boarded, highlighting another possible source of false probe reports. One planning staff person recalled that, at one point, a "Driver Alert" notice was issued reminding bus drivers about the importance of accurately keying in passenger information. Accuracy seemed to improve for awhile, but no ongoing effort has been made to encourage diligence in this area.15

Still another contributing factor to variances could be that the 10-key pad is difficult to operate repeatedly without error, although it must be noted that other Texas transits use similar pads. Accurately accounting for the various types of identification cards, passes, and tickets may prove difficult for some bus operators, particularly newer ones who are not yet familiar with Capital Metro's variety of fares. New operators are tested on farebox procedures and must pass the test with 80 percent accuracy; they are allowed one retake of the test.

Bus operators are not held accountable if the farebox probe report does not match the actual money and tickets collected, nor can they be when the actual revenue is combined before the count, as is current practice. Yet the fixed-route superintendent told TPR that it was reasonable to expect bus operators to be accountable for the information they enter into the farebox keypad. In fact, he indicated that trainee operators often are surprised (and no doubt relieved) when they learn that they will not be held responsible for farebox information and the money that should be in the farebox at the end of their shift.

Still another factor in reconciliation problems may be inaccurate fares that are inserted in the farebox without correction by the driver. Capital Metro has commissioned several external performance audits of its bus operators in recent years, and revenue collection is one focus of these reviews. A large random sample of drivers were observed to determine if they would challenge a fare that was less than the correct amount or a pass that was no longer valid. Many drivers failed to challenge incorrect fares (Exhibit 14).

EXHIBIT 14
Operator Recognition of Invalid Payments Capital Metro


Bus operator failed to challenge: 1/96 Review 10/96 Review 12/97 Review
Short Change 62% 79% 39%
Invalid Pass 52% 44% Not reviewed

Source: Infosource International.

StarTran's president responded to the two 1996 reports by indicating that operator performance may be hindered due to the "large number of different fare media (passes, discount cards, I.D. cards, tickets) with which operators must be familiar." He also mentioned a perceived lack of support from supervisors when disputes arise with customers over correct fare payments. TPR could not obtain information on his response to the 1997 report.16

The collection and processing of farebox ridership and revenue data involves a variety of different personnel and departments. The drivers oversee the collection of the fares and key in data on passenger types; the Finance Department counts the money from the fareboxes; the Vehicle Maintenance Department repairs and maintains the fareboxes; Planning compiles the data and checks them for errors; and System Performance Analysis prepares systemwide reports on the data. It is not clear, however, who is ultimately responsible for ensuring the accuracy of this important information.

Farebox investigation

In September 1997, the Internal Audit, Operations, and Planning Departments conducted a study of the fareboxes to investigate inconsistencies in ridership and revenue data. For five days in September, Capital Metro hired temporary "ride checkers" to ride selected bus routes, both morning and afternoon, and record the number and types of passengers and associated fares on those routes. The study team compared the information collected by the ride checkers to farebox data for those routes, and found significant discrepancies.

The audit focused on data from the last four days of the study. The first day's data proved to be to invalid because the temporary ride checkers had problems identifying certain customer types, but this was corrected for the remaining four days. On average, ridership reports from fareboxes and ride checkers varied by 19.8 percent, with the lowest variance at 4.2 percent on one day and the highest at 28.3 percent on another. In each case, the fareboxes indicated more passengers than the ride checkers counted. The probe reports also indicated that $232.50 should have been collected over the four-day period, while the ride checkers estimated collections at $227.50; interestingly, the revenue room cash count showed only $164.55.

Capital Metro believes these discrepancies resulted from a combination of factors. The study team reported that passenger count inconsistencies may have been tied to problems related to the farebox equipment, the probe, or the hardware or software on the computer terminal that captures the data, but could not isolate any specific technical causes. In all, however, the team cited operator error as the biggest problem area. Some drivers simply hit the wrong keys on a regular basis. The authority also indicated that drivers' ridership counts tended to be higher once they knew they were being observed as part of the study.

The study team recommended several steps to address the variance problem, including new procedures for farebox data processing and reporting; a regular preventive maintenance program for fareboxes; driver training and regular monitoring of key entry; and better software support from the Information Systems Department for the computer system that captures and analyzes the farebox data.17 However, in the ten months since the audit was completed, Capital Metro has taken no action on any of these recommendations.18

Planning Department ride checks

The Planning Department collects sample ridership data to study specific routes, particularly those that may have problems. The department samples only about three-quarters of a percent of the authority's ridership each year, a statistically insignificant sample that cannot be used to make valid estimates of total ridership.19

Unfortunately, and rather surprisingly, Planning's detailed examinations of ridership data for particular routes have never been compared to farebox figures for the same route. Yet such a check could be valuable in studying farebox accuracy.

TPR was told that the two sets of data are not easily comparable. For example, the Planning Department's sampling data generally concern a single route, while farebox data are compiled from all the routes a bus travels in a day. Most of Capital Metro's buses do not run on a single daily route, but rather on pairs of routes, switching from one inbound route to another to limit the number of transfers riders must make. Therefore, farebox data are aggregated by bus, while Planning's ridership numbers are aggregated by route.

However, the drivers are told to enter a new key setting at the beginning of each new route. Therefore, Planning could break the farebox data into individual routes for comparison to its route-sampling data. The department simply must compile the data differently than usual. As of July 1998, Planning staff indicate that they are working to develop the appropriate data conversion programming and anticipate it will be operational within two months.20

Ridership may be overestimated

In September 1997, Capital Metro engaged an outside consulting firm, LKC Consulting Services, Inc., to conduct an extensive evaluation of its fixed-route bus system. The study includes an origin-destination survey (noting where people are boarding and exiting the bus) aimed at gathering accurate information on riders' travel patterns.21

In an effort to find some other means of estimating Capital Metro's ridership, TPR contacted LKC to determine if its survey work could shed any light on ridership variances. LKC's contract did not call for validating Capital Metro's overall ridership numbers, but the company agreed to share its weekday and weekend ridership estimates with TPR. These data were for fixed-route service (both direct and contracted services), UT shuttle service collected in fall 1997, the Downtown Dillo, and Park and Ride services.22

Based on this data, TPR estimated annual ridership of 23.8 million boardings-5.2 million, or 22 percent, below Capital Metro's 29.0 million figure based on fiscal 1997 farebox data. (The 29.0 million figure is based only on totals for similar services as those surveyed by LKC, not all of Capital Metro's ridership.) LKC notes that its sampling can result in some undercounting of ridership, particularly on very busy or crowded buses, and also cautions that it is a small sample for estimating overall ridership, particularly given seasonal variations. However, Capital Metro consistently reports its highest ridership in the fall, so one might assume that projections based on fall 1997 figures might overcount, not undercount, total ridership; ridership drops in the spring and even more in the summer.

While TPR's estimate is hardly precise, it falls significantly short of Capital Metro's overall ridership figure. This difference, in addition to other concerns already outlined, further suggests that Capital Metro needs some other means of validating its ridership information.

Reliability key to selecting new farebox

Capital Metro staff is investigating the purchase of an entirely new farebox system. The authority has begun researching options although funding has not been officially budgeted for the purchase.

This acquisition is a major investment that could require two to three years to complete. The authority must decide what capabilities it desires from a new system and what technologies best meet those needs, and then develop specifications and evaluate various proposals. The authority also will be obligated to arrange for appropriate computer and maintenance support; even the actual installation of new units into hundreds of buses will require extensive effort.

Farebox technology can be as simple as a locked fare collection box and mechanical counter that costs less than $800 per unit, or as complex as a high-tech farebox that records and transfers information automatically with minimal driver involvement, at a cost of $11,000 each.23 However, even the most sophisticated equipment still depends in part on the driver and more elaborate systems may involve higher maintenance costs.

Capital Metro favors an extremely sophisticated new farebox system that minimizes driver involvement to reduce the possibility of human error. The model Capital Metro is examining, in conjunction with Dallas' DART, automatically records data from a magnetic card, and costs $8,900 per unit. With more than 400 buses to equip, this could cost the authority about $3.6 million.

TPR suggests that reliability-accurate counting of riders and revenue, as proven by a clear track record of good service with other bus systems-should be the key deciding factor in selecting a new farebox system. The latest technology can be risky if it is not proven in the field. Often systems that work well in subway systems, for instance, do not function as well when mounted on buses that encounter road shocks. Electronic equipment, moreover, does not always adjust well to the dirt and dust encountered on a bus. Maintenance costs must be considered too, particularly if a bus must be pulled out of service when its farebox system goes down.

In the meantime, Capital Metro's farebox data continues to be very questionable. During the interim period before the purchase of any new system, the authority should attempt to improve the accuracy and reliability of its existing equipment. If the data's reliability cannot be improved to a reasonable level, Capital Metro may need to decide whether it should continue to collect ridership data from fareboxes at all.

One alternative would be to simply require drivers to collect an accurate count of the total number of passengers, without recording details on passenger status (regular adult rider, senior citizen, mobility-impaired person, regular student, UT student, special AISD student, preschooler, employee, police officer, special events rider, or transfer). Data on different types of riders could be obtained by sampling a statistically valid proportion of the ridership and extrapolating this to the total number of riders. This method of sampling is used to prepare estimates of UT shuttle ridership, which represents more than 25 percent of the total. Sampling would be a departure from Capital Metro's traditional practice for calculating ridership, but it could provide more reliable overall ridership data, and should be considered.

Practices in other jurisdictions

TPR surveyed five Texas transit authorities on their farebox and revenue collection practices. San Antonio and Corpus Christi both indicated that their farebox reports usually reconcile well with the money and tickets collected. VIA's bus operators are not held accountable for the accuracy of their fareboxes, but Corpus Christi's operators are, to an extent; operators' performance evaluations may be affected by frequent discrepancies. Fort Worth indicated that it has problems with reconciliation, but attributes them, without elaboration, largely to old fareboxes.24

Dallas said its farebox revenue and actual cash counted reconcile consistently with a variance of just 0.2 percent. Dallas investigates any negative variance (in which the farebox report shows more money than is actually counted) for possible mishandling or farebox malfunctions. Dallas also audits individual fareboxes to confirm their performance, because like Capital Metro it combines all farebox proceeds daily. Any major variance in these audit tests results in maintenance and a subsequent retest of the unit.

Dallas' transit system also investigates any reports of operators taking cash over the top of the farebox (taking the money from the passenger instead of letting them drop it in the box), and compares collections by route and operator to look for patterns of possible abuse. For example, if one operator consistently collects much less in cash than other operators on the same route, Dallas would consider the possibility of revenue mishandling. Operators found to be mishandling money are subject to disciplinary action including possible discharge. When investigating such financial discrepancies, the Dallas authority also may investigate the revenue agents in the authority's counting room and the maintenance staff who have access to fareboxes.25

Houston said its farebox revenue figures reflect actual collections to within 1 to 2 percent. Houston's operators are not responsible for farebox receipts, but if significant discrepancies between farebox and actual revenue occur, the authority investigates the problem.26 In 1993, Houston Metro purchased new fareboxes that read a magnetic card that the rider passes through the farebox upon boarding the bus. The farebox registers payment of the fare as well as the type of rider based on information encoded on the magnetic card. While not all of Houston Metro's riders have these cards, the authority is moving in that direction. These new fareboxes have been met with mixed reviews; planning and management staff are thrilled with the data that can be obtained and analyzed from these devices, but maintenance workers have found it difficult to keep up with the frequent malfunctions that occur when sophisticated electronic equipment rides the bus all day.

Houston Metro, in conjunction with its purchase of this new farebox system in 1993, implemented an extensive system of checks to test the validity of farebox data. Even with brand-new equipment, the authority initially experienced many problems, everything from bad sensors on boxes to problem software to driver difficulties with the new system. To check the reliability of its fareboxes, Houston Metro developed a test set of 500 random types of riders and appropriate fares for each. A small sample of buses were checked to see if the fareboxes accurately reported the correct amount of fares and types and numbers of riders. Many adjustments had to be made to the software to ensure accurate readings. Houston Metro continues to monitor its boxes monthly, stating that continuous checking and double-checking of the information is necessary, even with advanced equipment.

Houston Metro reports that its farebox ridership figures are within 3 percent of audited ridership figures. Houston Metro has automated much of its monitoring process and programmed its computers to highlight any apparent inconsistencies in the data, which are then checked against other independent sources of information. For example, significant increases in systemwide ridership are checked to see if the amount of systemwide revenue also has increased. On its Park and Ride routes, Houston Metro staff count the number of cars in each parking lot daily, periodically estimating the number of "Kiss and Rides," in which a rider is dropped off or picked up from the lot. These numbers are compared with farebox ridership counts for these routes. Houston Metro also has employees ride a random number of bus trips each month to double-check the actual number of passengers against the number of riders reported by the farebox.

When discrepancies are found, any corrections or modifications to the data are clearly documented and reported back to the operations staff on those routes. Initially, a driver will be sent for additional training, but if he or she has ongoing problems with unreliable data, the problem is explored more fully in conjunction with the driver's supervisor. Particular focus is placed on ensuring that the drivers appropriately log onto the system because this ties the data back to the driver and allows for cross-checks with the timekeeping and scheduling records. Drivers with ongoing problems can be subject to disciplinary action. 27

Recommendations

A. The general manager and chief financial officer should establish tighter controls of the revenue room, especially during the cash counting. Revenue room personnel should be required to wear pocketless coveralls during cash counting; the revenue room should be monitored by the security coordinator; and money that has been counted should be moved into the revenue room vault.

Capital Metro's policy requiring revenue room personnel to wear pocketless coveralls should be enforced at the risk of disciplinary action. The practice of allowing employees to count cash while wearing a smock over street clothes presents a security threat. Capital Metro's policy specifically prohibits this and the policy should be followed. The revenue room is currently not monitored at all. The security coordinator should immediately begin conducting video monitoring of the daily cash counting with existing Capital Metro equipment. Money that has been counted and is waiting for pickup by the armored car should be transferred into the vault rather than left sitting in the unsecure revenue room.

B. The general manager should immediately set up a cross-departmental team to develop recommendations to improve the reliability of farebox revenue and ridership data and report to the board on necessary changes no later than October 1, 1998.

Capital Metro's process for capturing farebox and ridership data involves a number of different departments that do not coordinate their efforts, so a cross-departmental team would ensure better communication and accountability for the process.

The departments that should be represented include, Transportation, Vehicle Maintenance, Finance, Planning, System Performance Analysis, Information Systems, and any other unit with a relevant role. The team's members should include management representatives and front-line employees, including bus operators, mechanics, treasury clerks, and planning analysts. The general manager should play a key role to ensure that concrete changes are made quickly across all appropriate departments to improve the reliability of the data.

The team should examine the entire farebox data process and develop consistent procedures for obtaining, processing, reviewing, checking, analyzing, and reporting the data. The team should review the farebox audit of September 1997 and consider the recommendations that team made, while examining other transit system's approaches to ensuring more reliable data from their fareboxes.

If the team can document discrepancies and adjust the authority's ridership estimates accordingly, Capital Metro should immediately modify its published estimates.

C. The general manager should require that individual farebox data be audited regularly throughout the year by Internal Audit, with the assistance of System Performance Analysis. The general manager should immediately set a reasonable goal for reducing the error rate for ridership and revenue data by September 1999.

Regular monitoring and auditing of farebox data is the only way to begin to control errors in revenue and ridership data. At least two types of routine audits should be conducted, one with drivers' assistance and involvement and another without their knowledge.

An audit should be conducted immediately to establish baseline statistics on variances in revenue and ridership data. The general manager should set a reasonable goal for reducing these variances.

D. The chief operating officer and chief financial officer should immediately require that activities related to the collection of information from fareboxes be performed daily.

These activities include farebox probing, cash counting, daily cash count summary, the probe/revenue analysis reconciliation, and the daily collection of ridership statistics on the number and types of passengers. TPR found that many of these activities are performed erratically at best. Requiring each of these activities to occur daily would give Capital Metro easily comparable data, allowing it to examine variances quickly and more effectively. Weekend data could be compiled on Monday, but each day should be accounted for separately.

In addition, the probe/revenue analysis reconciliation reports should be distributed to Planning and System Performance Analysis. These departments should be made aware of differences in the probe report and actual revenue, and should use this additional information to assist in verifying ridership statistics.

E. The manager of Planning should validate the accuracy of farebox ridership data on any route whose ridership is being sampled and examined in detail for potential changes in service.

The Planning staff should routinely validate farebox data against more detailed ridership numbers gathered in the course of considering potential changes in service. These comparisons should be reported back to the general manager and the farebox team for use in evaluating farebox ridership data.

F. The general manager should ensure that the farebox team evaluates a wide range of proven and reliable options for replacing the current farebox system, including low-cost, low-technology systems.

This approach should help ensure that Capital Metro gives full consideration to more basic and significantly less costly replacements for its farebox system.

G. The general manager should cease collecting ridership data through the current farebox system if its reliability cannot be improved by September 1999.

Continuing expenditures of staff time and resources on the collection of ridership data whose overall reliability cannot be validated would be pointless. This recommendation would give Capital Metro a year to improve its data collection methods. If this cannot be done, the authority should stop throwing good money after bad and use a more basic system. Drivers should simply count the total number of passengers. The authority then would use statistical sampling to estimate the percentage of different types of riders, as is done for the UT shuttle already, at least until such time as a new farebox system is in place.

Fiscal impact

These recommendations could be accomplished with existing resources.

PROPOSAL 23

Develop a strategic plan and cost-effectiveness criteria for improving bus operations.


Background

Capital Metro's Planning Department evaluates, plans, and designs its transit services. The Planning Department's budget for fiscal 1998 is $982,465 and provides for 20 employees, 15 full-time and five part-time.28 Planning staff collects sample data on ridership on routes to recommend changes in the alignment, scheduling, and level of service to improve the system's overall efficiency.

Service changes are made twice a year, every August and January, with minor changes in May to adjust for the end of the school year. Performance is judged principally by ridership, which in turn is measured by the number of passenger boardings per hour of service. Planning tracks the number of passengers by route and divides that by the route's number of revenue hours (the combined number of hours all buses operate on the route). Routes falling below a certain level are subject to review, realignment, or frequency adjustment.29 Any proposed changes go through an extensive public input process and full board hearing. No changes in service are made without final board approval.

Poorly structured route system

Capital Metro's route system is a frequent target of criticism from the community. A recent Austin American Statesman editorial said that Capital Metro is beginning to improve under the new governing board, "but the buses that are the heart of the operation still run on an outmoded spoke plan that too often requires people to take a bus downtown before being taken to their destination, a trip that can take more than an hour compared to minutes by private auto."30

As the board considers the highly visible and highly contentious issue of light rail, many customers have voiced concern that too little attention is being focused on how difficult the current bus system can be for its riders. Key problems include long waits between transfers and, as mentioned in the editorial, the number of routes that must pass through downtown before reaching their destinations.

As part of its review of Capital Metro, TPR circulated a written survey and conducted six public forums in Capital Metro's service area to allow citizens to express opinions about the authority in person (see appendices). More than 550 people completed the written survey, primarily persons who use the authority's services daily. When asked to pinpoint the major issues facing Capital Metro, the largest single response-cited by 36 percent of the respondents-was routing and scheduling issues.31 Numerous comments from the public forums also addressed routing, such as the following: "The most responsible way to address the issue of accountability and more bang-for-the-buck would be to concentrate on the needs of the ridership and proper routing of the existing bus services."32

Inadequate sampling of routes

According to its 1997 KPMG Peat Marwick LLP performance audit, Capital Metro cannot identify and report basic ridership information on individual routes or on the number of transfers occurring between routes.33 In May 1998, Planning developed a matrix to begin to tracking transfers by route, but overall difficulties related to identifying ridership by route have not improved significantly since 1997. One key problem is that Planning checks less than 1 percent of Capital Metro's passenger trips each year. Planning's "ride the buses and record the number of passengers getting on and off at each stop.34 Ride checks, however, are handled by just one full-time employee and four part-time workers, although the Planning Department has 15 full-time and five part-time positions in all. 35

By its own admission, Planning does not check a statistically significant portion of Capital Metro's bus routes each year. The checking that is done focuses largely on routes that have attracted complaints or have known problems; the sample does not represent a random selection of routes.36 Therefore, the data that are collected provide useful information about individual routes, but cannot be used to draw any valid conclusions about the system as a whole.

Planning's ability to provide useful ridership information is further limited by its computer system. The current system is more than ten years old; the department needs more up-to-date software to analyze ridership data accurately. Planning staff report that some of their analysis applications were rendered inoperable by an upgrade to their Oracle software.37

No service plan

Overall ridership depends on the effective planning of each individual route. Each route must be examined to determine its effectiveness, with clear standards determining when to revise or eliminate an ineffective route. While Capital Metro's planners have developed various standards over the years, these standards have not been adopted by or endorsed by the board, whose decisions appear to have been based largely on a simple desire to increase ridership with little or no examination of costs.

In 1994, Capital Metro purchased a Five-Year Service Plan to guide service adjustments and establish performance measures for bus routes. Capital Metro contracted with an outside firm, BRW Inc. to prepare the plan for $130,000.38 The ridership goal cited in the five-year plan is an average 35 passengers per revenue hour (actual performance for fiscal 1997 averaged 27.4 passengers per revenue hour).39 According to the Planning Department, the 1994 plan was developed without management's endorsement, was never formally adopted by the board, and has never actually been used.40

In 1996, Capital Metro contracted with a second company, Weslin Consulting Services, Inc., to obtain basic information needed to revise its service plan. The total cost of this contract was $192,271.41 Prior to the completion of this study, Planning changed managers and the new manager questioned the accuracy of the data, so this information was never used either-a recurring theme at Capital Metro.42




High operating costs

According to TPR's comparative data, Capital Metro's ridership 1995 was more than 70 percent higher than the average of its 18 peer authorities. Even after UT shuttle ridership is excluded, its ridership is almost 30 percent higher. However, Capital Metro's operating costs per hour, a key indicator of the system's efficiency, also were high-17 percent higher than the 1995 peer average and increased further in 1997 (Exhibit 15).

High ridership alone does not equate to efficiency. Capital Metro puts out more service than most systems, but it costs more per hour as well. This is cause for concern, given that the bulk of Capital Metro's operating costs must be paid from local taxes; they are only minimally offset by money from fares.43
Operating costs rising faster than ridership

Capital Metro often stresses the dramatic increase in its ridership. This picture, however, does not include the costs associated with gaining that ridership. TPR attempted to obtain a more complete picture by comparing Capital Metro's increases in ridership to the rise in its operating expenses over the last ten years (Exhibit 16).

EXHIBIT 16
Ridership and Cost Trends
Fiscal 1988 - 1997




From 1988 to 1992, Capital Metro's ridership rose by 127 percent. This was due largely to Capital Metro's assumption of the UT shuttle service in 1989 and its free fare program of 1990-91. Operating costs rose as well, but only by 47 percent. In the next five-year period, from 1993 to 1997, ridership increases slowed to only 17 percent, but operating costs rose by 50 percent.

Clearly, Capital Metro has seen the end of the period of "easy" ridership growth at least for bus services. Previous boards often continued routes with low ridership in an attempt to guarantee service throughout the entire service area. This approach has perpetuated numerous costly and unproductive routes and a system that continues to grow without regard for the efficient use of public resources.

New board's focus

The new board has focused its attention on controlling the budget and increasing Capital Metro's cost-effectiveness; most of its members were appointed with these goals in mind.

At the request of the board chairman, Capital Metro staff introduced several new performance indicators related to cost-effectiveness in a December 1997 report to the board. This report, the prototype of a new monthly report to the board, provides the cost of each type of service provided by the authority, the portion of this cost covered by fares, and the amount subsidized by Capital Metro.

These data make it clear that Capital Metro's costs per passenger vary significantly, from a low of 88 cents per trip for the UT shuttle to a high of $25.17 per trip for special transit services for persons with disabilities. The amount of these costs covered by fares also varies considerably, from nothing for the free Dillo service to a high of 86 cents per passenger per trip for van-pool services.

The percentage of these costs subsidized by taxpayers varies from 41 percent, for the UT shuttle, to 100 percent for the Dillo and 99 percent for STS and Teleride Services.

EXHIBIT 17
Cost, Revenue, and Subsidy
of Capital Metro Services First Half, Fiscal 1998


Service Types Passenger Trips Cost Per Passenger Trip Revenue Per
Passenger Trip
Subsidy Per Passenger Trip Percent
Subsidized
Local Fixed Route 8,142,827 $ 1.84 $ 0.l8 $ l.66 90%
UT Services 3,338,613 0.88 0.52 0.36 41
Downtown Dillo 479,202 2.35 0.00 2.35 100
Fixed Route Van 425,868 3.39 0.07 3.32 98
Park & Ride 228,535 4.95 0.14 4.81 97
Special Transit Services 199,644 25.17 0.24 24.93 99
Special Events/ Charter 121,666 2.77 0.58 2.19 79
Van-pool 175,002 4.61 0.86 3.75 81
TeleRide 38,369 15.98 0.18 15.80 99
           
Total/Average 13,149,726 $ 2.16 $ 0.27 $ 1.89 88%

Source: Capital Metro.

Ridership and route efficiency

The Planning Department last reported on ridership by route in fall 1997. As stated earlier, Planning tracks ridership by route and earmarks those falling below a certain level for evaluation. Ridership is measured by the number of passengers on a route divided by the number of revenue hours. By this measure, some of these routes clearly are very efficient, while others run many hours for the benefit of only a few riders.

Some routes carry thousands of riders a day. Capital Metro's North Lamar route carried 8,416 passengers a day in fall 1997, or an average of 68 passengers per revenue hour. (Note that the hours in service of each bus on a route are added to compute the route's total revenue hours; thus a route may have far more than 24 revenue hours in a single day.) The Govalle route averaged 5,253 passengers a day and 37 passengers per revenue hour. The Colony Park/Windsor Park route had 3,893 passengers in a day and an average of 76 passengers per revenue hour.

The UT shuttle routes also enjoyed very high ridership: the Intramural Fields route carried 5,288 students to campus in a day, for an average of 68 passengers per hour, while the West Campus route carried 5,247 students in a day, or 116 passengers per revenue hour.

In contrast, some regular routes report fewer than 200 passengers in a full day of service. For example, the Barton Hills route had only 143 passengers in a day, with an average of just eight passengers per revenue hour. The Southwest Oaks route had 167 passengers in a day and an average of 13 passengers per hour.

Most of the express routes also showed limited ridership. The Cameron Road Flyer carried only 42 passengers a day, with an average of 12 passengers per revenue hour. The Oak Hill Flyer carried 56 passengers per day for an average of 10 passengers per revenue hour. The Leander Express had 1,319 passengers per day but still only averaged 12 passengers per hour, due largely to the fact that buses on this route were piling up more than 100 revenue hours each day to provide service to an area so distant from downtown. The Planning staff counters that "passengers per revenue hour" is not a good measure for these routes since by their very nature they involve longer trips and do not stop as often; Planning considers total ridership more significant on such routes.

The worst-performing service that Capital Metro operates is its TeleRide service. TeleRide provides curb-to-curb service, generally from a person's home to a nearby park-and-ride lot. TeleRide services are available in suburbs including Anderson Mill, Leander/Cedar Park, Oak Hill, San Leanna/South, Wells Branch/Pflugerville, and Central Millwood. The six TeleRide zones report from 15 to 150 passengers a day, with 1.2 to 3.7 passengers per revenue hour.44

Cost allocation data needed for service changes

Capital Metro's board needs clear and concise cost information to make decisions on service changes, particularly if it is to meet its goal of improving the authority's cost-effectiveness. Capital Metro's staff need to be collecting this type of information to better evaluate and substantiate their service change recommendations to the board. Such decisions cannot be based on ridership alone.

Currently, cost data are available only on a systemwide basis, but not for individual routes. Meaningful cost information on different services and routes provided by Capital Metro can be obtained only if the costs associated with the service can be isolated and indirect costs that benefit the authority as a whole can be allocated to the service under examination. The Finance Department does not have the cost-allocation model it needs to accurately report detailed information to the board on the different services currently provided. The Finance and Planning Departments are working to develop such a model, but do not anticipate having it in place until later this year.45

Many other transit systems use cost-allocation methods to estimate the costs associated with their various services. Houston Metro assigns costs to different services such as regular local bus service, park-and-ride service, and express service, and uses this information to estimate costs down to the route level. The authority then can develop useful cost indices for individual routes such as "subsidy per passenger trip" (which is simply the operating costs minus the fare revenue, and divided by the number of passengers).46

This type of information would give Capital Metro's managers and the board a much clearer idea of the actual costs associated with each individual route when making decisions about service changes. Capital Metro staff members have met with their Houston counterparts to obtain more information on their approach.

TeleRide services inefficient

One exception to Capital Metro's general lack of detailed cost data is a special Planning Department study of TeleRide services. As noted above, TeleRide provides curb-to-curb service to people in certain suburban areas. Initially, the service was established to build interest in fixed-route service by introducing suburbanites to the authority's Park-and-Ride services, but according to Planning staff, it has become a subsidized taxi service for suburban residents.47 The fare for this specialized service is only 50 cents.

Planning's overall estimate of costs for TeleRide for the first quarter of fiscal 1998 was $16.41 per passenger trip (Exhibit 18). Generally, this is as far as Planning can go in breaking out estimated costs for a Capital metro service. Because TeleRide service is contracted, however, its costs can be analyzed more readily. TeleRide services were provided by the Greater Austin Transportation Company (GATC) for $1.3 million in fiscal 1997, with estimated fare revenues of about $13,000 about 1 percent of its costs.48 Staff analysis estimated the costs for the six TeleRide routes as follows:

EXHIBIT 18
TeleRide Services Cost Per Passenger Trip Fiscal 1998, First Quarter


TeleRide Zone* Cost Per Passenger Trip
Anderson Mill $ 9.84
South Austin 11.47
Cedar Park/Leander 13.50
Wells Branch/Pflugerville 15.04
Central Millwood 16.90
Oak Hill 27.82
Average for all TeleRide Zones $16.41

Source: Capital Metro.

While the average of $16.41 per passenger clearly is high, the breakout of these costs shows huge variation in costs by route area, from $9.84 to $27.82. Despite this variation, all of TeleRide's costs per passenger are higher than those of any of Capital Metro's other services except for federally mandated services for persons with disabilities. One contributing factor to TeleRide's high costs is the fact that 15 percent of its trips are "no-shows," in which the passenger is not present when the dispatched vehicle arrives. Planning has recommended that TeleRide services be eliminated or reduced three times in the last two years (including its current recommendations for service changes for January 1999), but the board has not yet acted on these recommendations.49

This example indicates how critical detailed cost information can be in evaluating the continuation, elimination, or expansion of certain services. Cost per passenger is simply a more compelling and generally understandable measure than passenger boardings per revenue hour.

For example, as stated earlier, TeleRide services report 15 to 150 passengers a day, with 1.2 to 3.7 passengers per revenue hour. While this clearly smacks of inefficiency, knowing that some of Teleride routes cost as high as $27.82 per passenger per trip gives the board a much more concrete tool in its efforts to cut back on unproductive, costly services, particularly when facing pressure from riders to continue the service regardless of the cost.

Opportunity to set new direction

In September 1997, Capital Metro engaged an outside consulting firm, LKC Consulting Services, Inc., to evaluate its fixed-route bus system. This study is to be completed this year at a cost of $457,000. Its key purpose is to develop system performance data to evaluate the effectiveness and efficiency of bus routes. The study includes an origin-destination survey designed to provide accurate information on the travel patterns of Capital Metro bus patrons. The study also will examine buses' on-time performance and determine how often riders must transfer and how long such transfers take. This information can be used in evaluations of various ridership models, demographic analyses of Capital Metro's passengers, and assessments of current and future market share.50 The Planning Department seems hopeful that this study will provide some of the baseline data needed to develop an effective five-year service plan that can be fully endorsed by the board.

For the first time in years, then, Capital Metro should have baseline data on its riders' needs. While these data are important, the study does not address and will not propose any modifications to Capital Metro's ongoing system for collecting and analyzing ridership data. In addition, this study is only a statistical "snapshot" in time. Basic survey work for the study was conducted in fall 1997 and sampled about 8 to 12 percent of the ridership over a five-week period. The usefulness of this data will depend entirely on whether the board becomes actively engaged in setting a clear policy direction for development of the next five-year plan.

Is the board's primary wish to increase ridership or efficiency? Would it prefer to build new ridership in the suburbs, or to increase service to inner-city residents who depend more heavily on public transit? Should the authority focus on increasing ridership downtown or in newly developing job centers outside the downtown area? Most importantly, should Capital Metro provide service to all areas, regardless of the cost, or should cost-effectiveness be a critical deciding factor?

The study should provide excellent data for evaluating such options, but the board, not a consultant, must set Capital Metro's direction and priorities.

Houston Metro's "focus on the passenger"

Houston Metro provides an example of a transit system that has adopted a realistic direction for improving its bus services. Houston Metro's board decided in 1996 to focus on existing customers by increasing service on the most popular routes. Metro's Deputy general manager is quoted in the Houston Chronicle as stating "our `focus on the passenger' effort is designed around folks already using our service, not the hardest to serve. These moves have caused a bigger ridership boost than extending new routes into distant suburbs without a strong ridership base."51

One goal is to allow riders to forget about bus schedules on its busiest routes and simply be confident that a bus will arrive within a few minutes. Metro also has targeted downtown employers with presentations to encourage bus ridership, particularly on their park-and-ride services.

The result is a more efficient system, with fewer empty seats and empty buses. In fiscal 1997, Houston Metro increased its ridership by 7.8 percent while increasing its total miles of service by only 2.9 percent. For fiscal 1998, the authority projects a 13.8 percent increase in ridership based on its experience in the first six months of the year, with only a 1 percent increase in its miles of service.52 Thus a clear direction and strategy from Houston's board has increased ridership without a similar increase in costs due to increased mileage.

TPR does not necessarily endorse or recommend a similar strategy for Austin; however, some clear plan and priorities are needed to guide the authority's decisions. Capital Metro's board has no such plan to guide it in making decisions concerning the evolution of basic bus service over the next few years.

Recommendations

A. The board should develop a clear policy and strategic direction, no later than September 1, 1998, to guide the development of its five-year bus service plan.

Regardless of the fate of light rail, Capital Metro's primary function for the foreseeable future will be its bus services. Capital Metro is investing almost a half-million dollars in an outside study to assist it in developing a five-year service plan. However, neither the consultant or staff can ensure that this plan will receive any more attention than a long series of previous plans that accomplished nothing, unless the board itself sets the policy and strategic direction to guide it. The services plan can outline how to get there, but the board must decide where it wants to go.

The board should provide its staff and the consultants it has engaged with a clear strategic direction to guide the development of its core bus services and their integration into any other services developed in the future. The strategic planning process must set specific priorities for bus operations to strike a balance between increased ridership and cost-effectiveness.

B. The board should develop clear, written goals and criteria for evaluating and improving the cost-effectiveness of its bus services no later than October 1, 1998.

The board must set goals to improve the cost-effectiveness of the system through its new plan, with clear written criteria to measure the performance for each type of service offered by the authority. These criteria will help ensure that the board will be in a position to measure if the staff's implementation of its policy is achieving the goals it hoped to achieve.

For every proposed route change, the Planning Department should estimate a financial impact. The lack of clear cost information presently makes it impossible for the board to make informed decisions. Operating costs have outpaced increases in ridership in recent years and further expansion has become ever more expensive. To make effective decisions regarding service changes and spending priorities, the board must have access to information on costs as well as ridership.

Establishing the new criteria before the beginning of fiscal 1999 would ensure that they could be used to evaluate route changes scheduled to go into effect for January 1999. High-cost, inefficient routes should not be maintained simply to serve a few scattered riders, if the board is to remain true to its goal of making Capital Metro more cost-effective. It is not equitable or responsible to pay more than $25 for one passenger trip when the same money could provide service to ten times as many people on more efficient routes. Clear criteria in the area of cost-effectiveness would allow the board to make tough decisions fairly and reallocate the authority's resources to routes with greater needs and a higher potential for increased ridership.

C. Capital Metro's chief financial officer and Planning Department should collaborate on the development of a cost allocation method to assign costs by service type no later than October 1, 1998.

A good cost allocation method is a must for effective decision-making in a transit system of this size. Several good models exist throughout the industry that Capital Metro could adapt to its needs. This would allow its staff to develop reliable estimates for the costs associated with various changes in service and routes, and in turn enable the board to evaluate these costs against increased ridership.

D. The board should eliminate TeleRide services no later than October 1, 1998.

Capital Metro pays more than $1.3 million a year to provide taxi service in certain suburban areas at costs ranging from $9.84 to nearly $28 per passenger trip, while recouping only 50 cents from each rider. Residents within other parts of Capital Metro's service area do not receive comparable service.

TeleRide clearly is the most inefficient service Capital Metro provides. Eliminating it would send a clear message that the new board is serious about making the authority more cost-conscious.

E. Capital Metro's general manager should reallocate resources within the Planning Department to support the collection of basic ridership data from each Capital Metro route at least twice a year to coincide with the two major annual service changes.

Capital Metro has not devoted enough resources to the basic data collection needed to evaluate the service it provides, or the computer support needed to compile and analyze these data once collected. Yet this basic information is critical to effective routing and scheduling.

The key resources needed include enough additional staff members to perform ride checks on each route at least twice a year. The Planning Department has 15 full-time and five part-time employees, but has devoted only one full-time and four part-time people to this basic function. Additional computer software support also is needed to accurately analyze ridership data. The general manager should reallocate staff resources to ensure an appropriate level of ride checking. The Information Systems Department should make it a priority to provide Planning with the support needed to upgrade its computer systems to process ridership data appropriately.

Fiscal impact

The savings to Capital Metro from eliminating the TeleRide services are estimated at $1,355,000 per year, based on costs for providing this service in fiscal 1997. The loss in revenues from fares is estimated at $13,000 per year. The net savings to Capital Metro would be $1,342,000 per year or $6.7 million over the next five years.

Better cost information should improve the board's ability to cut costs by reducing or eliminating unproductive routes, but the resultant savings would depend upon future board actions and cannot be estimated. The other recommendations in this proposal should be implemented within existing resources.

Fiscal Year Savings Loss in Revenues Net Savings
1999 $1,355,000 $13,000 $1,342,000
2000 1,355,000 13,000 1,342,000
2001 1,355,000 13,000 1,342,000
2002 1,355,000 13,000 1,342,000
2003 1,355,000 13,000 1,342,000

PROPOSAL 24

Reduce costs for providing the University of Texas Shuttle service.


Background

Capital Metro has an agreement with UT to provide shuttle services through August 2003. Passengers on UT shuttle routes make up almost 25 percent of Capital Metro's total ridership. In fiscal 1997, this involved over 7.5 million boardings by UT students on special shuttle routes. The University of Texas at Austin has provided student transportation services since 1969. Initially, UT provided the service by contracting directly with private transportation companies. In 1988, the university entered into an agreement with Capital Metro to provide shuttle service; Capital Metro in turn contracts with private companies for the service. UT's cost for the service is funded by student fees. UT will pay Capital Metro $4,252,398 for the service in fiscal 1998, based on a total of 145,398 annual service hours.53 Hours exceeding 142,500 are paid at a rate of $29.25 per service hour.54

Since 1991, Capital Metro has contracted for shuttle service with Dial Activated Vehicle (DAVE) Transportation Service, Inc. (Laidlaw, another transportation company, acquired DAVE in 1997 but continues to do business as DAVE for the purposes of Capital Metro's contract). As part of the contract, Capital Metro provides DAVE with 88 fully operational, air-conditioned, wheelchair lift-equipped transit buses. Capital Metro also provides major bus repairs, fuel, and "fixed costs" including payroll taxes and other taxes, worker's compensation insurance, group insurance for drivers, vehicle licenses, and facilities costs. Capital Metro's major repair costs totaled $116,434 in fiscal 1997 and $54,580 in fiscal 1998 to date. Capital Metro has budgeted $696,629 for fuel costs for the UT shuttles in fiscal 1998. Fixed costs were budgeted at $2,340,480 for fiscal 1998.55

In all, Capital Metro's contract with DAVE will cost the authority $6,310,097 in fiscal 1998, $2,057,699 more than what Capital Metro will receive from UT-meaning that the shuttle service is a substantial money-loser for the authority.56

Other costs

This $2 million figure represents a direct subsidy of UT's shuttle service by Capital Metro taxpayers. However, other costs actually add to this subsidy amount, including indirect administrative costs, bus replacement cost, and heavy repair costs.

Three Capital Metro employees provide administrative oversight of the shuttle operation, at a cost to the authority of nearly $86,000 a year (Exhibit 19). Capital Metro has assigned one individual as a liaison between Capital Metro and UT. Capital Metro has included the cost of this individual in their budget and UT provides rent free space on the campus from which this individual works.

EXHIBIT 19
Capital Metro Staff Cost UT Shuttle Program


Staff Salary Percent of Time Working on UT Shuttle Administration Adjusted Salary Benefits (40% x Adj. Salary) Costs
UT Liaison $55,255.51 90% $49,730 $19,892 $69,622
Planner 30,377.51 8% 2,430 972 3,402
Planner 36,554.21 25% 9,139 3,656 12,795
Total $122,187.23   $61,299 $24,520 $85,819

Source: Capital Metro.

The annual replacement cost, sometimes called depreciation costs, of the vehicles in the UT shuttle fleet is substantial as well; based on information TPR received on 80 of 88 buses in the fleet, these costs total about $191,724 annually (Exhibit 20).

EXHIBIT 20
UT Shuttle Fleet Projected Annual Replacement Cost


Number of Buses* Life Expectancy Original Cost per Unit Total Cost of all Units Annual Replacement Cost(Total Cost / # of Years of Life Expectancy)** Replacement Cost Less 80% Federal Share***
           
19 Buses 12 years $136,384 $2,591,296 215,941 $43,188
1 Bus 12 years 148,214 148,214 12,351 2,470
3 Buses 12 years 136,384 409,152 34,096 6,819
1 Bus 12 years 148,214 148,214 12,351 2,470
1 Bus 12 years 136,384 136,384 11,365 2,273
32 Buses 12 years 134,761 4,312,358 359,363 71,873
1 Bus 12 years 129,600 129,600 10,800 2,160
1 Bus 12 years 156,910 156,910 13,076 2,615
1 Bus 12 years 156,810 156,810 13,068 2,614
1 Bus 12 years 167,229 167,229 11,429 2,286
19 Buses 12 years 167,230 3,177,370 264,781 52,956
*80 Buses     $11,533,537 $958,621 $191,724

* Capital Metro's Finance Department provided information on only 80 buses in the fleet.
** The cost to Capital Metro to replace a vehicle at today's market value.
***Capital Metro's purchase of buses is subsidized by federal funding for 80 percent of the purchase price.
Source: Capital Metro.

Capital Metro's Maintenance Department reported $116,434 in heavy repair costs for UT shuttle buses in fiscal 1997, and $54,580 for fiscal 1998 through April. TPR projects total costs for fiscal 1998 heavy repairs at about $109,000.

These other costs, then, total about $386,543 annually. Thus Capital Metro's total cost for providing the UT shuttle service in 1998, beyond the payment it receives from UT, is estimated at more than $2.4 million.

Procurement problems

In spring 1995, two companies responded to Capital Metro's most recent request for proposals to run the UT shuttle operation. These were DAVE and TCT Transit Services Group of Knoxville, Tennessee. DAVE was awarded a two-year contract with a one-year option, despite the fact that TCT submitted a lower proposal price ($16.5 million versus DAVE's $18.8 million).57

An official at TCT requested a review of the proposal evaluation process. As a result of this request, Capital Metro's internal auditor conducted a review and completed his report in August 1995. The internal auditor found several improprieties with the procurement process, including:
  • omission of TCT from the original vendor list, despite the fact that TCT had participated in previous solicitations with Capital Metro.
  • selection of the project manager for the shuttle program on the evaluation committee who had working relationships with representatives of both DAVE and TCT, and who should not have participated in evaluating the proposals.
  • influence by this project manager on other members of the evaluation committee; the project manager apparently implied that giving the contract to DAVE was the most reasonable thing to do primarily because they held the previous contract.
  • excessive weight given to a "Reasonableness of Price" proposal evaluation factor, considering the nature of the contract.
  • lack of objective analysis of price as a factor of the proposal. Members of the evaluation committee had concerns about the TCT's low proposal price, but these never were brought to TCT's attention.
  • apparent favoritism. After the board awarded the contract to DAVE at the June 1995 board meeting, a TCT official asked the project manager why TCT lost; the project manager responded that "You just can't beat the home team."
  • The internal auditor recommended that Capital Metro consider canceling this procurement and readvertising it.58 Capital Metro managers acknowledged some improprieties in the procurement process, but did not feel they contributed substantially to the selection of DAVE over TCT.59 Subsequently, Capital Metro's board authorized the general manager to negotiate and execute a contract with DAVE Transportation.

    Again, TCT's proposal prices for each of the three years of the contract were lower than DAVE's. The evaluation committee chose to interpret one of the evaluation factors, "Reasonableness of Price," as a measure of how realistic the proposal was, and apparently decided that TCT's proposal was unrealistic. Interestingly, as pointed out by the internal auditor, the evaluation committee did not consider the relative prices of the proposals as a factor.

    Another approach: A&M's shuttle program

    Texas A&M University has a student shuttle system run directly by the university, with no outside contract. The A&M shuttle system uses about 60 school buses to provide shuttle services for about 40,000 students on and off campus. The service's operating budget for fiscal 1997 was $2.3 million; it has no capital budget. Revenues to support the shuttle include $1.2 million per year in fees from off-campus bus passes and about $800,000 in student fees intended to cover the on-campus shuttle. A&M's Parking Division makes up the difference in costs of about $280,000 annually. The service also makes some money from charters.60

    Shuttle data skews picture

    The UT shuttle routes are significantly more efficient than the rest of the Capital Metro system. For example, the UT shuttle carries about 53 riders per hour. All other routes averaged about 23.5 riders per hour. When averaged together, Capital Metro ends up with an overall system average of 27.4 riders per hour. The inclusion of UT statistics brings the system's overall performance totals up significantly, from 23.5 to 27.4 riders per hour.61

    Capital Metro routinely uses this higher figure to compare itself with other transit authorities. In the peer comparison TPR conducted, the inclusion of the UT data gave Capital Metro higher ridership-per-hour figures than any of 18 peer transit authorities. When the UT data are excluded, Capital Metro's average of 29.4 is still slightly above average, but falls much closer to the peer average of 27.3.

    Thus the inclusion of UT ridership data significantly changes the statistical picture of Capital Metro's effectiveness in comparison with its peers. Although several other transit systems provide university-related services, they are not as extensive as the UT shuttle service and often are managed as independent or stand-alone operations and are not integrated into overall community services.

    Capital Metro has initiated efforts to begin separating these data based upon requests from the new board. However, according to the Planning Department, "passenger trips made by UT students are not readily available for extraction from all contractors' reports."62 This inability to isolate data on such a significant service element is not conducive to informed decisions.

    Recommendations

    A. Capital Metro should routinely and systematically collect and analyze UT shuttle ridership data and information separately.

    This change would ensure that planning and operating decisions concerning the shuttle service are made with a full awareness of each element in its cost and effectiveness.

    Capital Metro also should ensure that, when making comparisons to other transit authorities, it presents systemwide data both with and without the UT shuttle data. Capital Metro also should develop a contingency plan assessing Capital Metro's opportunities and risks if UT were to give the shuttle contract to another vendor. The first plan should be completed no later than January 1, 1999, with periodic updates to coincide with the authority's budget planning process.

    B. When the contract between Capital Metro and DAVE ends in May 1999 and Capital Metro begins the solicitation process for a shuttle service provider, the authority should seek to reduce the cost of providing the service by selecting the best vendor at the lowest cost possible.

    The 1995 contract, which was supposed to end on August 22, 1998, has been extended to May 1999. Capital Metro should seek to provide UT shuttle services in the most cost-effective way possible, particularly since the service is a net money-loser for the authority.

    In evaluating bids, Capital Metro's evaluation committee should make overall price a significant factor. The solicitation process should be conducted in a way that encourages as large a response as possible. The evaluation process should be free of subjectivity and allow for equitable, comprehensive, and objective analysis of all proposals.

    C. When the contract between Capital Metro and the University of Texas comes up for renewal in August 2003, the authority should discontinue subsidizing the shuttle program and pass the full cost of operating the system on to the university.

    Fiscal impact

    Recommendation A can be accomplished with existing resources. Capital Metro should realize savings by focusing its solicitation and evaluation processes on cost-effectiveness, but these savings cannot be estimated.

    Capital Metro's contract with the University of Texas does not end until August 31, 2003. TPR's recommendation concerning the UT contract should generate significant savings starting in 2003.

    PROPOSAL 25

    Strengthen controls over the van-pool program.


    Background

    Alternative Transportation Operations (ATO), a section within the Purchased Transportation Department since March 1998, is charged with developing alternative transportation methods for the authority's entire service area, to reduce single-occupancy car trips and provide transportation choices that satisfy the community's needs and improve its residents' quality of life. ATO pursues its goals by providing and promoting activities such as carpooling, telecommuting, "Guaranteed Ride Home" programs and van-pool services, which provide Capital Metro vans for the use of groups of commuters for a small fee per passenger.

    ATO's fiscal 1998 operating budget is $1,642,627, while the capital budget, used for purchasing vehicles, is $3.4 million.63 Direct costs budgeted for the van-pool program comprise about $1.5 million or more than 90 percent of the total operating budget. Most of the tasks involved in managing the van-pool program are performed by a private contractor.

    Van-pool services

    Capital Metro contracts with Van Pool Services, Inc. (VPSI) for the management of its van-pool program. The current contract, in effect since May 1, 1994, extends through December 31, 1998.64

    The provisions of the contract require VPSI to
    . . . procure and deliver vehicles to drivers, provide full insurance coverage, handle fleet management, audit and invoice monthly gasoline reports, coordinate van-pool driver agreements, obtain driver license records, pay postage for and distribute monthly van-pool passes, provide cards to the van-pool drivers for gasoline purchases and bill van-pool drivers for personal usage, and collect passenger fares.

    VPSI also must enroll, qualify, and train participants, arrange for high-priority vehicle maintenance, supply project administration, and assist in selecting van-pool commuter routes.65

    Van-pool drivers and alternate drivers, as well as their spouses, are allowed unlimited use of the authority's vans. The only limitation imposed on this usage is a requirement that drivers notify and receive pre-approval from VPSI if they will be traveling more than 200 miles round-trip from home.66

    In January 1998, ATO conducted a cost-benefit analysis that compared a continued contractual arrangement for van-pool services versus an in-house operation, and estimated that Capital Metro could save $2 million over four years by bringing the service entirely in house.67 TPR found that this analysis contained questionable calculations and assumptions and so cannot confirm the benefits of in-house operation. Based on this analysis, however, the board recently approved $3.4 million for the purchase of new vans, with the intent of assuming VPSI's functions upon conclusion of the present contract. Capital Metro realized about $278,750 in savings by purchasing the vans (for $2.8 million) through the General Service Commission's (GSC's) Cooperative Agreement program, which purchases supplies, materials, services, and equipment for state agencies and other qualified entities.68

    In March 1998, the ATO coordinator asked Capital Metro's external auditors, Martinez, Mendoza & Colmenero P.C., to perform a new cost-benefit analysis of in-house operation. The auditor presented its analysis to the board on May 11, 1998, concluding that Capital Metro could save $1.1 million in general, administrative, and personnel costs along with contractor profit over four years through in-house management.69 This estimate-almost half the original savings estimated in January 1998-assumed Capital Metro would hire four additional full-time employees for the in-house operation. The board subsequently voted to extend the VPSI contract through December 1, 1998, to give ATO time to make the transition to an in-house operation.

    Recommendation

    Capital Metro should annually assess the cost-effectiveness of its ATO van-pool operations to ensure that the estimated savings of $1.1 million are achieved.

    Fiscal impact

    This recommendation would ensure that estimated savings are not automatically reallocated in the authority's budget.

    PROPOSAL 26

    Improve van-pool fuel management.


    Background

    Through VPSI, ATO manages 111 vans.70 VPSI provides Wright Express credit cards to van drivers for gasoline purchases and audits and remits monthly gasoline report invoices to Capital Metro. Capital Metro pays a $2 monthly fee for each of these cards that is included in the contract's monthly van-pool charges; the authority also pays for all gasoline purchases used during van-pool commutes from home areas to the work site and back. Drivers pay fuel costs for any personal use of the vans.

    According to the VPSI contract, passenger fares should cover all fixed and operating costs for the vans, third-party management expenses, administrative costs, free driver commute transportation, personal use of vans by drivers, spare vehicle ratio, and staffing requirements. Gasoline costs are not included in the fares, however; these are paid entirely by Capital Metro.71

    Unnecessarily high fuel costs

    Van-pool drivers purchase fuel separately at various gasoline stations throughout the service area. ATO receives no bulk purchase discounts off the price of the gasoline through VPSI's subcontract with Wright Express. The authority is purchasing 10,000 to 12,000 gallons of fuel per month at $1.15 per gallon, for an average annual fuel cost of $151,800.72

    The 1995 Internal Audit review noted this lack of fuel discounts and recommended that management amend the contract to allow Capital Metro to take over gasoline purchases.73 No action was taken on this recommendation.

    The Internal Audit review also stated that Capital Metro's contract with Diamond Shamrock for unleaded gasoline for its buses could be amended to accommodate van-pool purchases. TPR found, however, that it is not feasible to do so because the unleaded gasoline is shipped to Capital Metro's bus service island; it would be inconvenient for van pool drivers, who reside and work all over the authority's service area, to fuel their vans only at Capital Metro's facility.

    GSC administers a statewide retail fueling contract that was solicited through a request for proposals (RFP) in 1995 by the state's Council on Competitive Government (CCG). Through this contract, the state realized total savings of $134,783 from April 1996 through February 1998. CCG negotiated and contracted with four vendors to achieve these savings.

    Recommendation

    Capital Metro's fiscal 1999 budget should be reduced to reflect any discounts from retail fuel prices received through a new fuel services contract.

    Because the current VPSI contract is scheduled to end December 30, 1998, ATO released a request for proposal for a gas card. Included in this is a request for information on discounts off retail fuel purchases.

    Fiscal impact

    Because many vendors who provide fueling services do not charge a service fee per van per month like Wright Express, ATO could save up to $2,664 per year (111 vans x $2 per month x 12 months) by seeking a more competitive arrangement. In addition, many vendors offer a discount on retail fuel to customers who meet certain volumes of purchases. These discounts vary, but based on a conservative assumption of a 1 percent discount, ATO would save an additional $1,450 annually. Total annual savings would be $4621. The first-year estimate assumes nine months of savings for a new retail fuel contract $4,114.

    Fiscal Year Savings
    1999 $3,100
    2000 4,100
    2001 4,100
    2002 4,100
    2003 4,100


    Texas Comptroller of Public Accounts Window on State Government
    Contact Us
    Privacy and Security Policy