NATIONAL ROADS AUTHORITY

Project Appraisal Guidelines

Appendix 7 - COBA guidance

March 2008

Version 1.0

COBA guidance Changes made to document since previous version -

Date issued 01.03.2008

St Martins House, Waterloo Road, Dublin 4 Tel:+353 1 660 2511 Fax +353 1 668 0009 Email: info@nra.ie Web: www.nra.ie

Table of Contents

1 2

Introduction ..................................................................................................................... 3 Data requirements........................................................................................................... 5 General.............................................................................................................................. 5 Data categories ................................................................................................................. 5 Scheme costs.................................................................................................................... 6 Link and junction data ....................................................................................................... 7 Traffic data ........................................................................................................................ 8 Traffic growth..................................................................................................................... 8 Accident data .................................................................................................................... 9 Economic values ............................................................................................................... 9 Local or national data........................................................................................................ 9 Modelling........................................................................................................................ 13 General............................................................................................................................ 13 Extent of network ............................................................................................................ 13 Role of the conventional traffic model............................................................................. 13 Data for the traffic model................................................................................................. 13 Traffic model validation ................................................................................................... 13 Compatibility between the traffic model and COBA........................................................ 14 Compatibility between the assignment and COBA networks.......................................... 14 Defining the fixed trip matrix............................................................................................ 15 When to model junctions................................................................................................. 15 Journey time validation techniques................................................................................. 17 Parameters..................................................................................................................... 21 General............................................................................................................................ 21 Economic parameters ..................................................................................................... 21 Future forecasting of economic parameters ................................................................... 23 Route selection ............................................................................................................... 23 Preliminary design........................................................................................................... 28 Construction documents preparation and tender award................................................. 30 Final account / closeout .................................................................................................. 30 Summary ......................................................................................................................... 31 How to use and run the COBA program ..................................................................... 37 General............................................................................................................................ 37 Structure of COBA data file............................................................................................. 37 Detailed description of COBA data ................................................................................. 37 Format of data entry........................................................................................................ 38 Data preparation and editing........................................................................................... 40 COBA Output and Interpretation ................................................................................. 43 General............................................................................................................................ 43 Description of COBA output ............................................................................................ 43 Interpretation of COBA output......................................................................................... 44 Sensitivity testing ............................................................................................................ 44 CBA report ..................................................................................................................... 51 General............................................................................................................................ 51 When is a CBA report required? ..................................................................................... 51 Report contents............................................................................................................... 51 Auditing and checking.................................................................................................. 55 General............................................................................................................................ 55 Responsibilities ............................................................................................................... 55

Key audit areas ............................................................................................................... 55 Audit checklist ................................................................................................................. 57 9 Final account / closeout ............................................................................................... 59 General............................................................................................................................ 59 Responsibilities ............................................................................................................... 59 Timing 59 Data requirements........................................................................................................... 59 COBA assessment.......................................................................................................... 60 Comparison with construction documents preparation / tender award COBA ............... 60 References ..................................................................................................................... 63 Enquiries ........................................................................................................................ 65

10 11

Introduction

COBA guidance

1
1.1

Introduction
This document is an Appendix to the NRAs Project Appraisal Guidelines and provides information on the processes required to undertake a COBA assessment. It deals with data collection, the requirements of the traffic model, and the parameter values that should be used for assessments at the different phases of a road scheme. The Appendix also addresses the structure of the COBA input and how the data file may be prepared and edited for input to the COBA programme. After completion of a COBA assessment, it is necessary to submit a full report to the National Roads Authority (NRA), the contents of which are also outlined in this section. The steps to be undertaken in completing an audit of the COBA print out and the requirements for Final Account/Closeout are also explained. An Irish specific version of the COBA software has been developed by Transport Research Laboratory (TRL), who is responsible for developing and maintaining the UK version of the program. A report prepared by the TRL documenting the development of the Irish COBA can be found in Appendix 8, which is available to download from the NRAs website, along with other supporting documents. The Appendix is structured as follows: Section 2 Data requirements. This describes the data collection that is necessary to prepare a COBA input file for undertaking a CBA study; Section 3 Modelling. The key requirements of the traffic forecasting model and its interface with the COBA input file is discussed; Section 4 Parameters. The recommended approach for input of parameters into the basic section of the COBA input deck is set out. The approach is dependent upon the phase at which the scheme assessment is being undertaken, with national default values sufficient at Route Selection and local scheme specific values required at Preliminary Design, Construction Documents, and Final Account / Closeout. While national default values are generally acceptable, the use of local values must be supported by an explanation of their derivation; Section 5 How to use and run the COBA program. The structure of the COBA input file, and the type of information required by it is described; Section 6 COBA output and interpretation. The interpretation of the COBA output is critical for an understanding of scheme evaluation and also helps identify errors in the specification of inputs. Section 6 describes the outputs from COBA, what they mean and how they can be used to identify the key economic indicators. Additional inputs required for the economic assessment of a scheme that are not obtained from COBA, for example environmental benefits and assessments of benefits during construction and maintenance, are also discussed; Section 7 CBA appraisal report. This Section describes what should be included within the CBA Appraisal Report, when it is required and who is responsible for producing it; Section 8 Auditing and checking. Guidance is provided on what to check when validating a COBA printout and CBA Appraisal Report, when this should be undertaken, and by whom; Section 9 Final account / closeout. The methodology for undertaking a post-project review, along with a rationale for this process is presented; Section 10 References, and Section 11 Enquiries. A contact address for technical queries and comments is provided.

1.2 1.3

1.4

1.5

Data requirements

COBA guidance

Data requirements

General 2.1 COBA has the flexibility for a number of input parameters to be modified by the user to better reflect local conditions. However, the use of standard values and relationships is central to the COBA concept and, at the early stages of scheme assessment, this avoids the need for time consuming and costly data collection exercises. The decision to use local or national default values depends on the phase of the scheme assessment. At Route / Option Selection, local values may be used where available, although national values are normally acceptable. However at Preliminary Design, local data should be input where it is both reliable and significantly different from national COBA values. Details of how local data has been compiled for use in the CBA process should be clearly outlined in the CBA Appraisal Report. At Final Account/Closeout, the same parameters used at Preliminary Design should be adopted. This Section sets out the different categories of data required, how such data should be compiled and whether local or national data are more appropriate for the CBA process. Data categories 2.4 Data input into COBA can be classified into three categories, according to the source, as illustrated in Table 2.1: (i) (ii) (iii) 2.5 Data that are always local; Data that should be local if values are reliable and differ significantly from national values, and Data that should always be national.

2.2

2.3

The data input can, in essence, be grouped under five key headings as follows: Scheme costs; Link and junction data (geometry); Traffic data; Accident data, and Economic values. The compilation of the data within each of these headings is considered in turn below.

COBA guidance

Table 2.1 Summary of data required for COBA Always Local Scheme costs Link geometric characteristics Junction geometric characteristics Link flows Junction turning proportions Local or National Seasonality index E-Factor M-Factor Traffic mix proportions Flow groups Accident rates and casualty proportions Speed-flow curves Vehicle occupancy Always National Values of time Accident costs Economic growth Vehicle operating cost parameters Taxation Traffic growth profiles Discount Rate Consumer Price Index Relative Price Factor Carbon Costs

Scheme costs Components of scheme costs The total scheme cost is made up from the following elements: Construction costs; Land and property costs; Preparation costs (planning and design); Supervision costs, and Maintenance costs. 2.6 Scheme costs can be entered into COBA using two different input keys KEY054 user inputs Construction and Land Cost estimates and COBA calculates the equivalent costs in the present value year for construction, land, preparation and supervision and allocates them to the correct year, or KEY055 allows the user to enter total scheme costs according to a manually calculated profile of expenditure. 2.7 2.8 Both keys allow the user to specify the sector incurring the cost (i.e. Central or Local Government) and any contributions from private developers. In each case, costs are entered into COBA exclusive of indirect taxation, i.e. in the factor cost unit of account. COBA uses indirect taxation rates held within the program to convert factor costs into market prices. Costs are entered into COBA as undiscounted values. COBA uses information on the discount rate to derive discounted values. When using KEY055, costs are entered at the price base year value, i.e. after taking into account the effects of inflation. KEY054 converts prices to the price base year using information

2.5

2.9 2.10

COBA guidance

provided by the user on the Consumer Price Index at the base year and the year in which the estimate was made. 2.11 It is an NRA requirement that KEY055 is used to enter costs. Detailed guidance on how to derive the profile of scheme costs for input into COBA KEY055 can be found in Appendix 12. Link and junction data 2.12 The geometric characteristics for the links and junctions included in the COBA network should be obtained from suitably scaled plans of the Do-Minimum and Do-Something schemes, and supported by field measurements of those elements on the existing network that are to be included in the assessment. Link data The following data relating to links are required: Type of link to allocate the speed-flow relationship, with links defined in accordance with the classifications in Table 4.2 (see Section 4); Speed limit; Length; Width including widths of any hard shoulders, hard strips and verges; Hilliness; Bendiness; Number of major junctions or accesses per kilometre; Degree of development fronting the link, and Sight distances. 2.14 The list above is an outline of the information necessary. The complete set of information necessary for each link depends on the link type (e.g. rural, urban, suburban, dual or single carriageway etc.). Full details of the data requirements, including definitions and their derivation, are given in the UK DMRB Volume 13, Section 1, Chapter 5. Speed limits Since January 2005, speed limits displayed in imperial values have no legal basis under EU Regulations. All speed limit signs in the Republic of Ireland are now required to display metric values. Default metric speed limits are outlined below: a) Motorways: 120 km/h; b) National roads: 100 km/h; c) Non-national roads: 80 km/h, and d) Roads in urban areas (builtup area speed limit):50 km/h. 2.16 In the event that it is intended to apply a 120kph limit on a Type 1 Dual Carriageway, as provided for under Irish law, two separate scenarios should be studied: with a 120 kph speed limit applied, and without. The most likely speed limit to be applied shall be chosen as the baseline scenario. Special speed limits of 100 km/h, 80 km/h and 60 km/h are available for application by the local authorities as deemed appropriate. In addition, a special speed limit of 30 km/h may be applied at areas of pedestrian/cyclist concentrations where appropriate speed reduction measures are employed. Further information can be found in the Report of the Working Group on the Review of Speed Limits (2003), commissioned by the Department of Transport. Maximum link speeds are entered into the COBA program in kilometres per hour. Junction data 2.20 The requirement to model junctions in a COBA network will arise if the presence of junctions on the existing road significantly impacts on journey time through the network. An accurate replication of existing journey times through the network can therefore be difficult if this relies on

2.13

2.15

2.17 2.18

2.19

COBA guidance

link transit times alone. In such cases, the inclusion of junctions will allow a more accurate reflection of traffic behaviour. 2.21 For junctions that are to be modelled explicitly, the level of detail to be collected depends on the type of junction; i.e. signal controlled, roundabout, signalised roundabout, merge or priority junction that is being modelled. An indication of the data requirements for each junction type is provided below: For all junction types, the junction will need to be defined as either rural or urban and whether the entry arms are single or dual-carriageway; For priority junctions, the major and minor arms and any stagger should be identified. The width of the carriageway for the various turning streams should be measured along with the geometry of any central reserves and sight distances; For roundabouts, the user must specify if the junction is grade separated or not, and specify the geometry of the entry arms and of the circulatory carriageway; For signal-controlled junctions, including signal-controlled roundabouts, the user will need data on the staging arrangement at the junction along with the geometry and gradient of the approach arms. Signal- controlled roundabouts also require data on the saturation flows for each approach arm, and the user can model the effects of flared entry arms, and For merge junctions at grade-separated interchanges, the number of lanes downstream from the merge is required. 2.22 Full details of the data requirements for junctions are given in DMRB Volume 13, Section 1, Chapter 6. Traffic data 2.23 Traffic flow data, in the form of link flows and turning proportions are normally obtained from a traffic model, the complexity and extent of which will reflect the nature of the scheme. In most instances the traffic reassignment effects of a highway scheme cannot be predicted accurately without a conventional traffic model, developed using commercially available software. For the purpose of developing and validating the traffic model, the following items of data will be required: Origin and Destination surveys; Traffic count data, and Journey time surveys. 2.25 Journey time surveys allow validation of the base year COBA model by comparing modelled journey times with actual recordings. They may also provide data for locally adjusting the maximum delay parameter at junctions. For the traffic count data, long-term data (i.e. over a period of at least one-year) are required to establish scheme specific values for the following parameters: Seasonality index; E-factor; M-Factor, and Vehicle mix proportions. 2.27 Default values for these parameters, applicable to the assessment of Irish road schemes, and the derivation of scheme specific parameter values for the above parameters are discussed in more detail in Section 4. Traffic growth 2.28 For traffic growth, national default growth rates for both low and high traffic growth are to be applied; these growth rates are presented in Section 4.

2.24

2.26

COBA guidance

Accident data 2.29 Local data on the occurrence and severity of accidents should relate to a period when the conditions on the road have been broadly unchanged (for example, no abnormal changes in traffic flow, no changes in junction design or road geometry, etc.). Local data should ideally cover the five years previous to the COBA assessment but in all cases must cover a period of at least three years. In addition, accident data should be checked to identify any particular anomalies where it is suspected that they do not accurately reflect the accident history of the link. Where such anomalies do exist, accident data may need to be processed or adjusted before inputting to the COBA data file. One such example is where road works led to a significant change in the accident rate over a period of one year during a five year period. In such a case it may be appropriate to manually interpolate data for that year. The number of accidents in each year is input, including zero for those links or years where no accidents occurred, and COBA will then internally produce a local accident rate (accidents per million vehicle kilometres) for each link. If insufficient data are available to allow computation of local accident rates, then the default national default values should be used. Only those accidents recorded by the Garda will be manifest in accident data, however the Garda are not always in attendance at the scene of accidents, especially when no injuries occur. The COBA program contains factors that automatically adjust locally observed accident data to account for under-reporting. Economic values 2.34 The COBA user must make use of the most recent data on economic values relating to: Discount rate; Value of time; Indirect rates of taxation; Value of accidents; Vehicle operating costs, and Carbon costs. 2.35 National default values are applicable for all economic parameters regardless of the scheme assessment phase. These values are contained within the COBA program and must not be changed unless instructed to do so by the NRA. Local or national data 2.36 With the exception of those parameters relating to economic values, the decision whether to use local or national (default) values will depend on the phase of the assessment and the nature of the parameter. At the route / option selection phase, when different routes are being considered, it will be sufficient to use the national default parameter values to avoid the time consuming and potentially costly exercise of collecting local data. However, when a detailed assessment is being undertaken (at the Preliminary Design phase), then parameters based on local data should be used for the parameters relating to those items in column two of Table 2.1. The parameters in column three of Table 2.1 should remain at the national default values and will be provided within the default COBA input files. Periodically these values may change and to account for this: An updated version of the COBA program will be disseminated. COBA users should ensure that on commencing CBA that the most recent version of the software is being used; or In advance of any revisions of the COBA software being made available, the NRA will issue instructions to amend national values manually.

2.30

2.31

2.32 2.33

2.37

2.38

COBA guidance

10

2.39

One of the principal purposes of COBA is to provide a standard economic appraisal and a benchmark against which, if necessary, the sensitivity of returns to local variations can be measured. For this reason, a sensitivity test using national values as a benchmark may also need to be undertaken to ascertain the importance of the local variations and to allow comparison of schemes on a similar basis. The source, derivation and treatment of local and national parameter values are outlined in more detail in Section 4.

2.40

COBA guidance

11

Modelling

COBA guidance

13

Modelling

General 3.1 The COBA user is required to provide a description of the road network to cover both the existing situation and all the proposed improvements. The basis of this will be a conventional traffic model, from which link flows for input into the COBA assessment will be obtained. In this Section the structure of the COBA model and its interface with a conventional traffic model is described. For more specific advice regarding traffic modelling, Appendix 3 should be consulted. Extent of network 3.2 The COBA network should extend far enough from the improvement to include all links on which there is a substantial difference in the assigned traffic flows between the Do-Minimum and Do-Something networks. If the scheme is expected to result in a significant change in the flow level on a competing route, that route should be included in the network. This concept should be balanced by the consideration that, as the network spreads, benefits arising from the scheme in distant areas are inherently less plausible and more difficult to assess than local benefits. Generally, the extent of the assignment network will determine the size of the COBA network. Role of the conventional traffic model 3.3 3.4 One of the principal purposes of the conventional traffic model is to provide Do-Minimum and Do-Something traffic flows for input into the COBA model. It is desirable for all traffic modelling to be kept as simple as possible. For some simple schemes, such as the replacement of a rural road, manual assignment methods will suffice. For more complicated schemes a traffic model using commercially available assignment software, such as SATURN, TRIPS and EMME2 should be used for this purpose. It is usual for traffic models to cover both the morning and evening peak hours and an interpeak hour. It is also common for 12-hour or Annual Average Daily Traffic (AADT) models to be developed, especially when road conditions indicate a relatively flat (non-peak) traffic profile. When AADT models are not used, local traffic data will be required to enable the development of factors to synthesise a 12-hour, 16-hour or AADT traffic flow that can be input into the COBA model. Data for the traffic model 3.6 The following data are required in order to build and validate the traffic model: Geometric characteristics of links; Geometrical and operational details (in the case of traffic signals) of junctions in the modelled area if these are to be modelled explicitly; Origin and destination data; Journey time surveys, and Traffic counts. Traffic model validation 3.7 The traffic model should be fully validated for the base scenario, thus demonstrating that it forms a reliable platform for deriving both the Do-Minimum and Do-Something traffic flows.

3.5

COBA guidance

14

This validation will cover both count validation; demonstrating that the assigned flows accurately replicate existing conditions; and journey time surveys, confirming that modelled times on significant routes match those recorded on the ground. Guidance on the validation of assignment models is provided in the UK Design Manual for Road and Bridges, Volume 12: Traffic Appraisal of Road Schemes. The traffic model should be accompanied by a Traffic Forecasting Report, which describes the development, calibration and validation of the traffic model. 3.8 The Design Office Project Manager approves the validation of the traffic model prior to the cost benefit analysis being undertaken. Compatibility between the traffic model and COBA Fixed trip matrix It is important to ensure that the basis for estimating changes in traffic patterns is compatible with the economic assessment method used. COBA computes benefits on the assumption that the travel demand (the trip matrix) does not change as a result of the introduction of the scheme into the network. This assumption allows the simplification that only changes in resource costs need to be considered in the benefit calculation. Prediction of travel behaviour, on the other hand, will usually be based on perceived costs (also called behavioural costs). The distinction between resource and perceived cost parameters has important implications in the economic assessment. Firstly, traffic will in some cases prefer routes that are not the best in terms of national resource costs. Secondly, because resource and perceived costs may change relative to one another over time, the extent to which routes that are uneconomic in national resource terms are chosen will also vary over time. These effects will generally only be important on larger networks where significant movements have a choice between distance saving and time saving routes. Data output Traffic flows derived from the traffic appraisal can be used directly in the economic evaluation. However, the best estimate of journey times should always be used. It has been found that the best estimate will not always be produced from the traffic assignment model, whose main function is to produce validated traffic flows. The journey times from most traffic models will be inadequate for economic appraisal, since they represent averages that do not differentiate between travel at different times and in different traffic conditions. For some schemes, especially inter-urban ones, the flow group and speed-flow analysis incorporated in COBA is usually the most suitable method of calculating journey times. In areas where chronic congestion and interacting junctions are found, it is important that the predictions of the traffic model are thoroughly validated. Methods of determining the accuracy of journey times and the number of journey time runs required to achieve a given level of accuracy are contained in DMRB Volume 13 Section 1 Part 5, Chapter 10 (Local Journey Time Measurements). Compatibility between the assignment and COBA networks 3.13 The assignment network and the COBA network need to be compatible in three respects: (i) Extent of network; (ii) Description of links and junctions; and (iii) Incorporation of future changes to network. 3.14 3.15 The principles guiding the extent of the COBA network are discussed in further detail in the following paragraphs. While it is important that there are no major discrepancies between the assignment network and the COBA network, slight differences are inevitable, given that link and junction descriptions in the two networks may differ. The assignment network incorporates speed/flow effects only where capacity restraint assignment procedures are used. Also, assignment networks typically do not model explicit flow-dependent delays at junctions. Furthermore, the assignment network

3.9

3.10

3.11

3.12

COBA guidance

15

and the COBA network can operate for different parts of the day: the assignment network often represents a 12 or 16-hour day, whereas COBA models a 24-hour day using flow groups. The facility in COBA to print out journey speeds and times for each flow group in a specified year can be used to check that the two networks are broadly compatible. 3.16 Competing and complementary schemes can be incorporated into COBA by amending DoMinimum and Do-Something network descriptions and traffic flows in the appropriate year. Where assignments change, a new set of link and junction flows should also be entered. If the economic case for a scheme is dependent on the provision of a neighbouring scheme, sensitivity testing should be carried out. Defining the fixed trip matrix 3.17 A common source of error in the use of COBA occurs when the fixed trip matrix assumption is violated or when the wrong fixed trip matrix is chosen. This can happen when a large assignment network is simplified for economic assessment by: (i) Omitting links whose flows only change by a small amount between Do-Minimum and DoSomething, or (ii) Cordoning the assignment, i.e. drawing a cordon around the scheme in the network as modelled and forming a trip matrix of movements crossing the cordon in order to isolate reassignment effects within the cordon. 3.18 It is recommended that, as a general rule, the extent of the COBA network should be the same as the assignment network, to avoid possible bias from (i). Even the smallest differences in the matrix of trips used on the assessment networks can affect the results. Errors from (ii) can occur when the cordon is drawn too close to the scheme through links with a significant flow change. When defining the extent of the cordon it is necessary to judge whether flow differences on a link are significant in terms of the scheme being evaluated or simply "noise" in the traffic model. The recommended procedure to follow in order to determine the NPV of a Do-Something scheme relative to the Do-Minimum is, therefore, to consider the following issues: If the benefits / disbenefits to diverting traffic are likely to be significant, the COBA network should be extended to encompass the alternative routes; or If the benefits to diverting traffic are not likely to be significant, the Do-Minimum assignment may be cordoned and the comparable Do-Something flows should be derived by assigning the cordon point matrix to the reduced Do-Something network. 3.21 When comparing several Do-Something options that produce significant flow changes at the cordon compared with the Do-Minimum, but have similar effects on the cordon, it may be worthwhile to cordon point the Do-Something matrix. This procedure is likely to overestimate scheme benefits by overstating the Do-Minimum levels of congestion; however the incremental economic results of the alternatives are valid. When to model junctions 3.22 One of the benefits associated with the construction of a new road scheme can be the relief of the delay that vehicles experience at congested road junctions in the existing road network. The amount of delay that vehicles suffer when queuing at a junction is dependent on the capacity of that junction and the traffic flow level. On the existing network congestion delay occurs when junctions are operating close to, or above, their capacity for long periods. There will, of course, also be delay associated with junctions on the new road scheme. The decision on whether junctions need to be modelled explicitly depends on the classifications of the links joining the nodes in the modelled network.

3.19

3.20

3.23

3.24

COBA guidance

16

3.25

Speed-Flow relationships Junctions are represented in the COBA network description by nodes joining links together. If any of these links represent roads on which the rural speed/flow relationships apply, then the nodes should be classified and modelled as junctions. This is because the rural speed/flow relationships in COBA take no account of the delays to traffic at junctions, and it follows that junctions on rural roads should be modelled explicitly. It is generally only necessary to model the junctions where there is a significant flow change, but there are no fixed rules relating to the flow change that necessitates a junction being modelled. The user will need to make a judgement regarding the significance of each junction to the assessment of the scheme being studied. For urban and suburban roads the speed prediction formulae within COBA do take account of delays at junctions. In general, therefore, most junctions in urban areas do not need to be modelled explicitly. The exception is the junction that is heavily congested (either now or at some time during the life of the scheme) to such an extent that the journey times through the network resulting from the application of the standard urban speed/flow relationships are incompatible with actual timed runs. This junction should therefore be modelled explicitly in conjunction with speed/flow information from timed runs. Junction interaction COBA models all junctions in isolation, with all arrivals assumed to be random; no allowance is made for any junction interaction. Therefore, great care must be taken when modelling junctions in the same area of the network where the capacity of one junction controls the flow at another, either now or at some time in the future. If this occurs it may only be necessary to model the controlling junction. Double counting of delay must be avoided. Maximum delay factors The concept of maximum delay has an important bearing on junction evaluation in COBA. The form of the COBA junction delay formulae implies that, as the capacity of the junction is reached, delays increase very rapidly. However, COBA includes, as default, a maximum delay at junctions of 300 seconds for peak flow group types. If journey time evidence warrants a local adjustment for any individual junction, the user may change the value for the peak group upwards or downwards (maximum 900 seconds). The maximum delay is attributed to all vehicles but on an arm-by-arm or stream-by-stream basis. Maximum delays for the non-peak flow groups are calculated as a proportion of the peak flow group maximum delay, as shown in Table 3.1. Table 3.1 Maximum delay by flow group Flow Group Type Type 1 (off Peak) Type 2 (Adjacent to Peak) Type 4 (Peak) Proportion of Maximum Delay 0.4 0.6 1.0 Default Maximum Delay (seconds) 120 180 300

3.26

3.27

3.28

3.29

3.30

Without a maximum delay cut off, the formulae would predict much higher delays in the future at heavily congested junctions than have been generally observed at existing congested junctions. Such high delays are considered unrealistic, bearing in mind that it applies to all traffic in the flow group. The 300-second maximum delay for the peak flow group presumes that some vehicles will queue for considerably longer than this, some for a shorter period, with the average for all vehicles over the entire peak period being 300 seconds. In practice some drivers will act to prevent such long delays occurring, by travelling by a different route or travelling at a different time. The maximum delay acts as a substitute for detailed modelling and evaluation of such behaviour, for which COBA is not designed.

COBA guidance

17

3.31

Where junctions are explicitly modelled and COBA junction delay benefits are an important element, the realism of the magnitude of the junction delay benefits should be examined by considering the following: Do-Minimum improvements or, where Do-Something delays are large, Do-Something junction optimisation. Small changes in the coding of junction layouts can sometimes yield significant changes in junction delay costs; Comparison of COBA and measured journey times, for example, where junctions interact. In practice, it is more common for local maximum average delays to be less than 300 seconds rather than more, and Explicit modelling of critical junctions outside COBA. In exceptional cases, the COBA user may wish to consider whether in-depth analysis of a critical junction is necessary.

3.32

A particular problem with calibrating maximum delays using timed runs is that it is impossible to calibrate future scenarios. For example, a junction may be coded at 300 seconds maximum delay on the basis of present journey time evidence, but may give rise to longer delays in the future due to traffic growth. This may be tested using more sophisticated junction modelling techniques as found in congested assignment packages that explicitly model local diversionary reassignment. Advice from the National Roads Authority should be sought in such circumstances. Accident only and delay only junctions Delay only junctions are used to model user specified geometric delay at roundabouts or points in the network where the speed flow relationships do not apply. The specified delay is applied to all vehicle types and is constant over all flow groups. An accident only node may be used to model accidents at a node without modelling junction delays. This is important for urban junctions, where the delay is subsumed into the link speed/flow relationships. As national accident data is only available on a link and junction basis, accident only nodes are not to be used without prior approval of the NRA. Journey time validation techniques

3.33

3.34

3.35 3.36

The procedures recommended for the validation of the COBA output are outlined in Section 8. The following paragraphs relate specifically to the approach to journey time validation. Often local journey time measurements over the road network will have been made for the traffic modelling stage of scheme assessment. It is good practice to check observed journey times with those on the assignment networks and those computed by COBA. This comparison may bring to light specific instances where it is worth checking the COBA Do-Minimum journey times by carrying out more timed runs. (Note: COBA journey times for each flow group are printed out for a specified year in Phase 8 of the COBA output). Generally, journey times are only required on roads subject to the national speed limits for single carriageways and dual carriageways. On roads subject to a local speed limit, journey time measurements are only worthwhile if there are significant flow differences between the DoMinimum and Do-Something networks. Journey time measurements carried out for COBA should be used to estimate average speeds only; it is not possible to determine the speed/flow slope to any reasonable degree of confidence from a small-scale survey. The journey time survey should be geared towards estimating observed journey time on the whole of the bypassed section of route; individual estimates of speed on each link are not required. In general, the longer the section of route that is being bypassed, the lower should be the variability of observed journey times. Fewer observations should be needed to estimate the journey time over a longer section of route to a given level of accuracy. The most important consideration for local journey time surveys is that the journey times should be representative of conditions throughout the year. A large number of runs carried out on one day will usually be worth less than fewer runs spread over several days. Generally, measurements should be taken to cover both the peak and off-peak hours of the day.

3.37

3.38

3.39

COBA guidance

18

3.40

The moving observer method is the most widely used method of carrying out journey time measurements. Alternatively, registration number surveys should be considered as an alternative to the moving observer method these have the advantage of providing high sample rates. The results of the journey time survey should be sent with the COBA Appraisal Report and should include information on the number of runs carried out in each time period with an estimate of their accuracy and details of the level of traffic flow at the time. The survey results should be compared with the COBA modelled times and flows in each flow group. If observed and modelled journey times are not in reasonable agreement, then the speed/flow relationship that has been used for a particular link should be reconsidered, and possibly a local relationship defined. For full guidance on undertaking journey time surveys, and comparing with the model output, reference should be made to DMRB Volume 13, Section 1, Part 5, Chapter 10 (Local Journey Time Measurements).

3.41

3.42

COBA guidance

19

Parameters

COBA guidance

21

Parameters

General 4.1 An Irish version of the COBA program has been developed that contains a series of default values for parameters relating to items such as economic values (for example, time, accidents, vehicle operating costs and carbon costs), accidents (rates and severity), annual traffic flow patterns and vehicle composition (E and M-factors and flow groups). These parameters have been developed specifically for the assessment of Irish road schemes. This Section provides a general discussion on the Economic Parameters, followed by a more detailed discussion of the approach to selecting parameters at the different phases of scheme development. A table is provided at the end of this Section, which summarises the data source for each parameter, and whether local or national values should be used. The National Parameter Values Sheet is available from the NRA website. Economic parameters 4.5 The economic input parameters are fixed and do not change by project phase. Some notes on these parameters are included below. Discount rate A discount rate of 4% should be adopted; this is in line with latest Department of Finance guidance. Present value year The Present Value Year is the year to which all prices are quoted and discounted back to. The Present Value Year for COBA assessments shall be 2002. Consumer Price Index COBA runs in a price base year and all inputs into the program need to be in this year. Therefore, if costs are developed for a different year, they need to be changed to reflect the expenditure power of that sum of money in the price base year. The Irish Consumer Price Indices (CPI) are used to adjust prices to the price base year For example, to adjust a price from July 2003 to the present value year (for the purpose of this example, average 2002 prices) you would apply the following formula:

4.2

4.3 4.4

4.6

4.7

4.8

4.9

Cost in average base year prices = Latest Available Cost

4.10

CPI 2002 CPI July 2003

Up to date information on the CPI can be obtained from the Central Statistics Office website (http://www.cso.ie). Relative Price Factor The Relative Price Factor (RPF) is used together with the CPI to adjust construction costs to the present value year. The resulting cost is the cost that would occur if construction costs were at their long-term average at the time of construction. Current advice is for a value of unity to be adopted for the RPF at all project phases. Value of time Values of time are presented for three journey purposes. These are: Journey in the course of work (work trips); Commuting to/from work (commuting trips); and

4.11

4.12 4.13

COBA guidance

22

4.14

All other journey purposes (other trips). For undertaking CBA of National Road schemes different values of time are to be applied for work, commuting and other non-work journeys. No distinction is to be made between modes of transport for different journey purposes. For example, the value of work time for a bus passenger is considered to be equal to the value of work time for a car passenger. Vehicle Operating Costs fuel COBA calculates fuel costs using a mathematical function based on the average speed on each link. The function includes a number of constants known as the a, b, and c parameters. These vary by vehicle class to reproduce the different fuel operating cost characteristics of different vehicle types. The COBA vehicle operating cost formula is of the form:

4.15

4.16

C = a + bV + cV 2
where: C = cost in cents per kilometre per vehicle; V = average link speed in km/h, and a, b and c are vehicle category parameters. 4.17 The a, b and c parameters are contained in COBA both in cents/km and litres/km at resource cost. Conversion between the two is simply a case of factoring the parameters by the resource cost of fuel (cents per litre). Vehicle Operating Costs non-fuel Non-fuel vehicle operating costs include oil, tyres, maintenance and mileage. Only items that vary with the use of the vehicle are measured and parameters are presented by vehicle class. Non-fuel costs are calculated using the equation:

4.18

C = a 1 + b 1 /V
where: C= cost in cents per kilometre per vehicle; V= average link speed in km/h, and a1 and b1 are vehicle category parameters. 4.19 Accident costs Accident costs are a combination of the costs due to each casualty and the costs of insurance, Garda and damage to property. Insurance, Garda and damage to property costs are presented by accident type. Casualty costs have been developed for the following types of casualty: Fatality; Serious injury, and Minor Injury. 4.21 Carbon costs Carbon emissions are considered in terms of the change in the equivalent tonnes of carbon released as a result of implementing a highway scheme. Carbon emissions are estimated from fuel consumption in the Do-Minimum and the Do-Something options. The change in tonnes of carbon emitted and the monetary value given to the change is calculated in COBA. Maintenance costs Road maintenance costs have been developed for the following road classes: Standard 2-lane with hard shoulder; Wide 2-lane with hard shoulder; 2+1 road; Type 2 Dual 2 Lane Carriageway, and Type 1, Standard and Wide Dual Carriageway/Motorway.

4.20

4.22

COBA guidance

23

4.23

Maintenance costs are presented as a rate per kilometre per year. The maintenance type relating to each road class is shown in Table 4.2. Taxation COBA requires inputs on average tax on final consumption, tax on fuel (final consumption) and tax on fuel (intermediate consumption). This allows input costs to be converted from resource costs to market costs, and hence allows an assessment of the cost and benefit stream for different market segments, as described in more detail in Section 6 of this document. Future forecasting of economic parameters

4.24

4.25

In future years the real value of a number of parameters will change. Growth factors have been developed to take into account these changes. Value of time Forecast growth in real gross national product (GNP) per person employed has been used to determine future changes in the real value of time. The same growth factor is applied for work, commuting and other non-work time, although the facility exists to use different factors. Value of accidents Forecast growth in real GNP per person employed has also been used to determine future changes in the real value of accidents. The assumption is that the values of most elements of accident costs are proportional to national income. Vehicle Operating Costs Vehicle Operating Costs, both fuel and non-fuel, are assumed to remain constant in real terms over time. Carbon costs Carbon costs change with respect to time and these changes are included within the COBA program. Taxation Any change in taxation levels is proposed by Government, and the implications of such tax changes on various segments of society are considered independently from the CBA process. The CBA should therefore assume that tax rates remain static throughout the assessment period. Route selection Default parameter values When undertaking an assessment of different route options to inform the Route Selection process, default values may be used as inputs into COBA. These default values have been developed by the NRA specifically for application to Irish road schemes. The use of default values avoids the need for significant volumes of local data to be collected during the Route Selection phase. Where the default parameter values are expected to change over time, they are presented separately on the NRA National Parameter Value Sheet. These default values may be updated periodically when more data become available, so Appraisal Teams should confirm with the NRA that they are in possession of the latest version of the parameter sheet. Some notes relating to the use of the NRA National Parameters Value Sheet for Route Selection, are given below. Network classification Three network classifications have been adopted for the assessment of Irish road schemes, as described in Table 4.1. Users of COBA should select the network classification that best describes the scheme being assessed. Note - the UK abbreviations for network classifications are still used.

4.26

4.27

4.28

4.29

4.30

4.31

4.32

4.33

4.34

COBA guidance

24

Table 4.1 Network classification Network Classification Motorway National Primary excluding motorway National Secondary COBA INPUT MWY TNB PNB

4.33

Road class COBA uses a number of Road Classes to describe the nature of each link throughout the network. The road class determines the type of data that is to be input for each link, and therefore is a key parameter to be selected for each link. Up to 20 road classes are definable in COBA, with each representing a certain road description as contained within the DMRB. Road classes are summarised in Table 4.2. Accident types For the Route Selection phase, link and junction accidents have been combined to produce default values for accident rates, severity splits and costs, which are all attributed to links. Six accident rates, based on road type, have been considered, and their relationship to the road classes is shown in Table 4.2. It should be noted that accident type numbers are not the same as the road class numbers used to define the speed-flow relationships. Accident data provided for each accident type has been split according to the speed limit structure for roads. For National Roads (including National Motorways) and non-national roads in non-built up areas, accident data relating to speed limits of 80 km/h or greater are applicable; whereas for built up areas, data relating to speed limits not exceeding 60 km/h should be used. Unlike the UK procedure, no distinction has been made concerning the standard of the road, i.e. old or modern. Maintenance type Each link will have a maintenance type, which relates to the type of link defined. Maintenance costs can distinguish between wide and standard single carriageways, motorways and dual carriageways, which can be standard Type 1, Type 2, or 2+1. A total of six maintenance types have been defined, which relate to the road classes as outlined in Table 4.2. Accident change coefficients COBA takes into account the existing long-term declining trend in accident rates and severity. The program uses a -factor to model the reducing number of accidents in each assessment year and also to model the reduction in the average number of casualties, by severity split, resulting from each accident. The change in accident rates and number of severities per accident is explained by the relationship:

4.34

4.35

4.36

4.37

AN = A0 N
Where: AN = the accident rate or number of casualties per accident N years after base year; A0 = the accident rate or number of casualties per accident in the base year, and N = change coefficient raised to the power N. 4.38 For the assessment of Irish road schemes -factors have been developed to take into account Irish policy on the reduction of road traffic accidents. Default -factors contained within the COBA program are provided in the NRA National Parameter Value Sheet. The -factors are applied for any year between 1995 and 2010. Between 2011 and 2020, and between 2021 and 2030 the reduction factors are assumed to be one half and one quarter

4.39

COBA guidance

25

respectively of the 1995 to 2010 reduction. For example, if the coefficient is 0.9, then it is 0.95 for the period 2011 to 2020 (or [1+]/2), and 0.975 for the period 2021 to 2030. Zero change is assumed post 2030. 4.40 There is no facility to change how varies with respect to time as this is embedded within the COBA program. Traffic flows 4.41 Ideally, traffic flows should be input as Annual Average Daily Traffic (AADT) values, extracted from an appropriate traffic model (see Section 3). If flows are entered as 12 or 16-hour values then the default values for the E and M-factors will be used by the program to convert these to Annual All Vehicle Flows (AAVF). Table 4.2 COBA road classes, accident and maintenance type summary Road Description Rural Reduced Single (7.0m) Carriageway S2* Rural Standard Single (7.3m) Carriageway S2* Rural Wide Single (10.0m) Carriageway S2* Type 1 Rural Dual Carriageway (Standard)* Type 1 Rural All Purpose Dual Carriageway (Wide)* Type 2 Rural Dual Carriageway** Rural All Purpose Dual 3 or more lane carriageway 2x2 Motorway (Standard 7.0m)* 2x2 Motorway (Wide 7.5m)* 3x3 Motorway 4x4 Motorway Urban All Purpose Dual Carriageway (Central)* Urban All Purpose Dual Carriageway (Non Central)* Urban All Purpose Single Carriageway (Central) Urban All Purpose Single Carriageway (Non Central Small town All Purpose Dual Carriageway Small town All Purpose Single Carriageway Suburban All Purpose Dual Carriageway Suburban All Purpose Single Carriageway 2+1 Road (with central safety barrier)* * 2+1 Road - (without central safety barrier) * * User Defined all vehicle relationship User Defined light/heavy vehicle relationship COBA Road Class 1 1 1 2 2 3 4 5 5 6 7 8 9 8 9 10 10 11 12 13 14 15-16 17-20 COBA Accident Type 4 4 4 10 10 10 10 1 1 1 3 10 10 4 4 10 4 10 4 11 5 COBA Maintenance Type 1 1 1 2 2 2 3 4 4 5 6 2 2 1 1 2 1 2 1 2 1

4.42

Speed Flow curves In the absence of comprehensive local information, UK derived speed-flow curves have been applied to Irish road classes. The only exceptions are for Type 2 Dual Carriageways and 2+1 roads, where the approach is as described below: For Type 2 Dual Carriageways, a modified form of the rural speed-flow relationship has been adopted. The modifications have been implemented to reflect the capacity impact of junctions and lower speed limit that would be used on these types of roads. In the case of 2+1 roads, the speed-flow relationship has been derived based on research undertaken on behalf of the NRA by Malone ORegan Scott Wilson Consultants.

COBA guidance

26

4.43

Different speed-flow predictions will be made by allocating a link to the appropriate road class of Table 4.2. Further details on the nature of the speed-flows defined for each road class can be found by reference to DMRB Volume 13, Section 1 Part 5, Chapters 1 to 9 (Speed on Links). The definitions of the road classes highlighted by a * in Table 4.2 are consistent with those in NRA TD 27, which sets out the dimensional requirements for road cross-sections for new National Roads. The Type 2 Dual Carriageway and 2+1 roads are as defined in documentation available from the NRA. The remaining road classes complete the range of classes that are likely to be required in the COBA assessment and are consistent with the definitions provided in the UK DMRB Volume 13. The following explanatory notes are provided on the different road classes. Where speed limits are quoted these relate to the new speed limit structure recommended by the Working Group on the Review of Speed Limits. These limits are effective since January 2005: (i) Rural single carriageway and dual carriageway roads are normally subject to a speed limit of either 100km/h for National Roads or 80km/h for Non-National Roads. This includes all road classes 1 to 3. Motorways (Class 5 and 6) are generally subject to a speed limit of 120km/h. Classes 8 and 9 are used for roads in built up areas subject to speed limits of 50 km/h (31 mph). The distinction is made between central and non-central urban areas, with central areas defined as those including the main shops, offices and central railway stations, with a high density of land use and frequent multi-storey development consistent with a central business district (CBD). Streets containing commercial or industrial development but not of a high density CBD nature should not be included within the central area. Non-central areas comprise the remainder of the urban area. Suburban roads, classes 11 and 12, apply to the major suburban routes in towns and cities where the speed limit is generally 60 km/h. The main urban speed flow relationships do not apply to towns with populations of less than 70,000, for villages or for rural roads with short stretches of development. In such cases the small town road class (Class 10) should be used.

4.44

4.45

(ii)

(iii) (iv)

4.46

User-defined Speed-Flow relationships The facility exists for the user to define special speed-flow relationships. They should only be used in special circumstances where the normal ranges of speed-flow relationships do not apply. There are two types of user defined speed-flow relationships: (i) Special Road Classes 15-16: The user may define the relationships by the use of six speeds equated to specific flow levels. The relationships apply to both light and heavy vehicles and are independent of link geometric parameters; and Road Classes 17-20: Here the user can define the basic constants to generate relationships similar to the form used for the rural road classes. Light and heavy vehicles can be modelled separately.

(ii)

4.47

Traffic growth The default growth rates for cars, light goods vehicles (LGV), other goods vehicles (OGV1 and OGV2) and public service vehicles (PSV), for both low and high traffic growth scenarios have been developed. For PSV the values are UK default growth rates. The base year for traffic growth is 2002, and is used as the basis for all traffic growth forecasts in the CBA work. The growth rates are based on the information contained in the NRA Document Future Traffic Forecasts 2002-2040 published in 2004. The Low Growth scenario used in the model is the growth forecast for all public roads, whilst the High Growth scenario is the growth forecast for the national primary road network. These indices have been used to produce the default growth scenarios within the COBA program, and must be used for the evaluation of all national road schemes. A summary of this information is provided below in Table 4.3. Vehicle mix proportions Average vehicle proportions are defined for use within the three network categories as outlined in Table 4.1. Vehicle proportions are based on the five COBA vehicle categories (Car, LGV, OGV1, OGV2 and PSV) and are outlined in the National Parameter Value Sheet.

4.48

4.49

COBA guidance

27

Table 4.3 Future traffic forecasts 2002-2040 Low Growth Year 2002 - Base 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 Cars & LGV 100 104 107 110 113 116 119 122 125 127 129 131 133 135 137 139 141 143 144 146 148 149 151 152 153 155 156 157 158 159 160 161 162 163 164 164 165 166 167 HGV 100 103 106 109 111 114 117 120 123 126 128 130 132 134 136 138 140 143 145 147 149 151 153 155 157 159 161 163 166 168 170 172 174 176 178 180 182 184 186 PSV 100 100 101 101 101 102 102 102 103 103 104 104 104 105 105 106 106 107 107 108 109 109 110 110 111 112 113 113 114 115 116 116 117 118 119 119 120 121 122 Cars & LGV 100 105 110 114 118 123 127 131 136 139 142 144 147 150 153 156 158 161 164 166 168 170 173 175 177 179 180 182 184 185 187 188 190 191 192 194 195 196 197 High Growth HGV 100 105 109 112 116 120 124 128 133 137 140 143 146 149 152 155 158 161 164 167 170 173 176 179 182 185 188 192 195 198 201 204 207 210 213 216 219 222 225 PSV 100 101 102 103 104 105 106 107 108 109 110 112 113 114 115 117 118 119 121 122 124 125 127 128 130 132 134 135 137 139 141 143 145 147 149 151 153 155 157

4.50

Traffic flow groups To take into account the variations in the level of traffic flow and vehicle composition, the 8,760 hours of the year are divided into different proportions (numbers of hours) called flow groups. Each flow group represents a different level of flow. The peak period flow group contains those hours throughout the year that are defined as peak hours (Flow Group 4) and defines the proportion of annual traffic travelling during those hours. Other flow groups represent the adjacent to peak (Flow Group 2) and off peak (Flow Group 1) periods. When undertaking CBA of National Road Schemes in Ireland, only four flow groups are to be used no distinction is made between an hourly flow occurring during a weekday or a weekend. Each Flow Group Number should also be further defined by a Flow Group Type. Relevant Flow Group Types are Type 1 (Off Peak) and Type 3 (Peak).

4.51

4.52

COBA guidance

28

4.53

For modelling purposes the flow in each hour of a given flow group is considered to be at a constant proportion of the annual average hourly traffic (AAHT). This proportion is defined as the flow group multiplier or d, the value of which is shown in the table. The structure of flow groups for use in the assessment of Irish National Road schemes is outlined in Table 4.4. Table 4.4 Default flow group structure Network Classification Flow Group Number 1 2 3* 4 1 2 3* 4 1 2 3* 4 Number of Hours in Flow Group 4,600 4,060 0 100 4,600 3,660 0 500 4,600 3,660 0 500 Flow Group Hours 1 4,600 4,601 8,660 8,661 8,760 1 4,600 4,601 8,260 8,261 8,760 1 4,600 4,601 8,260 8,261 8,760 Flow Group Type: [1 = Ordinary Flow Group (off-peak); 3 = Peak Flow Group] 1 1 2 3 1 1 2 3 1 1 2 3 FLOW/ AAHT d 0.455 1.587 0.000 2.239 0.404 1.568 0.000 2.320 0.378 1.579 0.000 2.484

4.54

Motorway (MWY)

National Primary (TNB) National Secondary (PNB)

* The default flow group structure contains no Flow Group Number 3 (or corresponding Flow Group Type 2). However, the COBA program contains a proxy Flow Group Number 3 with 0 hours, for a Flow Group Type 2 with a multiplier (d) of 0.000.

4.55

Vehicle category proportions Based on the flow groups in Table 4.4 default national values have been derived for the vehicle category proportions in each of the flow groups, for each network classification. Category proportions have also been derived for vehicles falling into each of the three time modes. Vehicle occupancy National default values for the average number of people occupying a vehicle of a given category and time mode (work, commuting and other) have been developed from roadside interview data. The default values also give occupancy rates for two of the three flow groups, described in Table 4.4; i.e. Flow Groups 2 and 4. These relate to the adjacent-to-peak (intra-peak) period and peak period respectively. For Flow Group 1, which relates to the overnight off-peak traffic, the UK default value has been taken, since the source roadside interview data only covered the 12-hour period from 07:00 to 19:00. Preliminary design

4.56

4.57

4.58

At the Preliminary Design phase a detailed COBA assessment must be carried out using locally derived parameter values for the items listed below. For the remaining parameters, national default values can be used, as described above for the Route Selection phase. Seasonality index The Seasonality Index is an important descriptor of annual traffic flow patterns. It is defined as the ratio of the average August weekday (Monday to Friday) flow to the average weekday flow in the neutral months of April, May, June, September and October (excluding periods affected by bank holidays). Long-term automatic traffic counter data is required to derive local values. A good estimate can be arrived at by comparing the weekday traffic flows from a three-week continuous count in August with one from late May/June or October.

4.59

COBA guidance

29

4.60

E-Factor The E-factor converts flows entered into the program as 12-hour values into the 16-hour equivalent. Local relationships between the 12-hour and 16-hour flows can be derived from long-term automatic traffic counts. M-Factor The M-factor converts flows entered into the program as 16-hour values into an Annual All Vehicle Flow (AAVF). A local M-factor can be derived that relates the average weekday 16hour count in the month specified to the annual all vehicle flow. Long-term automatic traffic counter data will be required to do this. Vehicle category proportions and adjustment factors The proportion of each vehicle category should be computed based on the weighted average over the entire network, taking into account the lengths of the various links and the total flow on them throughout the year (i.e. vehicle kilometres) and should be representative of the proposed scheme. Traffic proportions can be input as 12 or 16-hour weekday values or the Annual Average Daily Traffic proportions (AADT). If either 12 or 16-hour weekday values are entered, the program will convert them into the AADT proportions using default adjustment factors (the A(I) factor) based on national default values. It is unlikely that the CBA Team will have sufficient information to derive local A(I) factors, hence the NRA National Parameter Value Sheet contains default values for when 12 or 16-hour traffic proportions have been entered. Vehicle category proportions by flow group From long-term traffic data the vehicle proportions can be derived for each flow group for the network under consideration. Classified traffic data from count sites in the vicinity of the scheme will need to be collected and should ideally contain data for each of the 8,760 hours of a year. Where this is not possible, due to missing data, some degree of infilling is permissible; for example, using data relating to the average from similar days, or hours within the same month. The complete years data should be arranged in ascending order, i.e. the largest total flow would be ranked at number 8,760 and the smallest flow ranked number 1. The average vehicle proportions relating to the yearly hours given in Table 4.4 should be calculated (where 1 relates to the hour with the smallest total traffic and 8,760 the hour with the largest total traffic). Vehicle proportions by time mode Local values for the proportions of cars and light goods vehicles in work, commuting and other time may be derived from roadside interview data where sufficient information is available. Vehicle proportions by time mode must be disaggregated by flow group. Since interview data will normally cover a 12-hour period between 07:00 and 19:00, vehicle proportions by time mode can only be developed for Flow Groups 2 and 4, the adjacent-to-peak (intra-peak) and peak hours respectively. For Flow Group 1, UK default values should be used, as presented in the NRA National Parameter Value Sheet. Vehicle occupancy Roadside interview data can be used to derive vehicle occupancy rates for work, commuting and other non-work time modes. These data should also be disaggregated by vehicle class and also by flow group. Since the interview data will normally cover a 12-hour period between 07:00 and 19:00, average occupancies can only be determined for Flow Groups 2 and 4. For Flow Group 1, UK default values should be used. Accident rates and casualty proportions Where local accident data are available and are considered to be reliable, these should be used in preference to the national default accident rates and casualty proportions.

4.61

4.62

4.63

4.64 4.65

4.66

4.67

4.68 4.69

4.70

COBA guidance

30

4.71

Local data on the occurrence and severity of accidents should relate to a period when the conditions on the road have been broadly unchanged (for example, no abnormal changes in traffic flow, no changes in junction design or road geometry, etc.). Ideally, local data should cover the five years previous to the COBA assessment and must cover a period of at least three years. The user can either calculate the observed accident rate in terms of the number of accidents per million vehicle kilometres (mvkm), or alternatively input accident numbers from which the program will calculate a local link accident rate. In the latter case, the number of accidents in each year must be input, including zero for those links or years where no accidents occurred, and COBA will then internally produce a local accident rate (for each link). If the user is calculating a local accident rate outside of the COBA program the following procedure should be followed: For each year over the five-year period all personal injury accidents (PIA) should be summed, i.e. all fatal, serious and slight accidents, to give values X1, X2, X3, X4 and X5 (where X1 is the total number of PIA in year 1 etc.). The total traffic flow on the link over this period, Y, (expressed in million vehicle kilometres mvkm) should also be calculated. The local accident rate is then defined by: PIA/mvkm = (X1+X2+X3+X4+X5)/Y

4.72

4.73

If default accident values are to be replaced with local values, the COBA user must demonstrate that the local severity split is significantly different in statistical terms from the default national averages, and not a result of one or two particularly bad accidents, the effect of which will be evened out by less extreme accidents as time goes by. Adjustments are made automatically within the COBA program to account for accident underreporting. It is recommended that the local accident rates be input as combined link and junction rates. Where no local data are available and for the Do-Something scheme components, the default national parameter values should be used. Construction documents preparation and tender award

4.74 4.75 4.76

4.77

The approach to CBA at construction documents preparation and tender award phase is the same as at preliminary design. Final account / closeout

4.78

The approach to parameter coding at final account / closeout, mirrors closely that taken at Preliminary Design, with the exception that observed Do-Something traffic flows should be input along with the outturn scheme costs. The collection of traffic data for undertaking a final account / closeout COBA can be an expensive undertaking if the network is particularly large. One approach to rationalizing the number of counts that need to be taken is to validate the Do-Something traffic model based on a sample of post scheme implementation counts, so that the model can be used to obtain revised Do-Something flows. One of the primary functions of the final account / closeout COBA is to enable predicted benefits to be compared with actual behaviour following opening of the new scheme. It is for this reason that no parameters should change between the undertaking of the Preliminary Design and Final Account / Closeout CBA, unless there is adequate justification for doing so.

4.79

4.80

COBA guidance

31

Summary 4.81 In Table 4.5, the approach that should be taken for each parameter is summarised for the different stages of the life of a scheme. Parameters are either based on the national default values or locally derived data. In a few instances, the existing UK default values have been retained since there are insufficient data at present to determine Irish parameters. Where it is indicated that the national default values are to be adopted, the user must not change the values already contained within the COBA program. For each parameter, the COBA KEY number is provided: this corresponds directly with the record on the COBA input deck.

4.82

COBA guidance

32

Table 4.5 Summary approach to parameter coding PARAMETER Present value year / appraisal period Traffic proportions Vehicle mix groups Seasonality index E-factor M-factor Growth of traffic Tax Rates Tax rate changes Discount rate Vehicle operating cost Accident costs Annual compound growth rates Annual compound growth for fuel and nonfuel Values of time per person Accident rates, severity splits for link / junction combined Accident reduction factors (-factors) Maintenance costs Accident proportions Adjustment factors A(I) Flow group multipliers Flow groups composition correction factors Annual % change in vehicle occupancy COBA KEY 003 006 007 008 008 008 009 013 014 015 016 017 019 020 021 023 023 024 028 029 030 031 032 ROUTE SELECTION NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL DEFAULT PRELIMINARY DESIGN/CONSTRUCTION DOCUMENTS NATIONAL LOCAL LOCAL LOCAL LOCAL LOCAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL LOCAL NATIONAL NATIONAL LOCAL NATIONAL NATIONAL LOCAL DEFAULT FINAL ACCOUNT/CLOSEOUT NATIONAL LOCAL LOCAL LOCAL LOCAL LOCAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL NATIONAL LOCAL NATIONAL NATIONAL LOCAL NATIONAL NATIONAL LOCAL DEFAULT For running COBA, user must enter proportions (1.00) for Flow Group 3. Until directed otherwise the UK parameter values are to be used. COMMENTS

COBA guidance

33

PARAMETER

COBA KEY

ROUTE SELECTION

PRELIMINARY DESIGN/CONSTRUCTION DOCUMENTS

FINAL ACCOUNT/CLOSEOUT

COMMENTS At Route Selection phase, National Default values are used for Flow Groups 2 and 4, with UK default values for Flow Group 1. At Preliminary Design phase local values can be derived for Flow Groups 2 and 4, but UK default values are still to be used for Flow Group 1. For running COBA a value of 0 must be entered for all entries relating to Flow Group 3. At Route Selection phase National Default values are used for Flow Groups 2 and 4, with UK default values for Flow Group 1. At Preliminary Design phase local values can be derived for Flow Groups 2 and 4, but UK default values are still to be used for Flow Group 1. Note when inputting proportions for vehicle category 1 (normally cars), the flow group to which the proportions relate must be specified. Data for all the flow groups used must be input. For the purpose of running COBA, it is necessary to input a value of 0.50 for work and non-work proportions in Flow Group 3.

Occupancy of vehicle / category / time mode

033

NATIONAL / DEFAULT

LOCAL / DEFAULT

LOCAL / DEFAULT

Proportions of vehicles in each time mode

034

NATIONAL / DEFAULT

LOCAL / DEFAULT

LOCAL / DEFAULT

Annual changes to proportions in each time mode for a particular vehicle category

037

DEFAULT

DEFAULT

DEFAULT

Until directed otherwise, the UK parameter values are to be used.

COBA guidance

34

PARAMETER

COBA KEY

ROUTE SELECTION

PRELIMINARY DESIGN/CONSTRUCTION DOCUMENTS

FINAL ACCOUNT/CLOSEOUT

COMMENTS At Route Selection, Option Comparison Estimate to be used. At Preliminary Design, the cost estimate should be the weighted average of the Target Cost 1 and Total Scheme Budget.

Scheme costs

055

LOCAL

LOCAL

LOCAL

If the Target Cost 2 estimate produced at main contract award phase exceeds the Target Cost 1 estimate, the CBA must be updated accordingly. The final outturn cost is to be used for the purpose of Final Account/Closeout CBA

COBA guidance

35

How to use and run the COBA program

COBA guidance

37

How to use and run the COBA program

General 5.1 This section describes the structure of the COBA data file and the methods for compiling and editing a COBA input file. Structure of COBA data file 5.2 The COBA input data file consists of a Project. This is the basic unit for which COBA performs economic analyses and has the following structure: Basic Data; Do-Minimum Scheme, and One or more Do-Something Schemes (optional). 5.3 The first section, Basic Data, comprises network wide data that either remain constant throughout the Project or change in the same way from the base year value in each scheme of the Project. The Scheme is the data unit which describes the network and its traffic engineering details, and is constructed as follows: Scheme title; Network data; Costs, and Classification subsection. 5.5 The program assumes that the Do-Minimum data also apply to all Do-Something schemes unless told otherwise. Therefore only Do-Something scheme data that differ from the DoMinimum need be entered. The classification subsection comprises: Declassifications; Traffic flows; Road classifications; Junction classifications, and Accident specifications. 5.7 5.8 These data types must be in this sequence since, for example, road classifications are forbidden for roads with zero flow. Classification is the term used when a link or node is described by geometric parameters in order that the costs of negotiating the link or node can be calculated. Declassification is the process whereby the link or node can be retained in the network but the user costs will no longer be calculated. Detailed description of COBA data 5.9 The COBA deck, defining a particular Project is made up of a series of KEY records, which can be grouped together into the following headings: Control records; Basic data; Network data;

5.4

5.6

COBA guidance

38

Scheme data: declassification; Scheme data: costs; Scheme data: link flows; Scheme data: link classifications; Scheme data: node classifications; Scheme data: accidents, and Final control records. 5.10 Following an overview of the format of the data entered into the COBA deck, this section provides a summary of the contents of each of the above KEY categories. However, a detailed description of how to code individual KEY records in COBA is found in DMRB Volume 13, Section 1, Part 7. Format of data entry Characters Data are entered in the COBA input deck using individual KEY records. Data items may be divided into three groups: Alphanumeric (A): Contains any alphabetic, numeric or other character valid to the computer. Real (R): Contains a number with a decimal point. If a decimal point is omitted, COBA will assume its position is either on the right hand edge of the box for numbers intrinsically greater than unity (for example flows), or at the left hand edge for numbers intrinsically less than unity. This is indicated on the coding sheets by the presence of a dot on one edge of the box. Blanks (spaces) are treated as zeroes. Integer (I): Contains a number without a decimal point. Again, all blanks in the box are interpreted as zeroes. 5.12 It should be noted that if a data field, which should contain a number, contains any non-numeric character, (except a decimal point or minus sign where appropriate), COBA will halt immediately with a computer system error. This is outside the control of the COBA program. Limits on data Most numeric data are required to be within certain limits. These limits are principally to guard against mistyping of data in the wrong columns. If the data item is discovered to be outside the limits shown in the description of that data item an error message is printed; for example: This data item MUST NOT LIE OUTSIDE THE RANGE x TO y. 5.14 5.15 Errors result in the Project being truncated to a data check only. Some data items have two sets of limits, in which case the error limits are shown with the warning limits in brackets. If the item is inside the error limits but outside the warning limits, a warning message is printed; for example: WARNING - data item is LARGER THAN x. 5.16 The warning limits are chiefly determined by the range of values of observations made during the research on which COBA is based. Reclassification Repeat Inhibitor, RRI In COBA there is a distinction between those Do-Minimum improvements which would be made whether or not the Do-Something were to be implemented, and those which are in some way alternative to implementing the Do-Something, and would not be necessary if the DoSomething improvements were carried out. Normally all reclassifications that occur in the Do-Minimum are repeated in each Do-Something in the year in which they take effect in the Do-Minimum. For each particular Do-Minimum mid-scheme data change, (that is, flow changes, link or junction classification changes, or accident rate changes), it is possible to prevent the repeat in each Do-Something by entering an X in column 5. This X is referred to as the Reclassification Repeat Inhibitor, RRI. Obviously, the RRI is meaningful only in a reclassification that occurs in the Do-Minimum. In all other circumstances, it is ignored.

5.11

5.13

5.17

5.18 5.19

COBA guidance

39

5.20

Link and node names To a certain extent, link and node names are a matter of common sense. The values permissible are numbers between 1 and 9999. Further limitations are imposed in particular sections, for example: For a link or node to be classified or de-classified, or for its flow or accident data to be specified, it must obviously be in the network; For a link or node to be declassified, it must first be classified; Connectivity between links and nodes must be preserved. In flow data if the towards node is specified for a one-way link, it is obviously nonsensical if link and node are not connected in the network. Similarly, during junction data, the link on a subsequent record must refer to a link actually attached to the node being classified; The list of the arms of a classified junction must correspond exactly to the network description of that node, and The number of nodes that may be classified will generally be smaller than the number of network nodes because of storage space limitations.

5.21

Control Records Control records are used to switch the program from one data section to another, e.g. from the basic records to the network data. Basic Data Basic Data are those data records that apply to the whole of the project; for example, the years for which the Project is to be evaluated. The order in which Basic Data records are entered is important since some items interact with others. In general, entering records in the order of their free-format Keys will be successful, although the following guidelines should be noted: Mandatory records: KEYs 001, 003, 004 and 005 are mandatory records and should be input in that order. If any of these records are omitted, then the program run will be reduced to a Data Check Only. If the Print Phases required are to be specified, then KEY 002 should be input in numerical order within the mandatory records. Scheme years: COBA allows users to redefine the present value year on the same record as the first and last scheme years (KEY 003 - mandatory). The program will therefore not accept a Basic Data record of any type that specifies a year unless the scheme years record has already been accepted. Thus, since the KEY 003 record is mandatory (and is almost the first record of a file), the program will only fail in this way if the record itself is unrecognised. Vehicle categories: Most COBA runs will be performed using the default set of vehicle categories. If any item of Basic Data is entered which specifies a vehicle category by number, COBA will assume this number refers to the sequence of the default set. Thereafter the vehicle categories cannot be redefined. The following Basic Data types relate to vehicle categories: - Traffic proportions; - Local growth factors; - Adjustment factors; - Occupancy and change in occupancy of category; - Proportion of category in time mode, and - Increment to vehicle mode split.

5.22 5.23

5.24

The essential point is that if vehicle categories are to be redefined, this must be done before the traffic proportions for the project are specified. Network data Network data are those records that define the structure of the Do-Minimum network and how it changes with each scheme, or at reclassification years. Scheme data: declassification Declassification is the process whereby the link or node can be retained in the network but the user costs will no longer be calculated. Scheme data: costs By convention, COBA costs are input for the whole Scheme at the very beginning of the Scheme Data. Because of this, Scheme Costs will not be accepted in the reclassification

5.25

5.26

5.27

COBA guidance

40

section. Values for scheme capital costs, traffic related maintenance capital costs and delays during construction and maintenance works may be input in present value year terms in the correct year on KEY 055. 5.28 It is an NRA requirement that KEY055 is used to enter costs. Detailed guidance on how to derive the profile of scheme costs for input into COBA KEY055 can be found in Appendix 12. Scheme data: link flows A single record is used to input link traffic flows, either as total vehicles or by vehicle mix group. It is on this record that one-way links and their direction of flow are defined. Scheme data: link classifications These records are used to define the characteristics of all the links in the network for which user cost calculations are required. Each link is defined by a separate record, which allows the user to define the speed-flow relationship applicable to the link, geometrical quantities such as lane widths, accident type and speed limits. Scheme data: node classifications Those junctions that are classified are defined. Junction type, geometric parameters, operational parameters (e.g. signal stages), delays and turning flows are entered. Scheme data: accidents Accident data must be the last data type entered into any Scheme Data section. These data records are used to define observed accident rates or numbers, to overwrite the default values held within the program. Links may have accident rates only if they are classified. Final control records Control records define the end of scheme data, the end of the project data (when more than one scheme) and the end of the program run. Data preparation and editing Data preparation There are several methods a user can adopt to prepare the COBA input data file: An experienced user may prefer to edit an existing COBA input file; Entering data in the strict format defined in the COBA input coding sheets. These require the user to enter each character of the input data into specific, right-justified columns, or Using the program CSCREEN to edit an existing COBA input file or create a new file in the COBA data format on screen. Some data checking is undertaken as the file is created. 5.35 Data editing Edits to the data file can be done through a simple text editor. The file should be in plain ASCII format (i.e. containing no special characters and each line terminating with a carriage return). Suitable text editors include: The NotePad or WordPad program included with Windows software; and The EDIT program in DOS. 5.36 5.37 More sophisticated word processors may be used but the user must ensure that files are saved as a plain text/text only file. A default COBA input file is provided in Appendix 9. This file should be used as the starting point for coding. It contains all the default parameter values discussed in previous Sections.

5.29

5.30

5.31

5.32

5.33

5.34

COBA guidance

41

COBA output and interpretation

COBA guidance

43

COBA Output and Interpretation

General 6.1 6.2 This Section explains how the program brings together the various elements of the appraisal and presents the results in a series of summary tables. Historically, the COBA program has accepted input and worked in resource costs; this is still the case with the Irish version of the software. However, the calculus currently being used is the Willingness To Pay (WTP) methodology with the program converting costs and benefits to market prices using appropriate tax correction factors. Following the COBA run, appraisal results are summarized at the end of the COBA output file. The information describes the Economic Efficiency of the Transport System and is expressed in Market Prices. This output is provided in the form of a number of tables, each providing separate measures of output. The interpretation of such output, and the processes behind the calculation of the performance indices is described in this Section. It is important to note that COBA is only able to allocate the elements of the appraisal that the program calculates. There may be other significant costs and benefits that should be included in the decision making process. Description of COBA output 6.5 6.6 The most significant tables of the COBA output occur in Phase 16. These are: Table 14, Phase 16 Conversion of Travel Costs to Market Prices by Vehicle Category This table (see Table 6.1) shows the calculations necessary to convert the time and vehicle operating cost changes calculated in resource costs to market prices. The individual components given in Tables 9A to 9F of the COBA output file are presented under the Transport Economic Efficiency (TEE) categories and converted to market prices by the appropriate tax correction factors. Table 15A, Phase 16 Economic Efficiency of the Road System in Market Prices (TEE Table) Table 15A in the COBA output is an adaptation of the TEE Table. COBA takes the input values for the construction delays and maintenance delay savings (expressed as resource costs), converts to market prices and allocates between consumers and business in proportion to the consumer and business user (Time and VOC) benefits of the scheme under normal operating conditions. Table 6.2 shows how the elements of the TEE Table calculated by COBA are referenced from Table 14 and combined with the delays during construction and maintenance delay savings to produce the Net Consumer User Benefit and Net Business Impact. Table 15B, Phase 16 Public Accounts This Table (see Table 6.3) shows the summary of Public Accounts and summarises the funding of the project. This table fulfils the requirement for an exchequer cash flow analysis. Table 15C, Phase 16 Analysis of Monetised Costs and Benefits Table15C (shown in Table 6.4) of the COBA output summarises the monetised costs and benefits as calculated by COBA. This effectively represents the scheme summary, and is the key output from the CBA assessment.

6.3

6.4

6.7

6.8

6.9

6.10

COBA guidance

44

Interpretation of COBA output 6.11 If all impacts of a scheme could be monetised and included in the CBA, the overall economic worth of a scheme could be summarised using one or more of the following measures: Net Present Value (NPV); Benefit / Cost Ratio (BCR), and Internal Rate of Return (IRR). 6.12 Net Present Value The Net Present Value (NPV) is the difference between the Present Value of Benefits (PVB) and Present Value of Costs (PVC). NPV = PVB-PVC Benefit to Cost Ratio The Benefit to Cost Ratio (BCR) is given by the ratio of the discounted sum of all future benefits to the discounted sum of all costs. Thus: BCR = PVB/PVC Internal Rate of Return The Internal Rate of Return (IRR) is the rate of discount that makes the present value of the benefits exactly equal to the present value of the costs. Put another way, the IRR is the rate of discount that makes the NPV of the entire stream of benefits and costs exactly equal to zero. The IRR is that for which the sum:

6.13

6.14

6.15

(1 + )
y =n y =0

By

y p

=0

Where By is the net benefit (undiscounted) in year n. 6.16 To solve for an iterative approach should be taken whereby incremental changes are made to the discount rate entered by the user into the basic data section of the COBA file and re-running the program until a value is found that yields a NPV equal to zero. It should be noted that there may also be other significant costs and benefits, some of which cannot be presented in monetised form. In such cases, whilst a value for the NPV and BCR can be calculated the results may be misleading and will not necessarily provide a reliable measure of overall value for money. Under such circumstances, the analysis should not be used as the sole basis for decisions. Sensitivity testing 6.18 No COBA result is exact: a risk exists that project costs and benefits might deviate from their expected values. Any investment decision, whether public or private sector, is bound to be subject to uncertainty. Decisions regarding long-lived investments with distant forecast horizons, such as national road proposals, are subject to a high degree of uncertainty. However, it is important for decision makers to have some idea about how robust the results may be in order to know what weight to attach to them. It is, therefore, necessary to consider a range of possible outcomes. One approach is to set bounds on the uncertainty by carrying out tests on key variables to identify those variables to which the results are particularly sensitive. These are the variables on which the decision makers' judgment should focus. Local variables At preliminary design, locally derived parameter values should be used where local data is both reliable and significantly different from national values (as discussed in Section 4). However, it will often be useful for the user to carry out sensitivity tests on these variables, especially where they are both uncertain in the local context and likely to affect the COBA result significantly.

6.17

6.19

6.20

COBA guidance

45

6.21

It is recommended that the sensitivity test should use the national values as a benchmark. This is to ascertain the importance of local variations and to allow comparison of schemes on a similar basis. Forecasting inputs Errors introduced by forecasting inputs are known to have a significant impact on NPV results, in particular the impact of Gross Domestic Product and fuel price assumptions on traffic forecasts and the associated values of time, accidents and vehicle operating costs used in COBA. It is therefore necessary to present COBA results based on both high and low traffic growth. This should be done for all CBAs. Scheme costs There is a great deal of uncertainty concerning the final capital costs of road schemes. This degree of uncertainty tends to reduce at later stages of the projects development, when outturn costs can be estimated with more confidence. Sensitivity tests should, therefore, be undertaken to assess the impact of changes in construction costs and land and property costs on the overall NPV of a scheme (see Project Appraisal Guidelines, Chapter 5 for more detail). In particular the testing should look at the impact of increased construction costs and the level of increase required to reduce the viability of a scheme. Other considerations It may well be the case that a sensitivity test highlights variation in the NPV, but that the sign of the NPV and ranking of options remains unchanged. This would clearly increase the weight that can be put on the economic results.

6.22

6.23

6.24

COBA guidance

46

Table 6.1 Conversion of travel costs to market prices by vehicle category (Table 14, Phase 16 of the COBA Output)
From Table VEHICLE CATEGORY WORK Personal Travel Car Private LGV TOTAL Adjustment MARKET PRICE PSV Adjustment MARKET PRICE 9B 9C 9D Freight Freight LGV OGV1 OGV2 TOTAL Adjustment MARKET PRICE Private Sector Operating Costs Adjustment MARKET PRICE TIME COMM/OTHER TOTAL TIME OPERATING FUEL WORK COMM/ OTHER OPERATING OTHER WORK COMM/ OTHER TOTAL OPER. COSTS

9A 9B

t = AA t = GG t = BB t = HH (row) (row) =
see note vi

= CC -

tF = DD

t = EE

tN = FF

(row)

9E

t = II

(row)

see note vi

= JJ

t = KK

(row)

9E

t =LL

(row)

see note vi

= MM

t = NN

(row)

Totals

THIS ANALYSIS IS BASED ON TRAFFIC GROWTH AND ECONOMIC GROWTH COSTS ARE IN ( PV Year ) PRICES IN MULTIPLES OF A THOUSAND EURO AND ARE DISCOUNTED TO (PV Year) EVALUATION PERIOD (evaluation period ) YEARS FIRST SCHEME YEAR CURRENT YEAR DISCOUNT RATE (discount rate) PERCENT FOR (years) YEARS THEREAFTER (discount rate) for (years) THEREAFTER (discount rate) PERCENT

Notes: (i) (ii) (iii)

(iv) (v) (vi)

t is the average rate of indirect tax on final consumption in the economy; tF is the rate of indirect tax on fuel as a final consumption good; tF is the rate of indirect tax on fuel as an intermediate consumption good; tN is the rate of indirect tax on non-fuel vehicle operating costs as a final consumption good; tN is the rate of indirect tax on non-fuel vehicle operating costs as an intermediate consumption good; The adjustment for Work Operating Fuel is: Total x ( 1 + t N ) ( 1 + t ) - Total

COBA guidance

47

Table 6.2 Economic efficiency of the road system in market prices (Table 15A, Phase 16 of the COBA Output)
IMPACT TABLE REF TOTAL CARS & PRIVATE LGVs GOODS VEHS & BUS. LGVs BUS AND COACH

CONSUMER USER BENEFITS User Benefits Travel Time Vehicle operating costs Travel time and vehicle operating costs During construction During maintenance See note (ii) See note (ii) NET CONSUMER USER BENEFITS BUSINESS USERS User Benefits Travel Time Vehicle operating costs Travel time and vehicle operating costs During construction During maintenance Subtotal Private Sector Provider Impacts Operating Costs Other Business Impacts Developer and Other Contributions (4) = (row) (3) (2) See note (ii) See note (ii) MM + NN = (row) = (row) AA CC + EE II JJ + KK GG + LL (1) = (column) = (row) = (row) BB DD + FF HH -

NET BUSINESS IMPACT TOTAL Present Value of Transport Economic Efficiency Benefits

(5) = (2)+ (3)+ (4) (6) = (1) + (5)

THIS ANALYSIS IS BASED ON TRAFFIC GROWTH AND ECONOMIC GROWTH COSTS ARE IN ( PV Year ) PRICES IN MULTIPLES OF A THOUSAND EURO AND ARE DISCOUNTED TO (PV Year) EVALUATION PERIOD (evaluation period ) YEARS FIRST SCHEME YEAR CURRENT YEAR DISCOUNT RATE (discount rate) PERCENT FOR (years) YEARS THEREAFTER (discount rate) for (years) THEREAFTER (discount rate) PERCENT

Notes: (i) (ii)

Items AA to NN referenced on Table 11/1 COBA takes the input values for the Construction Delays and Maintenance delay Savings, converts to market prices (multiplies by 1 + t ) and allocates between Consumers and Business in proportion to the Consumer and business User (Time and VOC) benefits of the scheme under normal operating conditions.

COBA guidance

48

Table 6.3 Public accounts (table 15B, phase 16 of the COBA output)
IMPACT LOCAL GOVERNMENT FUNDING Operating Costs Investment Costs Developer and Other Contributions NET IMPACT CENTRAL GOVERNMENT FUNDING Operating Costs Investment Costs Developer and Other Contributions Indirect Tax Revenues NET IMPACT Present Value of Costs (PVC) (8) (9) (1+ t) (Maintenance Expenditure Savings) (1 + t) (Scheme Costs less Maintenance Capital Cost Savings) (7) (1+ t) (Maintenance Expenditure Savings) (1 + t) (Scheme Costs less Maintenance Capital Cost Savings) TABLE REF TOTALS

THIS ANALYSIS IS BASED ON TRAFFIC GROWTH AND ECONOMIC GROWTH COSTS ARE IN ( PV Year ) PRICES IN MULTIPLES OF A THOUSAND EURO AND ARE DISCOUNTED TO (PV Year) EVALUATION PERIOD (evaluation period ) YEARS FIRST SCHEME YEAR CURRENT YEAR DISCOUNT RATE (discount rate) PERCENT FOR (years) YEARS THEREAFTER (discount rate) for (years) THEREAFTER (discount rate) PERCENT

Notes: (i) (ii) (iii)

Because tax rates change over time the values entered in this table cannot necessarily be derived from a simple factoring of the values in Table 11/1; Costs entered into COBA are in Resource Cost terms; t is the average rate of indirect taxation in the economy

COBA guidance

49

Table 6.4 Analysis of monetised costs and benefits (table 15C, phase 16 of the COBA output)
IMPACT TEE Benefits Consumer User Benefits Business Benefits Private Sector Provider Impacts Accident Benefits Emission Benefits Present Value of Benefits (PVB) Government Funding Present value of Costs (PVC) Overall Impact Net Present Value (NPV) (11) (9) (11) / (9) Benefit to Cost Ratio (BCR) (9) (1) (2) (3) (10) (10A) (11) (1) + (2) + (3) + (10) + (10A) TABLE REF TOTALS

THIS ANALYSIS IS BASED ON TRAFFIC GROWTH AND ECONOMIC GROWTH COSTS ARE IN ( PV Year ) PRICES IN MULTIPLES OF A THOUSAND EURO AND ARE DISCOUNTED TO (PV Year) EVALUATION PERIOD (evaluation period) YEARS FIRST SCHEME YEAR CURRENT YEAR DISCOUNT RATE (discount rate) PERCENT FOR (years) YEARS THEREAFTER (discount rate) for (years) THEREAFTER (discount rate) PERCENT

Note: There may also be other significant costs and benefits, some of which cannot be presented in monetised form. Where this is the case, the analysis presented above does NOT provide a good measure of value for money and should not be used as the sole basis for decisions

CBA report

COBA guidance

51

CBA report

General 7.1 Having carried out a cost benefit analysis, the CBA Team is responsible for producing a formal CBA Report for submission to the DOPM. The purpose of the report is to detail and justify the methodology, provide detailed information on the data inputs and to present the results of the economic appraisal. The CBA Report is the primary output from the CBA process, and will contain all the information required by both the NRA and the DOPM. This includes: the validation checklist; node-link diagrams; maps; COBA *.dat and *.prn files for all different scenarios / sensitivity tests undertaken; and any other information requested by the NRA or DOPM. When is a CBA report required? 7.3 7.4 A CBA Report will normally be required at route selection, preliminary design, construction documents preparation / tender award and final account / closeout. The level of detail included in the report may reflect the phase at which it is prepared. For example, if default parameter values have been used during route selection, Section 5 of the report (as detailed below) will only need to state that default values have been used, with reference to the National Parameter Value Sheet. Conversely, a report produced at preliminary design, will have to contain full documentation on the derivation of local parameter values. At final account / closeout the report should compare the results with the preliminary design COBA results. Report contents 7.5 The CBA appraisal report shall contain a main report detailing the methodology and assumptions of the CBA process, and a set of appendices containing supporting information. The main report shall be structured as follows: Section 1 Introduction 7.6 An introductory section should set the scene for the report, identifying at what stage the assessment is being undertaken (route selection, preliminary design or final account / closeout) and broadly describing the nature of the scheme(s) being appraised. A map of the scheme should be included. Section 2 Software specification 7.7 The report should confirm the version of the CBA software that has been used in the assessment, and the version of the National Parameter Values Sheet. Section 3 COBA network 7.8 The rationale behind the extent of the area modelled in COBA will be set out. The schematic diagram of the COBA node-link diagram should also be presented. The COBA node-link diagram will clearly illustrate the following data: Network layout and road names; Node numbers: nodes should be numbered starting with node number 101 and increasing in increments of 1, e.g. 101, 102, 103 etc. Where schemes are separated into specific sections or regions the first digit can be changed to distinguish each region clearly within the COBA output file, e.g. 201, 202, 203, 301, 302, 303; Link numbers: link numbers should be four digits in length and broadly relate to each node number into which they run, i.e. links 2011, 2012, 2013, and 2014 will run into node 201.

7.2

COBA guidance

52

Where 2-way links are coded, link numbers can be defined by running into either node at the end of the link. For example, a link joining nodes 201 and 202 can be defined by either 2011 or 2021; Do-Minimum and Do-Something traffic flows; and Road type of each link. 7.9 The required format for the network diagrams is provided as an example in Figure 7.1. The justification for a standard coding methodology is to facilitate easier interpretation of COBA output data files by the Design Office and the NRA. A separate diagram showing Do-Minimum and Do-Something flows is also required. Section 4 Data collection 7.10 Details on the level of data collection that was undertaken should be outlined. This would cover traffic flow information and journey time surveys. The compilation, checking and interpretation of this data should also be described. Section 5 CBA input assumptions 7.11 All data input assumptions should be described, with specific reference to the treatment of the CBA parameters. Significantly, where NRA default values have not been used in the assessment, details documenting the derivation of these values should be provided. Information relating to the calculation of the scheme costs is also required, describing the approach taken in the derivation of the costs. Completed spreadsheets as provided in Appendix 13 and 14 should be included. Section 6 CBA validation 7.13 The process of validation of the COBA model should be presented, which examines in particular the base year journey times. Any significant deviation from journey times between modelled and surveyed data sets should be explained and justified. Section 7 CBA results 7.14 The results section will contain the summary of results of the assessment. This will provide a brief overview of the economic case for the road scheme. This section should also include reports of any sensitivity tests that may be necessary and an incremental analysis of link and junction standards if required. The report should contain a separate section on the Impact on the Public Accounts summary table (shown in Table 6.3), which fulfils the requirements of the exchequer cash flow analysis. For a final account / closeout CBA, section 7 should also contain a discussion of how the results compare with the construction documents CBA. Appendices 7.17 The appendices provide all the supporting documentation that accompanies the CBA appraisal report. The appendices should include: Cost estimate summary (e.g. Total Scheme Budget and Target Cost); The COBA input files; The COBA output summaries; A copy of the COBA schematic network diagrams; A summary of any relevant data collected to validate the COBA model, and A CD containing all the above information in digital form, including a full copy of the COBA / TUBA output file and the scheme drawings used to derive the geometrical data. Information on traffic flow inputs 7.18 Given the importance of traffic flow input to CBA, the basis of this input should normally be documented. This information is contained within the Traffic Modelling Report.

7.12

7.15 7.16

COBA guidance

53

Figure 7.1 Example COBA network diagram

KEY
1011 101 102 1021 103 1031 104 1041 113 1131 105 1052 106 1051 101 1011 Single Carriageway 2+1 Road Dual Carriageway Motorway Node Node Number Link Number

1061

107 1072 1083 117 1082 1141 1144 122 114 108 1081 116

1071 115

120

1142 121 109 1091 118

1143 1093 119

1092 1121 110 124 1103 1102 1101 123

111

1122

112

Auditing and checking

COBA guidance

55

Auditing and checking

General 8.1 Auditing and checking of the COBA output is an essential element of the cost-benefit analysis process and due consideration, in terms of resources allocated to the task and the time allowed, should be given in the project management of the scheme. The COBA output provides a detailed analysis of the scheme appraisal. Within the program a wide range of data and consistency checks are carried out. However it is not possible for the program to detect all possible errors. It is, therefore, important that the output be reviewed to determine whether or not the appraisal is robust. Responsibilities 8.3 The responsibility for undertaking the auditing and checking occurs at three levels: Firstly, the Appraisal Team Leader undertaking the COBA analysis should satisfy him/herself that the assessment has been carried out in a robust manner prior to submitting the appraisal report to the DOPM; Secondly, the DOPM or his/her team should undertake a review of the work themselves, and Lastly, the NRA will conduct an external review of each COBA produced. 8.4 8.5 CBA Team and Design Office audits should take place at each project phase that CBA is required. Persons involved in the appraisal process must have had some prior experience or attended an approved formal training course on either project appraisal or CBA. This holds true for both members of the Appraisal Team and for the person responsible for reviewing the appraisal work in the Design Offices. The Appraisal Team comprises an Appraisal Team Leader and Appraisal Team Members. The required experience within the Appraisal Team and the Design Office is defined below. Note that the requirements for the Appraisal Team are relevant regardless of whether the team is drawn from the Design Office, or employed as an external consultant. The Design Office Project Manager is responsible for the approval of the CBA Team. Appraisal Team Leader: The Appraisal Team Leader must be able to demonstrate knowledge of the appraisal process, including the use of standard computer programmes used in traffic modelling and economic assessment. The Team Leader must have been on an approved training course and have detailed experience of appraisal on at least three highway projects. Appraisal Team Member: The Appraisal Team member must have been on an approved training course or seminar and/or have practical experience in undertaking appraisal. At least one Appraisal Team member should be included on each Appraisal Team. Key audit areas 8.9 DMRB Volume 13, Section 1, Part 3, Chapter 11 provides guidance on what to check when validating a COBA printout; reference should be made to this document when undertaking the validation. In the first instance, there is a requirement for the CBA Team to provide several key pieces of information, including scaled scheme plans, node and link diagrams, cost estimates and COBA printouts. The COBA printouts can be lengthy, so electronic copies of the input and output files are required enabling all data to be viewed on screen.

8.2

8.6

8.7

8.8

8.10

COBA guidance

56

8.11

The key areas that require checking are described below. Where modelled networks are large, a sampling approach to the checking may be adopted. For example, for networks consisting of over 100 links, it would be appropriate to check in detail the coding of 10% of the links and junctions. Basic variables If local rather than default values have been used for parameters such as traffic growth, traffic proportions, seasonality index, E- and M-factors, flow groups and accidents, then the basis for the derivation of these values should be examined and any calculations confirmed. Coded network structure There should be a direct correspondence between the coded network, the node link diagram and scheme plans for both the Do-Minimum and Do-Something schemes. Link data A check should be made on the data pertaining to each link, such as the link length and other geometric characteristics, speed and accident rates, with those values falling outside the range of typical values for the link type being investigated further. Junction data The audit should check that flow patterns and turning matrices are reasonable. Maximum delay coding can be compared against timed runs. Traffic flows Flows entered for non-neutral months should be queried and the flows should be crossreferenced against those contained in any Local Model Validation Report. There should be no links with zero flow. Scheme costs The scheme costs input into COBA should be checked carefully. The audit should confirm that all items (construction, land, preparation and supervision costs) have been included and the cost estimates input are in accordance with the most recent estimates agreed with the NRA Cost Estimation Unit. The auditor should check whether the appropriate scenario testing / sensitivity analysis has been undertaken on costs as well as in terms of traffic growth forecasts (see Chapter 5 of the Project Appraisal Guidelines for more detail). Any indices used to take historical costs to the price base year, such as the CPI and RPF should be reviewed for appropriateness. The Auditor should also consider whether the assumed timing of the scheme costs is correct. Costs and benefits The items listed above all relate to data input by the user and are relatively straightforward to check. A more complex task is the interpretation of the costs (and thereby benefits) that these data give rise to in the context of travel time, vehicle operating costs and accidents, and the changes thereof over the appraisal period. The process should seek to confirm that the cost and benefit stream relating to each is realistic and also in line with the aims of the proposed scheme. For example: Travel time the audit should investigate links and junctions going over capacity and when this occurs during the appraisal period. The audit should also confirm whether changes in total vehicle kilometres accord with expectation, and should examine the resulting changes to time spent on each link and at each junction. Vehicle Operating Costs fuel and non-fuel related Vehicle Operating Costs will change as a result of the general increase in vehicle kilometres aggregated over the entire network over time. The auditor should confirm that the changes in vehicle operating costs accord with expectation. Accidents are large accident reductions realistic? Can they be reconciled with the problems for which the scheme is designed to address?

8.12

8.13

8.14

8.15

8.16

8.17

8.18

8.19 8.20 8.21

8.22

COBA guidance

57

Audit checklist 8.23 A checklist that summarises the key elements to be considered during the audit process is included in the summary checklist for validation of COBA printout, is contained in appendix 2. This list is based on the information contained in DMRB Volume 13, Part 1, Chapter 11, and includes the more important items in each output phase that should be reviewed and commented upon if necessary. It is recommended that auditors use this list, or similar, to ensure that all costs and benefits attributed to a scheme have been checked. If the audit is being undertaken by an independent consultant, any findings, comments and recommendations may be expanded upon in a separate audit report, with the summary checklist.

8.24

Final account / closeout

COBA guidance

59

Final account / closeout

General 9.1 The undertaking of a final account / closeout CBA as part of a post project review is a useful exercise that should be scheduled approximately 12 months after the implementation of the scheme. This should provide for enough time to enable traffic patterns to settle and to get a reasonably firm indication on final outturn costs. The main objective of the review is to draw lessons for the future, with the intention of improving the accuracy of subsequent assessments. The Department of Finances Guidelines for the Appraisal and Management of Capital Expenditure Proposals in the Public Sector define two main areas on which post project reviews should be focused, namely: Project outturn, and Appraisal and management procedures. 9.3 In the context of Cost Benefit Analysis, it is those elements affecting the project outturn that are of most interest. The final account / closeout review should, therefore, concentrate on establishing whether the bases on which the costs and benefits attributed towards the scheme were correct and drawing conclusions on whether the expected benefits are likely to materialise. The review of appraisal and management procedures relate to the Post Project Review and are dealt with in Chapter 8 of the NRAs Project Appraisal Guidelines. Responsibilities 9.5 An Appraisal Team within the Design Office or an independent consultant acting on behalf of the sponsoring agent will be responsible for undertaking the final account / closeout CBA review. Timing 9.6 The Post Project Review should be undertaken once sufficient time has elapsed to allow the project to be properly evaluated. A minimum period of approximately 12 months should elapse before traffic and journey time data are collected, to allow traffic patterns and network conditions to settle. Prior to undertaking the review final outturn scheme costs must be established. After 12 months the majority of the scheme costs should be known to allow a CBA to be undertaken. If a substantial proportion of the costs are still outstanding then the undertaking of a CBA can be postponed with the consent of the NRA. Costs to be used for the appraisal of Public Private Partnership schemes should be agreed with the NRA. Data requirements 9.8 The final account / closeout CBA is a data intensive exercise and should only be undertaken for a sample of schemes. The DOPM should contact the NRA at this phase to ascertain whether a final account / closeout CBA is required for their particular scheme. To undertake the review it will be necessary to collect information relating to: Scheme costs The actual costs and when they occurred in delivering the scheme can be obtained and compared against the estimates and time schedules entered into the COBA program.

9.2

9.4

9.7

9.9

COBA guidance

60

9.10

Link flows Traffic flows on the scheme can be collected and compared against those obtained in the traffic model and used as input to the COBA program. Journey times Journey times on the scheme can be collected and compared against the COBA output. Accident data When assessing accident rates it is common practice to use accident records over an extended period (3 to 5 years) to reduce the impact of one off events on overall trends. For this reason an assessment using observed accident data from the completed scheme cannot be included in the post-opening assessment. COBA assessment

9.11 9.12

9.13 9.14

The key inputs for a COBA assessment at final account / closeout are: Scheme costs The actual costs and when they occurred in delivering the scheme should be entered into the COBA program, using KEY055, which requires all costs to be net of indirect taxation and at the price base year. Appendix 16 contains an Excel file which converts from actual outcome costs to a format suitable for entry into COBA (i.e. in factor costs to a given base year). Traffic data For a simple COBA network, it would be possible to undertake traffic counts on all links in the network and input these into a new COBA deck. This would include both counts on the scheme and on links in the rest of the network, particularly where flows were forecast to change significantly. For more complicated networks it is too onerous to collect traffic count data for every link in the network. In these instances it may be necessary to revalidate the traffic model to post-opening conditions using a programme of traffic counts. The output from the traffic model could then be input as data for the final account / closeout COBA. Comparison with construction documents preparation / tender award COBA

9.15

9.16

9.17

A comparison should be undertaken with both the route selection and preliminary design COBA outputs. The comparison should look at the overall economic indicators including PVB, PVC, NPV, BCR and IRR. A summary table should be produced showing the values, at constant prices, for these indicators at each phase. Detailed reporting requirements are outlined in Section 7 of this document.

COBA guidance

61

10 References

1 0

COBA guidance

63

10

References
Transport Research Laboratory: COBA11:Cost Benefit Analysis Program. Department of the Environment, Transport and the Regions: Guidance on Methodology for Multi-Modal Studies: May 2000. Goodbody Economic Consultants in association with Atkins: Cost Benefit Parameters and Application Rules for Transport Project Appraisal: August 2004. Goodbody Economic Consultants in association with Atkins: Parameter Values for Use in CostBenefit Analysis of Transport Projects: September 2004. Department of Finance: Guidelines for the Appraisal and Management of Capital Expenditure Proposals in the Public Sector: July 1994. Highways Agency: Design Manual for Roads and Bridges, Volume 12: Traffic Appraisal ofRoad Schemes: Section 1 Traffic Appraisal Manual: HMSO 1997. Highways Agency: Design Manual for Roads and Bridges, Volume 13: Economic Assessment of Road Schemes The COBA Manual: TSO 2002. National Roads Authority Design Manual for Roads and Bridges: NRA TD 27/00 CrossSections and Headroom: December 2000. National Roads Authority Interim Advice Notes for 2+1 Roads and Type 2 Dual 2 Lane Carriageways. National Roads Authority: Future Year Traffic Forecasts 2002-2040: January 2004. National Roads Authority: National Road Needs Study: July 1998. National Roads Authority: National Roads Project Management Guidelines: March 2000. CSF Evaluation Unit: Proposed Working Rules for Cost Benefit Analysis: June 1999. Department of Transport: Report of the Working Group on the Review of Speed Limits: September 2003. Department of Transport: Road Safety Strategy 2004-2006. European Communities: Saving 20,000 Lives on our Roads A Shared Responsibility: 2003.

11 Enquiries

1 1

COBA Guidance

65

11

Enquiries

All enquiries should be directed to: Transport Economist National Roads Authority St Martins House Waterloo Road Dublin 4.