Teik Soon Looi Di Pan George Sun Land Transport Authority, Singapore

Transcription

1 INTEGRATED FARE STRUCTURE: ANALYSING THE IMPACT ON TRAVEL BEHAVIOUR USING SMART CARD DATA Teik Soon Looi Di Pan George Sun Land Transport Authority, Singapore 1. INTRODUCTION In 2010, Singapore implemented a new integrated distance based throughfare structure (i.e. Distance Fares) for the entire public transport system to replace the fixed transfer rebate which had been adopted for almost 20 years. Under Distance Fares, commuters pay according to the distance travelled in a whole journey on bus and rail (i.e. mass rapid transit or light rail transit), regardless of whether they travel direct using a single mode, or make transfers involving same or different modes. It effectively removes transfer penalty on fares that comes with each transfer within a single mode (e.g. busbus transfer) or between two different modes (e.g. bus-rail transfer). Distance Fares offer commuters with more choices and flexibility to decide on the best route to reach their destinations without incurring additional boarding charge at each transfer. A commuter who prefers to wait for a particular direct service can choose to do so. Another commuter who wants a faster journey can choose to hop onto the first bus that comes along and then make transfers along the way to reach the same place. In other words, these two commuters (one is a direct journey and the other a transfer journey) pay the same fare for the same journey distance, regardless of transfers. This study uses smart card data to examine the effect of Distance Fares on commuters travel behaviour. Instead of depending on surveys to ask commuters to report their travel itineraries over a long period of time, we mined the smart card data collected through the integrated ticketing system over a period of several years before and after the implementation of Distance Fares. The results showed that the new fare structure had indeed encouraged commuters to pay by distance and choose their travel routes that best suit their needs. The paper presents a brief introduction to the integrated fare structure and the smart card ticketing data collected in the PLANET, an enterprise data warehouse developed by the Land Transport Authority (LTA), Singapore. The methodology and findings are discussed and some observations that were of interest to policy practitioners were also presented. 1

2 2. INTEGRATED FARES AND SMART CARD DATA 2.1 Integrated Bus and Train Fares From the outset of the introduction of mass rapid transit (MRT) system in 1987, Singapore started to develop an integrated bus and MRT fare structure. The aim was to bring greater convenience to commuters, attract and quickly build up the MRT ridership to help sustain the operating viability of the MRT system. In 1991, leveraging on the stored value magnetic farecard technology then, Singapore implemented its first integrated fare structure for public bus and train. Under this integrated fare structure, a commuter using the same magnetic farecard was given a fixed amount of transfer rebate on fare when he or she made a transfer between bus services or between bus and train within the permissible time limit. This fixed amount of transfer rebate was set to defray the boarding charge that came with each transfer, thereby reducing the transfer penalty faced by the commuter. The transfer rebate was first set at 15 (Singapore) cents and was increased in stages to 25 cents by The time limit in between each transfer was also increased from 30 minutes to 45 minutes 1 in As technology evolved, the stored value magnetic farecard ticketing system was replaced with Contactless Smart Cards (CSC) ticketing system in This system was further enhanced in 2009 to be the Symphony for e-payment (SeP) system 2 that would support the implementation of Distance Fares in The vision of Distance Fares was conceived in the 2008 Land Transport Master Plan (LTMP)(LTA, 2008). Under the plan, the LTA adopted a more commuter-centric approach and focused on improving commuters experience on the entire journey, from the origin to the destination. To the commuters, the whole public transport system would be treated as a whole, not as separate bus or train modes. This means making the transfers as seamless as possible even-though a journey may need a few transfers to be made. One way to make the transfer more seamless is to remove the fare transfer penalty. This means that when commuters make transfers, there should be no penalty on fares. While the 25 cents transfer rebate had helped to reduce the cost of each transfer, a journey involving a few transfers was still more expensive than a direct journey. Therefore, a decision was made in 2008 to remove fare penalty on transfers and charge commuters according to the total distance travelled in one journey from origin to destination. The transition to Distance Fares was progressively done over a three year period (2008 to 2010) as part of the nnual fare adjustment exercise. In 2008, the transfer rebate was increased from 25 cents to 40 cents. In 2009, it was further increased to 50 cents. The final switch to Distance Fares was in July 2010 to replace the fixed transfer rebate which had been adopted for almost 20 years. Under Distance Fares, commuters are charged for the same fare 2

3 for the same distance travelled, regardless the number of transfers, as long as valid transfers are made. The staged increase of transfer rebate from 25 cents to 50 cents and the final switch to Distance Fares had effectively lower the fares of journeys involving transfers. This benefitted the commuters. Figure 1 shows an example of fare saving to commuters who made one transfer in a 6km journey. With such reduction in fares for transfer journeys brought about by the implementation of Distance Fares, the effects on commuters travel behaviour were examined in this paper. $1.80 $1.50 $0.25 $0.40 $0.50 $1.20 $0.90 $0.60 $1.64 $1.39 $1.24 $1.14 $1.08 $0.30 $0.00 No transfer rebate (1990) 25cent rebate ( ) 40cent rebate (2008) 50cent rebate (2009) Distance-through Fare ( ) Based on the fare level of 2014 Transfer Rebate Journey Fare (Bus 2km, Train 4km) Figure 1: An example of change fare for a 6-km journey ( ) 2.2 Smart Card Data and PLANET Embarking on the integrated ticketing system as early as in 1991 was strategic move that laid the groundwork for capturing temporal, spatial and pricing data to support transport planning and management as well as policy development. Recognising the benefits of having such data, when the Contactless Smart Card (CSC) was introduced in 2002, the ticketing system was designed to require commuters to tap-in and tap-out when getting on and off buses or passing through and the faregates at train stations. This way, the ticketing system was able to detect whether a boarding belongs to a new journey or a transfer of a journey, compute the distance travelled and deduct the correct fares automatically. Dovetailing CSC introduction, LTA began to develop its first simple data warehouse in 2003, focusing on report generation for management s use. By 2006, it was recognised that LTA needed a more comprehensive enterprisewide data warehouse that could combine ticketing and fare data, traffic data, vehicle (bus and taxi) data and geographic data to generate timely information for more robust analysis and policy formulation. In 2010, a new enterprise data warehouse, the Planning for Land transport NETwork (PLANET) was launched. 3

4 Today, smart card data has been widely used in LTA for planning, policy making and performance monitoring. The PLANET 3 provides a critical platform for analysts and planners to understand issues and constraints, monitor performance and evaluate impact so that they are able to determine trade-offs among competing demands, refining and introducing new mechanism to influence commuters behaviours. Transactions from the ticketing system already contain information required to compute the fare for a journey, such as card ID, location (bus stop or train station), fares deducted, card-type (whether it is a stored value card or a season pass), patron-type (child, student, senior citizen or adult), service-type (a trunk bus service, feeder bus service, and express bus service, etc), time of tap-in and tap-out, etc. In the PLANET, all transactions belonging to a trip (a pair of boarding and alighting) and a journey (a series of trips) are identified in sequence and marked using unique ID. To reduce the processing time in analysing billions of records, the frequently used information, such as in-vehicle time and transfer waiting time, is also designed to be pre-processed and stored in the PLANET. With such designed data structure, the PLANET is able to process a one-month s collection of data within one minute, facilitating prompt analysis of vast amount of data collected over period of time. In this study, PLANET data collected on all bus and train journeys taken over a three year period is examined. 3. DATA SELECTION METHODOLOGY On a daily basis, almost 2 million smart cards were used to take buses and trains. The cards made more than 4 million journeys a day or more than 6 million individual trips on buses and trains. Over a month, there were round 5 million unique cards used in taking buses and trains, generating almost 130 million journeys. At any day, some new commuters would start to use buses and trains and some existing commuters would not (e.g. they might switch to cars). Some cards were used by tourists and they were captured in system for a short period only. Some cards were faulty and replaced by new cards. Therefore, in order to examine commuters travel behaviours before and after the implementation of Distance Fares and minimise the influence from other factors, such as demographic changes and short-period users, we had to ensure a) a selected card represented one commuter; b) selected cards can represent general public transport commuters; and c) the cards had been used before and after the implementation over a period of several years. The selection of cards was thus conducted in two steps. The first step was to create a population of cards which had been used over three years from 2010 to In this study, adult and senior citizen cards that were used to pay bus and train fares in all three periods: 2 weeks in 2010, 4

5 same period in 2011 and 2012, were selected to create a population of cards. Almost 1 million cards were identified in this step. The second step was to use stratified random sampling method to randomly pick up 10% of identified cards. 92,450 cards were selected in this step, containing both adult and senior citizen cards, and were proportionally to the population. Stratified random sampling ensured characteristics in the sample that were proportional to the overall population (Gravetter & Forzano, 2012). The identified cards were grouped into two groups based on their holders characteristics: adults and senior citizens. Each group was considered one stratum. Samples were selected randomly in each group according to their share of total identified cards. The random sampling method ensured results were representative. Having selected these 92,450 cards, all transaction records between September 2009 and June 2012 on bus and train made by such cards were extracted for the longitudinal study. The analysis was conducted in two perspectives. First, the aggregated effect on all selected commuters (i.e. cards) was reviewed. The purpose was to probe whether the overall travel patterns on public transport had been changed after the fare structure change. Second, a comparison of behaviours of individual commuters before and after the implementation of Distance Fares was conducted to understand at the individual level how commuters responded to fare structure change. The number of journeys taken by a same card was normalised to a scale of 100 to remove the fluctuation in travel frequencies. 4. EFFECTS OF DISTANCE FARES Distance Fares has totally removed transfer penalty on fares. By removing the travel cost due to transfer penalty, it is expected that commuters will have less resistance to make transfers and thus there will be more journeys made that involve transfers. In addition to savings on travel cost, shorter waiting time at transfers can also be expected on bus-bus transfers. This is because such bus commuters will now have more choices of bus service to hop on and make transfers down the road instead of waiting for a particular direct bus service to arrive. Such commuters prefer to be on the move rather than waiting. As the ticketing system only starts to record a journey from the first tapping of the smart card on bus or rail, the data on first waiting time (i.e. prior to the first boarding) is not captured for analysis. Nonetheless, the transfer times that are captured can provide a good data source to examine the changes on transfer times before and after Distance Fares. For clarity, in this analysis, transfer time is the time interval between the time of tap-out for alighting a bus or exiting the rail system in the previous ride and the time of tap-in for boarding a bus or entering the rail system in the following 5

6 ride. It includes the walking time in between bus stop and train station. With this definition, transfer time is computed for both bus-bus transfers and busrail transfers. It was expected that transfer time would become shorter under Distance Fares. These two effects of Distance Fares, i.e. on transfer journeys and on transfer time, are discussed in the ensuing paragraphs. 4.1 Increased Transfer Journeys A journey is defined to be a series of trips made on bus and train from the origin to the destination. Journeys can be categorised into direct journeys (i.e. a single bus or train trip), bus-bus transfer journeys (all transfers are between buses) and bus-rail transfer journeys (there is at least one transfer between bus and train). In the year after Distance Fares implementation, the number of transfer journeys in the public transport system grew much faster than the overall PT journey growth rate. A question was thus raised on whether the faster growth on transfer journeys was directly related to the implementation of Distance Fares or other factors. To answer this question, two time series were plotted for the selected cards: the percentage of bus-bus transfer journeys over total journeys (see Figure 2a) and the performance of bus-rail transfer journeys over total journeys (see Figure 2b). a) % bus-bus transfer journeys b) % bus-rail transfer journeys % Bus-bus transfer journeys (selected cards) Percentage point change due to CCL Percentage point changes not due to CCL % bus-rail transfer journeys Pre-DF 10% Q4 2010Q1 2010Q2 2010Q3 (CCL1&2) (DF) 2010Q4 2011Q1 2011Q2 2011Q3 2011Q4 2012Q1 (CCL4&5) Figure 2: Percentage of bus-bus and bus-rail transfer journeys 2012Q Q4 2010Q1 2010Q2 (CCL1&2) 2010Q3 (DF) 2010Q4 2011Q1 2011Q2 2011Q3 2011Q4 (CCL4&5) 2012Q1 2012Q2-1.2 Figure 2a showed that, for the selected cards, the percentage of bus-bus transfer journeys increased by 8% immediately after the implementation of Distance Fares. This result confirmed that the increase of bus-bus transfer journeys was directly related to the implementation of Distance Fares. This finding is consistent with literatures on the relationship between transit demand and transit fares which showed that fare reduction can increase the demand on public transport (Litman 2014, Holmgren 2007, Turnbull and Pratt 2003). Furthermore, it suggested that transit fare is one of factors affecting commuters decision on their route choices. A direct journey is attractive because of not only its convenience (i.e. less hassle) but also it entails only one boarding charge (hence lower fare) as compared to transfer journey. If 6

7 No of Transfers Cumulative % of transfers there is no fare difference, commuters may consider making transfer journeys for other benefits, such as shorter waiting time at bus stops. Commuters do not have to stick to waiting for a direct service just to benefit from lower fare. For example, they may hop on the first bus that comes along and transfer to another bus further down the road to reach their destinations faster. The increase suggested that more commuters had discovered the benefit of transfers that best suit their needs. When the new fare structure was implemented in 2010, some adjusted their travel behaviours immediately. Others took some time to get familiar with new rules and to appreciate the advantages of new fare structure. Different from bus-bus transfer journeys, the finding on bus-rail transfer journeys was less conclusive. While the percentage of bus-rail transfer journeys increased over the years (see Figure 2b), the net increase was more likely to come from the opening of Circle Line (CCL), a new MRT line starting revenue service around the same period. The new MRT line also generated demand on bus-rail transfer journeys and mode shift from bus to rail. Although some commuters probably made bus-rail transfers involving CCL because there was no additional cost on transfer journeys, we were unable to conclude which one was the driving factor, i.e. the new CCL or Distance Fares, by mining the ticketing data alone. 4.2 Shorter Transfer Time The total number of transfers in public transport system had increased significantly. More transfers were made within 10 minutes after the Distance Fares (See Figure 3a). In terms of the distribution of transfer time (the maximum allowed for a valid transfer is 45 minutes), in total, around 95% of the transfers were able to be completed within 30 minutes (See Figure 3b). As expected, the same 95% percentage of bus-train transfers was able to be completed within 20 minutes compared to bus-bus transfers of 35 minutes. This suggested that there scope to fine-tune the limit of transfer time so as to minimise commuters taking advantage of transfers to run quick errands by breaking up a journey into trips. a) Number of transfers by transfer time b) Cumulative distributions of transfer time 100% Pre DF (2010) Post DF (2013) 80% 60% 40% Bus->Bus Bus->Rail Rail->Bus Total Transfer time (minutes) Figure 3: Transfers distributions by transfer time Transfer time (minutes) 0% A comparison between bus-bus transfer journeys and direct bus journeys made by the same group of commuters for the same pair of origin and 20% 7

8 destination bus stops (O-D) showed two effects: (a) average transfer time had dropped under Distance Fares (see Table 1); and (b) average travelling cost (fare paid by commuters) was the same between direct and transfer journeys for same O-Ds under Distance Fares. The shorter transfer time (20.6% reduction in transfer time) after Distance Fares implementation confirmed that commuters spent significant less time in waiting with no difference in fares between transfer and direct journey. Table 1: Average fare and travelling time for journeys with the same O-D Average Fare ($) Before After Direct journey $0.93 $0.94 Transfer journey $1.07 $0.94 Difference between direct and transfer journey $0.14 $0 Average Transfer Time (minutes) (20.6% ) Notes [1] Bus stop pairs were not same in the two periods (before and after DF) [2] Bus fare was raised in addition to the DF implementation. Before DF, adult fare on 1 st fare band was 69 cents, while it was raised to 73 cents after DF012. Interestingly, in terms of distance, almost a quarter (24%) of such commuters were even willing to the make a transfer journey that were slightly longer (by within 1km) in exchange for shorter waiting, if the journey fare did not cost them more (See Figure 4). Figure 4: Proportion of direct and transfer journey by distances 5. INFLUENCES ON COMMUTERS SOME OBSERVATIONS This section discusses some observations on commuters travel behaviour changes that were associated to Distance Fares. 5.1 Commuters Adjusting Travel Behaviours for Distance Fares The core concept of Distance Fares is to provide commuters an option for making transfer journeys without having to pay higher fares. While commuters enjoy lower fares for transfer journeys under Distance Fares, not 8

9 % Commuters making transfers where direct services available % commuters making transfers where direct services available all transfer journeys have resulted from travel behaviour changes in response to Distance Fares. This is as commuters may deliberately make transfer journeys (where direct services are available) in order to run quick errands within the transfer time. Such behaviour may already exist even before Distance Fares. In order to understand the influence of Distance Fares on commuters travel behaviour using PLANET data, a comparison can be made using the percentage of the same commuters (i.e. represented by each unique smart card) that make transfer journey where direct journey is available, before and after Distance Fares. Assuming that the propensity to make quick errands by each smart card remains unchanged, any changes in the percentage of transfer journeys according to such unique smart cards before and after Distance Fares will reflect the adjustment made in response to Distance Fares. PLANET also allows unique smart card data to be mined to track such snap shot changes over a time period (e.g. on a quarterly basis) to reveal the adjustment trend in the initial period of Distance Fares implementation. Figure 5a showed the percentage of such transfer journeys had increased significantly after Distance Fares and it was also trending up. As at end of second quarter of 2012 (i.e. 2012Q2), the percentage was more than 2% point higher than that before Distance Fares. This result suggested that more than 2% bus commuters had adjusted their travel behaviours. The up-trend between 2010Q3 (after Distance Fares taking effect) and 2012Q2, i.e. the first 2 years, indicated that more commuters had discovered the benefits of transfer journeys that were better suit their travel needs. It also confirmed that the removal of transfer penalty on fares had effectively influenced commuters preference between direct and transfer journeys and resulted in travel behaviour changes on taking buses. a) Overall percentage b) Adult and senior citizen Before DF 2010Q3 2010Q4 2011Q1 2011Q2 2011Q3 2011Q4 2012Q1 2012Q2 Figure 5: Commuters making transfer journeys where direct services available Figure 5b showed the trends of smart cards breakdown by passenger types, i.e. adults and senior citizens 4. Notably, higher percentage of senior citizens had increased their transfers to enjoy faster journeys, as compared to adult commuters. In the same period, the senior citizens also showed a faster uptrend than adult commuters. Such uptrend showed increasing adoption 0 % Adults % SCs Before DF 2010Q3 2010Q4 2011Q1 2011Q2 2011Q3 2011Q4 2012Q1 2012Q2 9

10 (i.e. acceptance and action) of both adults and senior citizens towards making transfer journeys. Higher cost reduction and more non-commuting trips were the two likely factors leading to higher adoption in senior citizens. First, the perceived cost saving on transfer journeys was higher for senior citizens than adults. This was because senior citizen concessionary fare was capped at a distance of 7.2km for the whole journey, while adult fare continued to increase for every 1km until 42km. Second, compared to adults, senior citizens tended to make more non-commuting or discretionary journeys which were more sensitive to fare changes (McCollon and Pratt 2004). 5.2 Improvement on Smartcard Tapping Behaviour In addition to travel behaviour changes, improved smartcard tapping behaviours 5 on buses were also observed under Distance Fares. The drop of improper tapping, i.e. those who tapped out quickly after boarding and those who failed to tap out upon alighting (see Figure 6), suggested that commuters were more conscious of the negative consequence of improper tapping that they would not be able to enjoy the benefit at transfer journey fares under Distance Fares. The increased awareness could be a result of public education to remind commuters to tap-out properly that was undertaken at the launch of Distance Fares. %bus rides where commuters tapped for alighting immeidately after boarding % bus rides where commuters did not tap for alighting 2.50% 1.33% 1.20% 1.22% Before DF After DF Before DF After DF Figure 6: Improvement on tap-out behaviour From PLANET data capturing point of view, this improved tap-out behaviour meant that there were more data with complete information on both boarding and alighting that would enable more accurate data analysis and data analytics be done for bus service performance monitoring. For example, the data analytics has recently enabled the production of island-wide heat-map that tracks bus passenger loadings (or crowdedness) which is computed using the number of passengers getting on and off a bus at each bus stop along a bus trip. Another use is the tracking of actual in-vehicle bus time and distances (see Figure 7) which can be used for determining that spread of time variation by distance, estimating journey time for dissemination information to commuters, and estimating speed for bus capacity planning. 10

11 Time (minutes) Time (minutes) Single-Ride Travel Time Tranfer-Journey In-Vehicle Time y = 2.79x R² = y = 2.80x R² = Distance (km) Distance (km) Figure 7: Relations between in-vehicle time and distance 5.3 Commuters Took Advantage of Distance Fares Literature suggests that given an innovation, individuals may create some new ways of using it, which is known as re-invention (Rogers 2003). Similarly, for a policy change, individuals may discover some ways to maximise its utilities. The study found that there was a small group of commuters who took advantage of Distance Fares in a manner not designed for. More commuters took different bus services within 45 minutes of transfer time in order to make return trips back to the same origins, so as to trick the system into treating the to- and fro- trips as a continuous single journey instead of two separate journeys, thereby saving on fares. Such fare saving was higher as compared to the 50 cents transfer rebate before Distance Fares. In particular, if the combined travel distance was less than 3.2km, which was still within the first distance fare band, there would be no additional fare deducted for the return trip. Commuters might find that taking a bus would be a more attractive alternative than a short walk to run a quick errand down the road. As a result, a significant increase was observed on bus journeys which consisted of two bus trips (i.e. to- and fro-) with the same origin and destination pair as a single journey.(see Figure 8). From fare policy perspective, such return trips would contribute to fare leakage as only one direct journey fare was charged instead of two direct journeys. % bus journeys with return trips 0.16% 0.21% Before DF After DF Figure 8: Bus Journeys with return trips 11

12 6. SUMMARY The enterprise data warehouse, the Planning for Land transport NETwork (PLANET) captures all smart card travel and payment transactions made on buses and trains. It is part of the big data that is developed to allow practitioners and researchers to carry out data analysis and data analytics for policy making, planning, performance monitoring and understanding travel patterns and behaviour. Instead of relying on a longitudinal interview survey, this study mined the PLANET data over three years to examine the effects of Distance Fares and infer commuters behaviour changes. The results showed that the total number of transfers had increased significantly (8%) and transfer time had decreased (by 20.6%) after the implementation of Distance Fares. More transfers were made within 10 minutes after the Distance Fares, with around 95% of the transfers completed within 30 minutes. The percentage of bus-bus transfer journeys increased by 8% immediately after the implementation of Distance Fares. Comparing busbus transfer journeys and direct bus journeys, average transfer time had dropped and average travelling cost (fare paid by commuters) remained unchanged under Distance Fares. Some commuters were prepared to take a slightly longer transfer journey. Observations showed that more than 2% of commuters had adjusted their travel behaviours, suggesting that Distance Fares had effectively influenced commuters preference between direct and transfer journeys and resulted in behaviours changes when taking buses. Commuters were willing to hop on the first coming bus and transfer to another bus further down the road, to reach their destinations. This meant that there is scope to restructure a long bus routes into shorter legs so as to improve bus service reliability. The increasing adoption trend at initial period showed that such major policy change required some time to take effect and its impact was related to the extent that commuters had actualised the benefits of Distance Fares. Public education had improved the awareness of the policy change. Tapping behaviour was improved. There was unintended behaviour observed as commuters sought to maximise their utilities. Although the behaviour changes were inferred from travel patterns obtained from PLANET data, the actual reason for a transfer is not known. For example, this data analysis will not reveal whether a commuter make a busrail transfer because of Distance Fares or rail network expansion (a new MRT line). Nevertheless, the findings suggested that commuters would adjust their travel behaviours when there is a change in fare policy. Further studies are therefore needed to examine commuter s travel behaviour change in response to new MRT lines and/or new bus routes. REFERENCES Gravetter, F. J. & Forzano, L. B. (2012). Research Methods for the Behavioural Sciences (4th Ed.). Wadsworth Cengage Learning 12

13 Holmgren, J. (2007), Meta-analysis of Public Transit Demand, Transportation Research Part A: Policy and Practice, Volume 41, Issue 10, Dec. 2007, pp Land Transport Authority (2008), Land Transport Master Plan 2008, Singapore Litman, T. (2014) Transit Price Elasticties nd Cross-Ealasticisties, Journal of Public Transportation, 7 (2), pp.37-58; ( McCollon and Pratt (2004), Traveler Response to Transportation System Changes: Chapter 12, Transit Pricing and Fares, Transit Cooperative Research Program Report 95, Transport Research Board ( Public Transport Council & Land Transport Authority (2010), Expiry of 3% Fare Rebate & Launch of Distance Fares on 3 July 2010, Singapore Rogers, E. M. (2003). Diffusion of Innovation, 4 th ed. Simon & Shuster Inc. Turnbull, K. F. and Pratt, R. H. (2003), Transit Information and Promotion: Traveler Response to Transport System Changes, Chapter 11, Transit Cooperative Research Program Report 95, Transportation Research Board ( NOTES 1 In addition, a transfer was also governed by a time limit of 2 hours measured from the first tap-in of the card at the start of each journey to the current tap-in of the card for the transfer. 2 SeP adopts the Specification for Contactless e-purse Application (CEPAS), a national standard SS 518 gazetted in January CEPAS is a nationwide interoperable micropayment platform that bridges multi-industry sectors, both the transit and retail e-payment space. It is a result of an interagency collaboration, with the LTA as one of the key partners. All smart cards used in the integrated ticketing system have to comply with CEPAS. 3 Data coverage is >13 million daily; data accuracy >95%; data analytics generates >27,000 dashboard charts and >150 management reports monthly. 4 Typically, adults constitute some 80% of the total daily journeys while Senior Citizens (i.e. concession card holders whose ages are 60 and above), constitute some 10%. The remainder are students, children, etc. 5 Singapore s bus and train ticketing system is a closed-system whereby commuters have to tap-in (i.e. entry process) and tap-out (exit process) their smartcards upon boarding and alighting buses or passing through fare gates at train stations. Unlike trains, smart card tapping at entry and exit processors on buses can be subject to misuse. Some commuters tap out quickly after boarding so as to underpay the fares. Those paying flat fares may be lazy to tap out as it makes no difference to their flat fares paid by them. 13