Section 5: Analysis Procedures for Transit Operations

Transcription

1 Section 5: Analysis Procedures for Transit Operations Basic Fleet Size Relationships Fleet Size Problems Solutions to Fleet Size Problems Transit Scheduling Transit Scheduling power point Transit Demand Analysis Transit Demand Analysis power point Transit Cost Analysis Transit Cost Analysis power point

2 Basic Speed/Fleet Size Relationships For a given route, where policy is used to determine the level of service, the number of vehicles necessary to provide service at a given headway and average speed can be found as follows: 1. Cycle time is the time for one vehicle to make a complete cycle of the route. It is twice the route length (L) divided by the average speed (S) plus layover times at each end of the route. A B L 60 Cycle Time = 2 * Length * Speed 12* L Cycle Time = S 12 * L Cycle Time = + TA + TB S 2. In order to allow for layover and terminal activity, add the layover or terminal time in minutes at each end of the route (TA, TB). 3. Vehicle requirements (NV) can be calculated as a function of cycle time and headways. Number of Vehicles = Cycle Time Headway or, taking the cycle time as given before, Number of 120 * L Vehicles = S * H + TA + H TB H:\Projects\Transit course\fleet relationships.doc 11/14/2006 1

3 Fleet Size as Determined by Policy A transit board of commission can set headways by policy. For example the board can declare that all vehicles operate at 30 minute headway. Once that policy set, then the number of vehicles needed is directly determined given the cycle time on each route. The actual number of vehicles provided is found by rounding up to the next whole number. Thus if there is a cycle time of 100 minutes and a 30 minute headway, four vehicles would be needed. This is somewhat inefficient. The route cycle time could change (a 120 minute cycle time for four vehicles or 90 minutes for three vehicles) or the headway could be changed (to 25 minutes) to make the route more efficient. Fleet Size as Determined by Demand Vehicle needs can also be determined as a function of vehicle capacity and peak point travel demand. The peak point demand is a triple peak, at the peak load point on the route, in the peak direction and at the peak time of day, given in passengers per hour. 1. Calculate the required headway from vehicle capacity divided by the peak point demand (D). Vehicle capacity is the number of seats (C) times the allowable load factor (LF) at the peak point. The equation is multiplied by 60 to convert hours to minutes. Headway C * H = Seating Capacity * Load Factor = Peak Point Demand LF * 60 (in min utes) D * The number of vehicles needed to service a route with a given length (L), terminal times (TA + TB), and average speed (S) can be determined by combining the equation above with the previous equations. 2 * L * D NV = + S * C * LF D * TA C * LF * * TB 60 The capacity of a route can be determined by rearranging terms in previous equations. Note this is the level of peak point demand that can be accommodated on the route (D), not individual vehicle capacity (C). D D m m = = C * LF * 60 H 60 * NV * S * C * LF 120 * L + S(TA + TB) H:\Projects\Transit course\fleet relationships.doc 11/14/ of 7

4 Vehicle Hours, Vehicle Miles Given that you know the number of vehicles, average speeds and lengths of time periods when service is provided, vehicle hours and vehicle miles can be calculated. 1. Calculate daily vehicle hours (VH) as a function of the length of the peak (P) and the length of the base period (B) and the vehicles operated in each period (NVP, NVB). Vehicle Hours Peak = No. of Vehicles Peak * Length of Peak Vehicle Hours Base = No. of Vehicles Base * Length of Base Total Vehicle Hours = Vehicle Hours Peak + Vehicle Hours in Base VH = NVP * P + NVB * B If you want to adjust total vehicle hours by the ratio (PP) of total pay hours to total platform hours (hours of providing service), then the total vehicle hours is as follows: VH = PP * (NVP * P + NVB * B) 2. Daily vehicle miles (VM) are given by multiplying the vehicle hours by the average speed (S). VM = S * (NVP * P + NVB * B) If there is also a ratio (DH) of deadhead miles to total miles you would adjust vehicle miles by that factor as well. VM = DH * S * (NVP * P + NVB * B) H:\Projects\Transit course\fleet relationships.doc 11/14/2006 3

5 TRANSIT ROUTE WORKSHEET DATA Route Length - miles - (L) Average speed - mph - (S) Layover - (TA + TB) COMPUTE Cycle time = (120 * L/S) + TA + TB DATA Peak headway - minutes - (PH) Base headway - minutes - (BH) Seats per bus - (SE) Load factor - (LF) COMPUTE Peak vehicles = cycle time/ph Base vehicles = cycle time/bh Capacity = SE * LF * 60/PH DATA Length peak - hours - (P) Length base - hours - (B) Deadhead - portion - (D) Pay/Platform ratio - (PD) COMPUTE Daily vehicle hours = (Peak veh. * P + Base veh. * B) * PP Daily vehicle miles = vehicle hours * speed * (1 + D)/PP H:\Projects\Transit course\fleet relationships.doc 11/14/ of 7

6 TRANSIT ROUTE WORKSHEET DATA Route Length - miles - (L) Average speed - mph - (S) Layover - (TA + TB) 10 miles 15 mph 8 minutes COMPUTE Cycle time = (120 * L/S) + TA + TB (120 * 10/15) + 8 = 88 minutes DATA Peak headway - minutes - (PH) Base headway - minutes - (BH) 15 minutes 30 minutes Seats per bus - (SE) 50 Load factor - (LF) 1.2 COMPUTE Peak vehicles = cycle time/ph Base vehicles = cycle time/bh Capacity = SE * LF * 60/PH 88/15 = 6 buses 88/30 = 3 buses 50 * 1.2 * 60/15 = 240 passengers/hour DATA Length peak - hours - (P) Length base - hours - (B) 4 hours 12 hours Deadhead - portion - (D).10 Pay/Platform ratio - (PD) 1.1 COMPUTE Daily vehicle hours = (6 * 4) + (3 * 12) * 1.1 (Peak veh. * P + Base veh. * B) * PP ( ) *1.1 = 66 hours Daily vehicle miles = vehicle hours * speed * (1 + D)/PP 66 * 15 * (1 +.10)/1.1 = 990 miles H:\Projects\Transit course\fleet relationships.doc 11/14/2006 5

7 Example: Racine Spreadsheet Analysis, Estimated Route Performance for Next Year Route Red (1) Brown (2) Yellow (3) Blue (4) Pink (5) Grey (6) Purple (7) Orange (8) Green (9) Average Length (miles) Headways Weekdays peak Weekdays base Sat./holidays Speed (mph) Hours peak Hours base Hrs. Sat./holidays Ridership Totals Weekdays Sat./holidays % chg weekdays % chg Sat./hol Calculations Totals % change fares 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Mod weekdays Mod Sat./holidays Weekdays Peak veh (calc) Peak buses Base veh (calc) Base buses Veh-miles Veh-hours Fare revenue $378 $366 $849 $774 $452 $190 $652 $259 $150 $4,069 Cost $1,651 $1,727 $2,417 $2,484 $1,675 $1,683 $1,963 $1,197 $1,250 $16,047 Subsidy $1,274 $1,362 $1,568 $1,710 $1,223 $1,493 $1,311 $938 $1,100 $11,978 Subsidy/pass $2.23 $2.46 $1.22 $1.46 $1.79 $5.18 $1.33 $2.40 $ Oper ratio Yearly cost $421,017 $440,424 $616,312 $633,362 $427,145 $429,188 $500,646 $305,192 $319,811 $4,092,096 H:\Projects\Transit course\fleet relationships.doc 11/14/2006

8 Spreadsheet Analysis, Transit Routes Estimated Route Performance for Next Year Route Red (1) Brown (2) Yellow (3) Blue (4) Pink (5) Grey (6) Purple (7) Orange (8) Green (9) Average Sat./holidays. Buses (calc) Buses used Vehicle-miles Vehicle-hours Fare revenue $222 $212 $307 $586 $239 $152 $482 $140 $0 $2,340 Cost $1,270 $1,327 $1,318 $1,330 $1,288 $1,294 $1,228 $918 $0 $9,971 Subsidy $1,048 $1,114 $1,010 $744 $1,049 $1,142 $746 $778 $0 $7,631 Subsidy/pass $3.13 $3.47 $2.17 $0.84 $2.91 $4.96 $1.02 $3.67 $0.00 $2.46 Operating ratio Yearly cost $72,366 $75,620 $75,106 $75,791 $73,394 $73,736 $69,969 $52,354 $0 $568,336 Total cost =$4,660,432 Performance Pass/bus hr.- daily Pass/bus hr. - Sat H:\Projects\Transit course\fleet relationships.doc 11/14/2006

9 Fleet Size Problems 1) How many vehicles are needed to maintain a 20 minute headway on a route 10 miles long with an average speed of 12 mph? 2) How long should the route be if only four vehicles are available for the route? 3) What would the average speed have to be to maintain a 10 mile route with 4 vehicles at 20 minute headway? 4) For No. 3, what level of demand peak can be accommodated with 50 passenger buses at a 1.0 load factor? 5) How many buses would be needed to accommodate a peak flow of 250 persons (50 passenger buses, 1.0 load factor, speed = 15 mph, 10 mile route)? 6) For No. 5, what is the resulting average headway? 7) For No. 5, what load factor would allow the cutting of one bus? H:\Projects\Transit course\fleet problems.doc 4/5/2006

10 Solutions to Fleet Size Problems: 1) How many vehicles are needed to maintain a 20 minute headway on a route 10 miles long with an average speed of 12 mph? 120 * L 120 * NV = = = = 5 vehicles S * H 12 * ) How long should the route be if only four vehicles are available for the route? 120 * L 120 * L L NV =, 4 =, 4 =, L = 8 miles S * H 12 * ) What would the average speed have to be to maintain a 10 mile route with 4 vehicles at 20 minute headway? 120 * L 120 *10 60 NV =, 4 =, 4 =, S = 15 mph S * H S * 20 S 4) For No. 3, what level of demand peak can be accommodated with 50 passenger buses at a 1.0 load factor? Seats * LF * *1 * H =, 20 =, PP = = 150 passengers/hour PP PP 20 5) How many buses would be needed to accommodate a peak flow of 250 persons (50 passenger buses, 1.0 load factor, speed = 15 mph, 10 mile route)? NV 2 * L * PP 2 *10 * = = = 6.6, S * Seats * LF 15 * 50 *1 15 = use 7 buses 6) For No. 5, what is the resulting average headway? 120L 120 *10 80 NV =, 7 =, H = = 11.4 minutes S * H 15 * H 7 7) For No. 5, what load factor would allow the cutting of one bus? 2 * L * PP 2 *10 * NV =, 6 = =, LF =, LF = 1.11 S * SE * LF 15 * 50 * LF 15 * LF 90 H:\Projects\Transit course\solutions TO FLEET SIZE PROBLEMS.doc 4/5/2006

11 Interpretation of the Examples: With the basic data, it takes 5 vehicles to operate a route that is 10 miles long on a 20 minute headway and with an average speed of 12 mph (Problem 1). This can be reduced to four vehicles by either making the route shorter at 8 miles (Problem 2) or by operating at a higher average speed at 15 mph (Problem 3). With four vehicles the route has a capacity of 150 passengers per hour (Problem 4). If the demand was 250 passengers per hour, the route will require 6.6 buses, this is rounded up to seven (Problem 5). Then the average headway is 11.4 minutes (Problem 6). Finally, if you wanted to operate six vehicles and accommodate the demand, the load factor at the peak load point would be 111% (Problem 7). H:\Projects\Transit course\solutions TO FLEET SIZE PROBLEMS.doc 4/5/2006

12 TRANSIT SCHEDULING Basic Approaches Level of Service Policy Orientation: Provide a basic service level and maintain it or do nothing at all, i.e. 30 minute headways in off-peak 15 minute headways in peak no matter what demand is. Generally used to set minimums for a system or provide enough service to a certain area of the city. Demand Orientation Look at current level of demand and provide enough seats to meet that demand. At times or places where the demand doesn t exist to cover costs of service don t provide it. Management Viewpoint Want to meet demand at lowest possible cost. Costs are proportional to vehicle hours and vehicle miles; travel varies by time of day, day of week, time of year. To accommodate peak demand, this may mean excessive service at other times of the day. It may be in operator s best interest to discourage peak ridership not worth the cost of adding another full time driver and vehicle to be used for a short time. Need to explore methods for peak shaving lower prices during the off peak for certain users or overall, use car pool/van pool service, taxi substitute, etc. Money isn t saved unless an entire run or piece of work is eliminated. Driver Viewpoint Drivers want schedules that are easy to meet, convenient layovers, short work days, convenient starting time, pleasant passengers, good days off, high pay. They need to take breaks during the day and have meal times. Labor contracts with drivers typically specify constraints on operations and scheduling. For example, a drivers working a split shift will need to be paid extra if the work day exceeds a certain amount (i.e. 10 hours). They are paid a spread penalty a percentage of the hourly rate. Contract provisions have a major effects on schedules, costs, spread penalties, layover provisions, split shifts, etc. Example Contract Provisions: Guarantee time: 8 hours Overtime premium: ½ hour for each hour over 8 hours Spread time premium: ½ hour for each hour over 10 hours All breaks less than 20 minutes are paid H:\Projects\Transit course\scheduling.doc 1 of 3

13 Only one unpaid break No pieces of work greater than 5 hours without a 30 minute break Pull out, turn in paid Customer Viewpoint Customers need to be at their destinations at a given time, want to avoid waiting, long travel time, want reliable service, a safe place to wait, schedules that are easy to remember, availability of service at all hours, weekends, holidays. User oriented transit operates directly from origin to destination (no transfers), convenient, reliable schedules, reasonable fare. Disutility of travel from mode choice studies place a high value on out-of-vehicle time, for example: Disutility of a trip = in-vehicle time * out-of-vehicle time + fare/value of time. Out-of-vehicle time (walking, waiting, transferring) is 1 ½ to 7 times as important as in-vehicle time. Data Needs Running Times How long does it take for vehicles to travel each route segment by time of day? How much variation is there in travel times? What is a reasonable time for recovery at the end of the route (layover)? Possible improvements Use passes, speed up loading and unloading, eliminate indirect routes, add express service, skip stops, increase stop spacing, parking restrictions preferential treatment, bus loading bays, signal preemption. Peak Load Point Counts Count the number of people in the vehicles at the route peak load point, in the peak direction and at the peak time of day (triple peak). Conduct on-off counts to give route ridership profile and information to locate peak load point. Service should be set to provide enough buses to accommodate peak load point demand. Policy maximum headway span of service evenings, weekends loading standards maximum load factor (number standing) and/or maximum standing time on the bus Stop spacing H:\Projects\Transit course\scheduling.doc 2 of 3

14 Example: ROUTE PROFILE BY TIME OF DAY AT PEAK LOAD POINT AND BUS ALLOCATION Number of Buses Number of Passengers Midnight 6 a.m. Noon Midnight The above requires a base service of three buses for two straight shifts each, two split shifts and one afternoon tripper. Six vehicles are required and nine pieces of work. The driver and vehicle used in the tripper may have other assignments on other routes. Computer scheduling: Most transit systems use a computer based system to do their scheduling. A good manual schedule will be as close to optimal as a computer based schedule, but will take much longer to prepare and not be easily adapted to changes. For more information see Transit Cooperative Research Program Report 30: Transit Scheduling Basic and Advanced Methods : available from the TCRP web site H:\Projects\Transit course\scheduling.doc 3 of 3

15 Transit Scheduling E. Beimborn, University of Wisconsin-Milwaukee 1 Basic Approaches Level of Service Policy Orientation: Provide a basic service level and maintain it or do nothing at all, i.e. 30 minute headways in off-peak 15 minute headways in peak no matter what demand is. Generally used to set minimums for a system or provide enough service to a certain area of the city. Demand Orientation Look at current level of demand and provide enough seats to meet that demand. At times or places where the demand doesn t exist to cover costs of service don t provide it. 2

16 Management viewpoint Want to meet demand at lowest possible cost. Costs are proportional to vehicle hours and vehicle miles; demand varies by time of day, day of week, time of year. To accommodate peak demand may mean excessive service at other times of the day. Discourage peak ridership? Peak demand may require costs of another full time driver and vehicle to be used for a short time. Peak shaving lower prices during the off peak, use car pool/van pool service, taxi substitute, etc. Money isn t saved unless an entire run or piece of work is eliminated. 3 Driver viewpoint Drivers want schedules that are easy to meet, convenient layovers, short work days, convenient starting time, pleasant passengers, good days off, high pay. They need to take breaks during the day and have meal times. Labor contracts with drivers typically specify constraints on operations and scheduling. Contract provisions have a major effects on schedules, costs, spread penalties, layover provisions, split shifts, etc. 4

17 Customer viewpoint Customers want: to be at their destinations at a given time, to avoid waiting, to avoid long travel times, reliable service, to have a safe place to wait, to have schedules that are easy to remember, to have service at all hours, weekends, holidays. 5 Data needs Running Times: by route segment, by time of day, variation, Peak Load Point Counts: number of people in the vehicles at the route peak load point, in the peak direction and at the peak time of day (triple peak). Policy: maximum headway, span of service evenings, weekends, loading standards, stop spacing 6

18 Example: ROUTE PROFILE BY TIME OF DAY AT PEAK LOAD POINT AND BUS ALLOCATION Number of Buses Number of Passengers Midnight 6 a.m. Noon Midnight 7 How many drivers, vehicles? The above requires: a base service of three buses for two straight shifts each (six drivers) two split shifts (two drivers, two vehicles) one afternoon tripper (one driver, one vehicle). Six vehicles are required and nine pieces of work (drivers). The driver and vehicle used in the tripper may have other assignments on other routes. 8

19 Computer scheduling: Most transit systems use a computer based system to do their scheduling. A good manual schedule will be as close to optimal as a computer based schedule, but will take much longer to prepare and not be easily adapted to changes. 9 Acknowledgements Some of this material was developed as part of work being conducted by the Great Cities University consortium under the lead of the University of Alabama at Birmingham using funds provided by the Federal Transit Administration of the U.S. Department of Transportation. The opinions expressed are the product of independent university work and not necessarily those of the sponsoring agencies or of the agencies supplying data for the project. 10

20 Transit Demand Analysis Purpose To determine impact on ridership and revenues of any change in service policy. Need to know how demand estimates will be used. May not require an elaborate analysis in many cases. Just try it and find out level of demand. Role of MPO or State DOT Generally have expertise in demand forecasts, full scale system wide mode split analysis using computerized techniques. Demand estimates for a new start or major project need to use advanced methods as part of a regional travel simulation while demand estimates for operational changes can use simplified methods and rules of thumb. Relation to Service Design Services should be designed to attract users successful service/user oriented transit. Transit will be successful when it has the following characteristics Concentrated trip ends: Activities that relate to transit should be located close to transit stops. A quality access system: Provide safe, direct and easy access to transit by pedestrians, bicyclists and automobile users. Minimize distances from vehicle door to buildings. Transit oriented street patterns: Permit through routing, direct service, few turns. Control through automobile traffic if necessary. Market orientation: Services are designed to maximize customer satisfaction and needs. Operate directly between origins and destinations without transfers, convenient schedules, competitive price, clean, comfortable vehicles, good user information. Services should be designed to meet the needs of customers. User oriented transit means that there is:: Direct service from the users trip origin to destination No transfers Schedules that match customer needs Reasonable cost Users ride with similar users Good access on both ends of the trip Two factors need to be determined with transit estimates: market size and market share Market Size What is the total number of potential users for your service? Need to define where the market is and then determine how many people this represents. For example, university students, workers in CBD, elderly without a car, etc.. Analysis using census or better yet through data from major trip generators along the route. H:\Projects\Transit course\transit DEMAND ANALYSIS.doc 1 of 13

21 Market size can be estimated using geographic information systems and census data to identify locations of groups that are more likely to be transit users. Groups that are likely transit users such as low auto ownership households, low income households, households with high proportions of elderly or young can be mapped to show locations of potential demand. Transit propensity analysis uses a weighted sum or index to combine several factors. For example, Dr. Steve Polzin of the Center for Urban Transportation Research (CUTR) at the University of South Florida suggests the following equation: Score = 1.0 * Population in zone * Total employment in zone * Number of households with zero vehicles * Service employment in the zone. Transit supportive zones would be places with higher scores. Source: Florida Transit Information System, 2004, Users Guide - Application Transit Supportive Areas, p 3-43 A series of examples are on the following pages Atlanta, Georgia : Identified transit potential based on four factors: Is emand greater than frequency provided? Are there transit supportive land uses? is there an environmental justice concern? Is it a location with congested highways? Source: Regional Transit Action Plan chapter 3, June 2003 La Crosse, Wisconsin: Looked at locations of zero vehicle households, minority populations and low income households. A traffic analysis zone would have a medium potential for transit if it was within one standard deviation of the average value for all zones, if it was beyond one standard deviation for two or the factors it would be rated as high or low potential, if it was beyond one standard deviation for all three factors, it was rated very high or very low. These could then be mapped as show on a following page. La Crosse Propensity table: H:\Projects\Transit course\transit DEMAND ANALYSIS.doc 2 of 13

22 Atlanta Propensity and demand potential maps H:\Projects\Transit course\transit DEMAND ANALYSIS.doc 3 of 13

23 La Crosse Propensity map H:\Projects\Transit course\transit DEMAND ANALYSIS.doc 4 of 13

24 Market Share What percent of the market will use your service? A well designed service aimed at a key market can capture a high share. Travel Behavior Travelers are constrained by time, money, social and family conditions. All travelers in the long run have choices: auto, bus, move residence, change jobs, etc.. Travelers will choose that mode which they believe will minimize the negative aspects of travel. These include total trip time, time spent waiting, walking, transferring, cost of trip, discomfort of vehicle, feeling of insecurity. TECHNIQUES FOR DEMAND ESTIMATES Simple Methods A) None If it is a well designed service with a high potential market, simply begin the service and see who shows up. Need flexibility to quickly add service, or ability to drop it later if it doesn't work. Set goals, label as an experimental route. B) Judgment Based on past experiences, what do you expect the ridership to be? Football pool Rules of thumb First guess methods learned from working with the system Examples: Passengers/Bus Mile = (0.03) * # of residents within service area of route Passengers per day = (1.2) * park-n-ride spaces Park and ride will attract 30% of the market. A 1% increase in fare produces a 0.3% decrease in ridership. A 10% change in frequency of service produces a 5% change in ridership. C) Non-committal survey Would you ride the bus if...? Rely on stated intentions, requires extensive new data collection, useful for modifying ridership estimates from other techniques, stated intention known to badly overestimate ridership. D) Similar Route Method H:\Projects\Transit course\transit DEMAND ANALYSIS.doc 5 of 13

25 Find a similar route and make adjustments for any differences, e.g. population density, automobile ownership, route length, headway difference. Adjustment factors must be derived or assumed. Mode spit analysis As part of a regional travel forecast, mode split models provide estimates of transit use. These tend to not be used for route level demand estimates, simpler methods are used. But they can provide estimates of elasticities if well calibrated with good data. Elasticity Method Definition of elasticity: percent change in ridership divided by percent change in something (headway, fare, etc.). Elasticity is the slope of the demand curve at a particular point, often a simpler method is used, called the shrinkage ratio. The shrinkage ratio normally stays the same at all levels of demand Different ridership groups have different elasticities. Elasticities can be for different things transit characteristics such as fares, headways; other mode characteristics such as gasoline prices, parking costs, etc. Simple Application of Elasticities A transit system has a current ridership of 4,000,000 fares per year at a current fare of $0.75 This yields a current revenue of $3,000,000 per year. If the fare elasticity is -0.3 and the new fare is $1.00 per trip, what happens? The % increase in the fare is 33.3% The % ridership change is -0.3 * 33.3% = -10% The new ridership is 3,600,000 (a 10% decrease) The new revenues are $3,600,000 (a 20% increase) Sources of information: Transit agencies can keep track of ridership changes following changes in other factors. These are most useful when they are isolated events and do not occur at the same time as other changes TCRP report 95: Traveler Response to Transportation System Changes has individual chapters that track how ridership has changed in response to other things. See: (search for 95) H:\Projects\Transit course\transit DEMAND ANALYSIS.doc 6 of 13

26 FIGURE 20. MEAN BUS FARE ELASTICITY VALUES RIDERSHIP LOSS (FARE ELASTICITY) FARE INCREASE U.S. CURTIN RULE EQUIVALENT GREAT BRITAIN (79 VALUES) WEST GERMANY (13 VALUES) OTHER COUNTRIES (5 VALUES) Note: Values from some fare decreases are included in the foreign data, but fare increases predominate. H:\Projects\Transit course\transit DEMAND ANALYSIS.doc 7 of 13

27 RIDERSHIP LEAST SENSITIVE TO FARE CHANGES LARGE DENSE CITIES RAPID TRANSIT HIGH COST OF DRIVING HIGH TRANSIT MODE CHOICE PEAK PERIOD SMALL URBAN AREAS SPARSE TRANSIT SERVICE FEEDER SERVICE LOW COST OF DRIVING LOW TRANSIT MODE CHOICE OFF-PEAK WEEKENDS RIDERSHIP MOST SENSITIVE TO FARE CHANGES Observed Differential Responses to Fare Changes H:\Projects\Transit course\transit DEMAND ANALYSIS.doc 8 of 13

28 AVERAGE FARE ELASTICITIES BY SUBMARKET Type of Fare Change Fare Increase Fare Decrease City Size Populations greater than 1 million Populations 500,000 to 1 million Populations less than 500, Transit Mode Bus Rapid Rail Time Period Peak Off-peak Income Group Less than $5, $5,000 to $14, More than $15, Age Group 1-16 years years years years More than 65 years Trip Purpose Work School Shop Source: Patronage Impacts of Changes in Transit Fares and Services, H:\Projects\Transit course\transit DEMAND ANALYSIS.doc 9 of 13

29 1% DECREASE IN FARE 1% INCREASE IN BUS MILES OR TRIPS AVERAGE 0.6% INCREASE IN PATRONAGE 1% INCREASE IN BUS TRIPS AVERAGE 0.9% INCREASE IN PATRONAGE AVERAGE 0.3% INCREASE IN PATRONAGE FARE CHANGE FREQUENCY OR COVERAGE SERVICE CHANGE SERVICE CHANGE ACCOMPANYING EPRESS BUS INTRODUCTION Figure 17. Patronage increases attributable to transit system changes. H:\Projects\Transit course\transit DEMAND ANALYSIS.doc 10 of 13

30 Changes in level of service (frequency of service) - The proportional change in transit patronage is less than the proportional change in service. Average Bus Service Elasticities Time Period Service Change Peak Off-Peak Headway In-vehicle travel time Source: Patronage Impacts of Changes in Transit Fares and Services, TCRP 95 values: H:\Projects\Transit course\transit DEMAND ANALYSIS.doc 11 of 13

31 H:\Projects\Transit course\transit DEMAND ANALYSIS.doc 12 of 13

32 Acknowledgements: Some of this material was developed as part of work being conducted by the Great Cities University consortium under the lead of the University of Alabama at Birmingham using funds provided by the Federal Transit Administration of the U.S. Department of Transportation. The opinions expressed are the product of independent university work and not necessarily those of the sponsoring agencies or of the agencies supplying data for the project. H:\Projects\Transit course\transit DEMAND ANALYSIS.doc 13 of 13

33 Transit Demand Estimates E. Beimborn, University of Wisconsin-Milwaukee Purpose To determine impact on ridership and revenues from new systems or any changes in service or policy. Need to know how demand estimates will be used. May not require an elaborate analysis. Just try it and find out level of demand. MPO or State DOT generally has expertise in demand forecasts, normally as part of a full scale system wide planning effort. 2

34 Travel Behavior Travelers are constrained by time, money, social and family conditions. Travelers will choose that mode which they believe will minimize the negative aspects of travel. Total time, waiting, walking, transferring, time, cost, discomfort, inconvenience. Choice and captive users consider different factors. 3 Fit method to the problem Demand estimates for a new start or major project need to use advanced methods as part of a regional travel simulation Demand estimates for operational changes can use simplified methods and rules of thumb. Service should be designed to attract users successful service/user oriented transit 4

35 Elements of Successful Transit Concentrated trip ends: Activities that relate to transit should be located close to transit stops. Quality access system: Provide safe, direct and easy access to transit by pedestrians, bicyclists and automobile users. Minimize distances from vehicle door to buildings. Transit oriented street patterns: Permit through routing, direct service, few turns. Control through automobile traffic if necessary. Market orientation: Services are designed to maximize customer satisfaction and needs. Operate directly between origins and destinations without transfers, convenient schedules, competitive price, clean, comfortable vehicles, good user information. 5 User Oriented Transit Direct trip origin to destination No transfers Schedules match needs Reasonable cost Similar users Good access on both ends of the trip 6

36 Transit Ridership Forecasting Methods Two factors: Market size How many people could potentially use the service? This depends on location and quality of the access system Market share What portion of potential users will actually use the service? This depends on the quality of service frequency, travel times, costs 7 Market size What is the maximum number of users that could make the trip? If the market size is small, the ridership will be small no matter how good the service. Must meet the six conditions for transit use connectivity, access, schedule, knowledge, boarding, and security. Examples: number of students at a university that live within walking distance of a bus route that they can take to their destination without a transfer, number of employees who meet similar conditions number of people who work in an area served by an express bus who pass a park and ride lot 8

37 Transit propensity analysis See TCRP 27 Transit Markets of the Future and 28 Building Transit Ridership Use geographic information systems and census data to identify locations of groups that are more likely to be transit users Weighted sum Population density high density Race non-white Gender - female Income low income households Auto Ownership - zero and one car households 9 Atlanta Source: Regional Transit Action Plan /pubs.htm chapter 3, June

38 Demand Potential Demand greater than frequency provided Transit supportive land uses Environmental justice concern Congested highways 11 La Crosse, Wisconsin Based on locations of zero vehicle households, minority population and low income households Medium potential - within one standard deviation of average High or low beyond one standard deviation 12

39 La Crosse Source: org/content/plan s/mtp/mtp.htm, Chapter 5 Appendix D 13 Market share What percent of the market is likely to use the transit service? Useful to do separate estimates of captive and choice users. 14

40 Simple Methods None: If it is a well designed service with a high potential market, simply begin the service and see who shows up. Rules of Thumb: First guess methods learned from working with the system. Based on past experiences, what do you expect the ridership to be? Non-committal survey: Would you ride the bus if.. Relies on stated intentions, known to badly overestimate ridership. Requires sophisticated data collection and analysis and current behavior to get good results. 15 Mode split analysis As part of a regional travel forecast, mode split models provide estimates of transit use Tend to not be used for route level demand estimates, simpler methods are used. Can provide estimates of elasticities if well calibrated with good data. 16

41 Similar Route Method Find a similar route and make adjustments for any differences, e.g. population density, automobile ownership, route length, headway difference. Adjustment factors must be derived or assumed. 17 Elasticity (shrinkage ratio) Method Shrinkage Ratio: percent change in ridership divided by percent change in something (headway, fare, gasoline price, etc.). Different ridership groups and trip purposes may have different numbers. Source: see TCRP Report 95: Traveler Response to Transportation System Change extensive case study data on how transit ridership changes in response to other changes. search for 95 18

42 Example Current Ridership = 4,000,000 Current Fare = $0.75 Current Revenues = $3,000,000 Fare Elasticity = -0.3 New Fare = $1.00 % Fare Increase = 33.3% % Ridership Change = -0.3 * 33.3% = -10% New Ridership = 3,600,000 (10% decrease) New Revenues = $3,600,000 (20% increase) 19 FIGURE 20. MEAN BUS FARE ELASTICITY VALUES FARE INCREASE RIDERSHIP LOSS (FARE ELASTICITY) U.S. CURTIN RULE EQUIVALENT GREAT BRITAIN (79 VALUES) WEST GERMANY (13 VALUES) OTHER COUNTRIES (5 VALUES) Note: Values from some fare decreases are included in the foreign data, but fare increases predominate. 20

43 RIDERSHIP LEAST SENSITIVE TO FARE CHANGES LARGE DENSE CITIES RAPID TRANSIT HIGH COST OF DRIVING HIGH TRANSIT MODE CHOICE PEAK PERIOD SMALL URBAN AREAS SPARSE TRANSIT SERVICE FEEDER SERVICE LOW COST OF DRIVING LOW TRANSIT MODE CHOICE OFF-PEAK WEEKENDS RIDERSHIP MOST SENSITIVE TO FARE CHANGES Observed Differential Responses to Fare Changes 21 AVERAGE FARE ELASTICITIES- 1 Type of Fare Change Fare Increase Fare Decrease-0.37 City Size Populations greater than 1 million Populations 500,000 to 1 million Populations less than 500, Transit Mode Bus Rapid Rail Time Period Peak Off-peak

44 AVERAGE FARE ELASTICITIES -2 Income Group Low Medium High Age Group 1-16 years years years years More than 65 years Trip Purpose Work School Shop Source: Patronage Impacts of Changes in Transit Fares and Services, Service elasticities 1% DECREASE IN FARE 1% INCREASE IN BUS MILES OR TRIPS AVERAGE 0.6% INCREASE IN PATRONAGE 1% INCREASE IN BUS TRIPS AVERAGE 0.9% INCREASE IN PATRONAGE AVERAGE 0.3% INCREASE IN PATRONAGE FARE CHANGE FREQUENCY OR COVERAGE SERVICE CHANGE SERVICE CHANGE ACCOMPANYING EPRESS BUS INTRODUCTION Figure 17. Patronage increases attributable to transit system changes. 24

45 TCRP headway elasticities 25 Acknowledgements Some of this material was developed as part of work being conducted by the Great Cities University consortium under the lead of the University of Alabama at Birmingham using funds provided by the Federal Transit Administration of the U.S. Department of Transportation. The opinions expressed are the product of independent university work and not necessarily those of the sponsoring agencies or of the agencies supplying data for the project. 26

46 Transit Cost Analysis Types of Costs Fixed Cost: does not vary with the amount of service provided in the short run. Variable Costs: change with the amount of service provided. Variable Cost Fixed Level of Service Average Cost: total cost (fixed and variable) divided by output i.e., cost per platform hour. Marginal cost: the change in total cost for each unit of output. Principles of Cost Analysis The purpose of the analysis dictates the way you analyze costs. Analysis of the costs of a service change: use the incremental, marginal cost -- which costs are affected by the change? Need to decide term of analysis next year? 5 years? Analysis of a new system: look at total costs, set up of administration, vehicle purchase, facilities, hiring, training and operations. Allocation of deficit, revenue: take costs and allocate them to individual routes. What is their share of revenue, deficit? Service changes are incremental -- you save money by cutting pieces of work -- value of one to seven hours of saving is zero. The extra cost of using vehicles in off peaks is likely only a portion of the per mile cost, i.e. extra fuel and maintenance costs. Key is what you use it for -- compare the world with it versus the world without it. Look at incremental cost change -- with addition or deletion of services use a total allocation system. H:\Projects\Transit course\transit COST ANALYSIS.doc 9/5/ of 11

47 Cost Model for Transit Planning Use a variety of approaches depending on the situation. From sketch planning, to detailed budget analysis. Simple extreme Average system cost per unit of service, i.e. total cost/hour or total cost per mile, used for quick calculation, analysis of a new system. This method will tend to overestimate potential savings of a service cut or costs of a service expansion since it includes fixed as well as variable costs. Complex extreme Reschedule the whole system to look at the effect of a change, run cutting, used to implement services and for budgeting for next system change. Intermediate Cost models with a limited number of variables per hour, per mile, and per vehicle. Cost Allocation Methods To calculate the costs of transit service, all costs of the system need to be allocated to given categories for example. Cost =.44 * VM = 12 * VH + 15,000 * Veh. VM = vehicle miles VH = vehicle hours Veh. = number of vehicles Need to go through your expenses and allocate each to miles, hours, and vehicles. per hour costs driver wages and fringe. per mile costs maintenance wages and fringe, parts, fuel, tires, accidents (insurance?) per vehicle management, advertising, legal fees, office supplies, training, overhead items, utilities, etc. (Some expenses could fall into multiple categories.) Note that a few items (driver wages and benefits, fuel, maintenance wages and benefits) are the bulk of the costs. Fully allocated model takes fixed costs and attempts to make them variable, especially the per vehicle portion, i.e., if you added 10% to your vehicle fleet administrative costs wouldn't rise by 10%. Think about how the change will take place. H:\Projects\Transit course\transit COST ANALYSIS.doc 9/5/ of 11

48 Transit Cost Allocation Procedures 1 Many performance indicators operating expense per vehicle hour, operating expense per one-way passenger trip, administrative expense as a percentage of total expense, and revenue per one-way passenger mile require accurate financial information. A complete performance evaluation requires revenue and expense estimates not only for the transit system as a whole, but also for the individual service components that are being analyzed. For example, to perform a comprehensive diagnosis of a system's operating problems, the manager needs a separate analysis of each service sector or type of service within the operation to determine if one part of the operation is dragging down the performance of the whole system. The costing issues that must be considered before applying the performance evaluation methodology described in this guide involve two primary topics: cost determination and cost allocation. Cost Determination: Cost determination is the process of identifying the total cost of providing the service. The goal of this process is to produce a statement of the revenue and expenses for the paratransit service for a particular period. The basic source of information for this cost determination is the accrual accounting system that will result in a listing of expenses such as that shown in the following table. Though the example expense listing in the table is for a twelve month period, performance evaluations also use monthly, quarterly, or semiannual information. The accrual accounting system, as contrasted to a cash accounting system, records revenue and expenses when they are due or incurred, rather than received or paid. An accurate performance evaluation requires that the accrual system be used so that revenue and expenses can be properly associated with the services provided and consumed. For example, if the accrual system is not used, an annual vehicle insurance bill paid in one month will overstate expenses and the related financial performance measures for the month when the bill is paid. Likewise, counting revenue in the period when it is received, rather than when it is earned, will improperly represent the true revenue per passenger, or overall cost recovery of the system. In addition to the operating revenue and expense data provided by the accrual accounting system, the system manager may, depending upon the purpose of the evaluation, need to make adjustments to the expense data. The need for such adjustments often arises when the evaluation involves comparing the performance of a privately operated system with that of a nonprofit or public agency-operated system. For this type of comparison, in addition to basic operating expenses, special treatment of costs may be required for costs incurred by the private operator but not by the public on nonprofit agency such as depreciation, profit, and certain taxes. 1 This material is adapted from a NTI course Improving Transit System Performance: Using Information Based Strategies developed at the University of Wisconsin-Milwaukee This material was written by Jack Reilly of the Capital District Transportation Authority (Albany, N.Y.), Edward Beimborn or UWM and Robert Schmitt of RTR Associates in Pittsburg. H:\Projects\Transit course\transit COST ANALYSIS.doc 9/5/ of 11

49 Sample Chart of Accounts Used for Cost Allocation Expense Object Class Annual Expense TRANSPORTATION EPENSE Driver Wages and Salaries $195,000 Driver Fringe Benefits 42,900 Fuel and Oil 42,500 Tires and tubes 6,500 Vehicle Insurance 39,500 Vehicle Lease 6,500 Purchased Transportation 46,900 Other 3,460 TOTAL TRANSPORTATION EPENSE $379,760 MAINTENANCE EPENSE Mechanic Wages and Salaries $23,000 Mechanic Fringe Benefits 4,830 Materials and Supplies 14,600 Contracted Maintenance 26,800 Facility Rental 6,000 Utilities 4,000 Contracted Services 8,900 Other 3,350 TOTAL MAINTENANCE EPENSE $91,480 CALL TAKING AND DISPATCHING EPENSE Dispatcher Wages and Salaries $31,500 Dispatcher Fringe Benefits 6,500 Telephone Expenses 6,600 Computer Expenses 4,200 Rent 3,600 Other 5,400 TOTAL CALL TAKING AND DISPATCHING EPENSE $57,800 ADMINISTRATIVE EPENSE Administrative Salaries $69,500 Administrative Fringe Benefits 15,500 Materials and Supplies 4,500 Nonvehicle Insurance 2,200 Professional Services 6,500 Travel 3,000 Office Rental 6,000 Utilities 3,600 Equipment Rental/Service 5,400 Other 3,300 TOTAL ADMINISTRATIVE EPENSE $119,500 TOTAL OPERATING EPENSE $648,540 TOTAL VEHICLES 14 TOTAL VEHICLE MILES 399,000 TOTAL VEHICLE HOURS 28,500 H:\Projects\Transit course\transit COST ANALYSIS.doc 9/5/ of 11

50 Because proper treatment of these cost differences is essential to a fair comparison of public versus private transit operations, the Federal Transit Administration (FTA) has sponsored several studies of the issues and published a number of reports and guides that explain how to properly determine these expenses. Fully Allocated Cost Analysis: Guidelines for Public Transit Providers, 3 a report prepared by Price Waterhouse, provides detailed information on how to construct fair and accurate cost comparisons of private and public transit services. The Price Waterhouse report describes how the three-variable unit cost model can be used to estimate total expenses and it also explains the adjustments that must be made to compensate for differences between the public and private sectors. For those persons interested in using a microcomputer model to estimate total as well as subservice costs, the UMTA-sponsored Public Private Transportation Network (PPTN) has prepared and distributed a cost allocation model 4 that can be used for either fixed-route or paratransit operations. The private-sector comparison principles proposed in the Price Waterhouse study are incorporated into the PPTN model. Because the cost determination and cost allocation methods described in the next section of this guide are consistent with those presented in these manuals, they are only summarized here and presented in the context of the needs of the paratransit performance evaluation process. The reader is referred to the other resources for more detailed information on cost allocation and determination of private versus public sector costs. 3 Fully Allocated Cost Analysis: Guidelines for Public Transit Providers, Prepared by Price Waterhouse for the Urban Mass Transportation Administration, April "Cost Allocation Model: A Microcomputer Software for Transit Service Costing," The Comsis Corporation, February H:\Projects\Transit course\transit COST ANALYSIS.doc 9/5/ of 11

51 Cost Allocation Models Nearly all performance evaluation studies require that total operating expenses such as those listed in the cost allocation table be allocated so that the cost of providing a particular type of service can be determined. The cost allocation process involves distributing each cost element among the service components. For example, to determine the cost of operating a particular vehicle or group of vehicles in a particular service sector, total operating costs must be apportioned among all vehicles and/or services. Individual vehicle or service cost estimates are useful not only to evaluate the operating and financial performance of the service component, but also, such disaggregation of expenses is necessary to evaluate the desirability of private versus public provision of service. The most common method of allocating operating expenses incurred by transit systems, and the one recommended here, is called the three-variable unit cost model. The unit cost model assigns actual operating costs experienced by a system to each sub-service (vehicle, route, service area, etc.) based on three service variables: vehicle hours, vehicle miles, and vehicles. The underlying assumption behind the allocation model is that the cost of operating a transit system is directly related to the number of vehicle hours of service provided, the number of miles traveled, and the number of vehicles required to provide the service. Therefore, the expense of providing service in a specific service sector can be determined by apportioning total expenses of the organization in proportion to the number of vehicle hours, miles, and vehicles required to provide the particular service. The model can be described as follows: Annual Total Expense = (Vehicle Hour-Related Expenses * Vehicle Hours) + (Vehicle Mile-Related Expenses * Vehicle Miles) + (Fixed Expenses/Vehicle * Vehicles) (1) This cost expression can be used to represent the entire paratransit operation for the entire year, or it can be used to calculate the operating expenses for a sub-service and/or for a shorter time period. The remainder of this section presents a simplified example that applies the unit cost model to the data presented in the Cost Allocation chart. A more detailed explanation of the unit cost model and various refinements can be found in the sources listed earlier in this chapter. Also, a recent report prepared for the Maryland Department of Transportation 5 provides a thorough explanation of how to apply the unit cost model to demand-responsive transportation services. The reader should consult this report for step-by-step instructions on applying the unit cost model to a variety of situations faced by demand- responsive operators. 5 Cost Analysis Methodology for Demand-Responsive Service, prepared for the Maryland Department of Transportation Mass Transit Administration by Comsis Corporation, October H:\Projects\Transit course\transit COST ANALYSIS.doc 9/5/ of 11