5. Review of Laurencekirk STAG Appraisal

5. Review of Laurencekirk STAG Appraisal

Introduction 

5.1 In addition to the wider outcome evaluation set out Chapter 4, an important element of the study scope was to review the economic elements of the original STAG appraisal, with a view to understanding:

  • why the passenger and revenue figures for Laurencekirk station exceeded those forecast in the STAG appraisal;
  • Revised STAG, including Transport Economic Efficiency / Benefit Cost Ratio (BCR) for the scheme; 
  • the ‘Wider Economic Benefits’ (WEBs) of the project; and
  • Wider Impacts of station reopening.

Review of STAG

5.2 Evidence from a range of new rail line and rail station openings demonstrates a repeated weakness in passenger demand and revenue forecasting.  This problem is prevalent across different geographies and station types – forecasts often significantly under or over-estimate demand with no clear reason as to why.  Laurencekirk is an example of where the patronage forecasts underestimated outturn demand as will be detailed below.  

5.3 This section comprises a brief review of the original STAG appraisal of August 2004 and the focus is on the quantification of costs and benefits with a view to comparing these with the outturn values derived in part from the survey of Laurencekirk station users.  

5.4 The purpose is also to understand the methodology employed in generating the original passenger and revenue forecasts with a view to explaining why passenger numbers were under forecast.  It is important to note however that the  guidance on transport appraisal has also been updated since the original STAG work of 2004 and parts of the methodology were not necessarily inconsistent with the guidance at the time.

Laurencekirk Station Train Service Appraised

5.5 The train services envisaged in the STAG was as follows:

  • 15 northbound station calls (i.e. Dundee–Laurencekirk–Aberdeen); and 
  • 14 southbound calls (i.e. Aberdeen–Laurencekirk–Dundee) per day.

5.6 These services were specified to be a mix of Glasgow-Aberdeen and Edinburgh-Aberdeen services, based on the timetables in operation at the time.  No other station stops were dropped to accommodate the additional stop at Laurencekirk[9].  

5.7 On a weekday, there are currently 12 northbound and southbound trains calling at Laurencekirk, so the outturn service is somewhat less than originally envisaged in terms of frequency.  In addition, there are only two direct trains per day between Laurencekirk and Glasgow with the other trains all being Edinburgh (or Dundee) services.  Indeed several respondents to the survey did note that more direct links to Glasgow would be beneficial.

5.8 The outturn service is therefore slightly less frequent to that considered in the appraisal.  Note though that STAG appraisal does not appear to have been linked to specific timetables / train times in the way that, for example a network modelling exercise would be.

Demand and Revenue Forecasting

5.9 The 2004 STAG report does not lay out the calculations underlying the demand forecasting in any detail.  However, the process is described as follows:

  • ‘Identify existing patterns of trip movements in each of the study corridors;
  • Identify existing origins and destinations of trips in each of the corridors;
  • Identify the modes used by the existing trips in each of the corridors;
  • Evaluate the numbers of car trips transferred from road to rail when the station is opened; and
  • Calculate the effects of these modal transfers.’

5.10 The report outlines that TEMPRO-based trips rates (TEMPRO v4.2, DfT, 2002) for 2001 were extracted for the Laurencekirk area, resulting in a trip rate of 5.5 trips per household per day (combined car driver and passenger, not stated whether these are one-way or return trips).  It is then stated that ‘in some respects, the application of aggregate forecasts to a specific rural area may be open to question, given the impact of location and relative distance to prime facilities and services’, so an assumed figure of 2.5 trips per household per day is used, cited as a conservative estimate.

5.11 From these trip rates, an estimated patronage of 36,000 one-way rail trip is derived which at an assumed £7.20 per trip provides a revenue estimate of £259,000 revenue (2003 prices).  The figure of 36,000 is significantly lower than the outturn passenger numbers discussed in chapter 4, which showed almost 75,000 passengers using the station in the first full year of 2010-11, rising to over 92,000 by 2012-13.  

5.12 No further details are provided as to the derivation of the forecasts.  It could be assumed that the trip rate of 2.5 trips per household has been applied to a catchment area of households.  An estimate of how many of these trips go to Dundee and Aberdeen could have been made and then some proportion of these taken as rail trips.  This raises a number of questions:

  • are all new rail trips assumed to have transferred from car?;
  • are all new rail trips from Laurencekirk new to rail, i.e. no transfer (via Park and Ride) from existing stations?;
  • are any entirely new trips generated?;
  • has any change in destination choice occurred?; and
  • is there an assumed loss of revenue to other modes (eg bus)?

5.13 The trip volumes are used to estimate the number of car km removed from the network, and a figure of 4.3 million vehicle kilometres per annum is reported.  As a result of these reduced traffic levels, accident savings (£69,700 per annum) and vehicle operating cost savings (£139,600 per annum) are also reported.

5.14 MOIRA was used to estimate the revenue impacts on existing services of stopping trains at Laurencekirk, and it was also used to estimate monetised delays to remaining passengers.

5.15 Using the above results, the total revenue and benefits per annum used the appraisal are as follows:

  • Revenue : +£259,000 per annum (at 2003 prices);
  • Accidents savings (accidents prevented due to reduced traffic levels): +£69,700 per annum (at 2003 prices);
  • Vehicle Operating Cost (VOC) savings: + £139,600 per annum (at 2003 prices);
  • Abstraction (reduced rail boardings due to increased journey times associated with the additional station stop): –£26,000 per annum (at 2003 prices); and
  • Delay (monetised time delays to remaining passengers due to increased journey times associated with the additional station stop): –£95,000 per annum (at 2003 prices).

5.16 This would suggest a net positive annual benefit of £347,300 per annum at 2003 prices.

5.17 The MVA Report of 2003 is included as an Appendix to the main STAG report.  It includes an estimated figure of 25,000 single journeys originating in Laurencekirk and a similar number of returns, i.e. a total of 50,000 journeys.  A total annual revenue of £240,000 is also forecast.  It is notable that both forecast revenue figures (£240,000 by MVA and £259,000 by Scott Wilson) are similar, yet these are derived from very different patronage forecasts (50,000 for MVA and 36,000 from SW).

5.18 The STAG Report notes that ‘however, there are several aspects of the methodology used by MVA that are worthy of highlighting as they have led to quite low estimates of demand and revenues’.  However the patronage forecast used by SW is actually lower than that developed by MVA, albeit the revenue forecasts are similar.  This suggests there may have been some misunderstanding in the interpretation of these figures between SW and MVA, but it is not possible to fully understand this apparent discrepancy with the information available in the STAG report.

5.19 This review of the benefits cited in the original STAG report suggests:

  • A fairly limited range of benefits were estimated, and these were primarily financial;
  • Most notably there was no estimate of travel time savings;
  • A very simple assumption of transfer of existing car trips to rail underlay the analysis.  No other aspects of behavioural change appear to have been considered;
  • The appraisal period, price base and discount rates have been updated since the appraisal was undertaken; and 
  • The key underlying assumptions and calculations underlying the demand forecasts in particular need to be more clearly laid out.  As it stands, it is not possible to follow a full trail of logic through to the derivation of the key figures.

5.20 The STAG Report does not therefore contain sufficient detailed information to allow us to determine precise reasons why forecast passenger numbers were lower than seen in the outturn.  

5.21 The approach taken in the STAG appears to have been fairly high level, and not based on, for example, specific population projections or planning data assumptions for Laurencekirk which may not have materialised.  As such, the previous approach cannot simply be re-run with new data to see how this would have affected the forecast.  

5.22 The key conclusion here is that the key underlying cause and effect assumptions and analysis must be recorded and laid out clearly in a STAG Report (or supporting annex) if a later evaluation process is to understand why outcomes may have diverged from forecast.

Outturn values

5.23 As outlined above, the Laurencekirk STAG study estimated economic benefits associated with:

  • Increased rail revenue (£357,000 per annum, 2013 prices[10]), based on assumed yield of £9.93;
  • Reduced vehicle operating costs (£193,000 per annum, 2013 prices); and
  • Reduced costs associated with road accidents savings (£96,000 per annum, 2013 prices).

5.24 Outturn values for each of these are estimated below, based on LENNON and the Laurencekirk user survey data.  

5.25 The STAG also estimated disbenefits associated with:

  • Delay to existing passengers due to the additional station stop at Laurencekirk (£131,000, 2013 prices); and
  • Loss of existing passengers due to the additional station stop at Laurencekirk (£36,000, 2013 prices). 

5.26 The STAG states that the time penalty associated with the extra station stop is three minutes, and it is assumed here that the above values are derived from this three minute penalty.  It is assumed that those trains which do not stop at Laurencekirk incur no time penalty.  In order to assess whether this timetable impact was borne out, historic MOIRA files have been examined for December 2008 (i.e. pre Laurencekirk) and December 2010 (post Laurencekirk opening).  

5.27 Figure 18 below shows the average timetable journey time across the day between Dundee and Aberdeen for 2008 and 2010.  The results are shown for all trains, then Aberdeen-Glasgow services and Aberdeen-Edinburgh services separately.  

5.28 It also shows the results separately in each case for trains which stop at Laurencekirk and for trains which do not stop at Laurencekirk.

Figure 18. Before and After Laurencekirk – Dundee-Aberdeen Travel Times (hours)

Figure 18.	Before and After Laurencekirk – Dundee-Aberdeen Travel Times (hours)

5.29 These figures therefore suggest that overall, there was little change in average journey times between 2008 and 2010 for all rail travel between Dundee and Aberdeen following the opening of Laurencekirk station (i.e. comparing ‘2008’ with ‘2010 average’), despite around one third of trains now stopping at Laurencekirk.  

5.30 Trains stopping at Laurencekirk see journey times increase by around five minutes, greater than the value assumed in the STAG.  However, other trains see faster journey times.  This suggests that following the opening of Laurencekirk, or around this time, there was a wider reorganisation of stopping patterns on the line to create ‘stopper’ trains (slower than 2008) and ‘express’ trains with fewer stops (i.e. faster than 2008).  Indeed the May 2014 timetable sees between two and six station stops between Aberdeen and Dundee, which supports this.  It should be acknowledged however that these changes may or may not have been solely attributable to Laurencekirk station coming into operation.

5.31 Though stopping in itself would obviously add a minute or two to journey times, this apparent wider reorganisation means it is impossible to isolate the outturn timetable impacts associated solely with Laurencekirk station.  However, using this analysis, average journey times between Dundee and Aberdeen increased by only eight seconds or 0.2% between 2008 and 2010, so overall the impact can be seen as negligible, when seen in the context of these wider timetable changes.  

Rail Revenue

5.32 LENNON data for the 2012/13 financial year reports a total of 92,500 single journeys and £571,000 of revenue associated with rail travel via Laurencekirk station, an average yield (i.e. total revenue divided by total journeys) of £6.70.  

5.33 The outturn patronage and revenue is therefore far higher than forecast although yield is significantly lower than envisaged.

5.34 However, analysis of the survey returns suggests that 18% of this total revenue figure has been displaced from other stations.  Hence the net figure for additional revenue is £469,000.  The STAG Report makes no reference to net / gross revenue in this way, but this net figure has been used in the benefit cost calculations which follow.

Change in Vehicle Kilometres: Vehicle Operating Costs & Safety

5.35 Both these sets of benefits (vehicle operating costs and safety) are directly aligned with the quantum of vehicle kilometres removed from the network as a result of the new station.  Overall, around 70% of rail trips made via Laurencekirk have come about through a switch from car-based travel, either to another station (15%, primarily Montrose and Stonehaven) or all the way to the final destination (55%).  

5.36 The resulting lower traffic levels will clearly lead to reduced vehicle operating costs and will also have led to a reduction in the number of accidents (as discussed in Chapter 4), primarily with respect to the A90.  

5.37 The STAG Report states that the opening of the station would reduce vehicle kilometres travelled on the network by 4.3m per annum, and this is associated with 36,000 rail passenger trips.  This implies that each single rail trip has taken 119km of car travel off the network.  This value does seem very high given that the road distance from Laurencekirk to Aberdeen (which accounts for around 75% of all trips) is only 50km.  In addition, this high figure does not align with previous MVA analysis undertaken in 2003 (and included in the STAG Report as an appendix) where very low rates of modal shift from car to rail are forecast (also in contrast to the outturn findings).  

5.38 The annual level of car kilometres associated with all of the trips recorded in the customer survey undertaken as part of this research has been estimated using details of the journeys’ origins and destinations, and also the frequency of travel.  This analysis has been carried out for both:

  • journeys undertaken in the present day using the new station; and
  • the equivalent journeys based on the same journeys undertake either before the station had reopened, or hypothetically if the new station had not reopened.

5.39 This analysis suggests a lower figure of around 2.6m vehicle kilometres per annum has been removed from the network.  Some reasons for this difference are:

  • the original STAG may have assumed that all new Laurencekirk rail journeys replaced end-to-end car journeys, whereas in reality only around half of rail trips have replaced end to end car journeys; and
  • the figures in the STAG analysis do appear high as noted above and the derivation of the figure reported in the STAG is not clear.

5.40 These figures are used to analyse vehicle operating cost and accident savings below.

Vehicle Operating Costs

5.41 The STAG report forecasts a reduction of 4.3m vehicle kilometres and a reduction in vehicle operating costs of £193,000 in 2013 prices.  This suggest a vehicle operating cost of around £0.04 per kilometre.  Current WebTAG guidance suggests a vehicle operating cost of £0.15 (non work) and £0.19 (in work).  Appraisal Guidance has changed significantly since the STAG was undertaken and these figures reflect that, particularly in light of several years of above inflation fuel price rises, although these have been compensated for to some extent by improved vehicle efficiency. 

5.42 Applying these new values to the estimated reduction in vehicle kilometres derived from the survey gives an annual saving of £432,000 in today’s prices.  This figure is higher than that in the STAG despite the lower estimate of reduced car kilometres due to the higher pence per kilometre saving derived from WebTAG.

Accident Benefits

5.43 The STAG report estimated an annual saving in accident costs of £69,700 per annum (2003 prices), sourced from the NESA Manual (DMRB Vol. 15, 2002).  STAG currently recommends that NESA / COBA values are used to determine the quantity, severity and cost of this reduction in accidents, and also references the NESA manual.  When the values in the current NESA manual are applied to the quantum of accident reductions reported in the STAG, the monetary values reported in the STAG are reproduced, i.e. the valuations of accidents by type have not changed since the original STAG was produced.

5.44 It is therefore reasonable to assume that the outturn accident benefits will be lower than those estimated in the STAG in proportion to the lower outturn reductions in vehicle kilometres as shown below:

  • STAG estimate: £96,000 per annum (2013 prices); and
  • Outturn estimate: £96,000 * 2.6m/4.3m = £58,000 per annum.

Benefit Cost Ratio

5.45 Using the above parameters, it is possible to re-calculate the Benefit Cost Ratio (BCR) on the same basis as the original STAG report.  A proportionate, relatively high level, approach has been taken to this task, focussing on the main parameters, and retaining consistency with the original STAG where possible.

Benefits

5.46 The STAG reported a Present Value of Benefits[11] (PVB) value of £5.0m based on 2003 prices, a 30 year appraisal period and a 3.5% discount rate (assumed discounted to 2002).  

Costs

5.47 The STAG Report developed a construction cost figure of £3.2m in 2003 prices for the new station, with this value including an optimism bias of 34%.  The optimism bias figure was adjusted downwards from 44% to 34% as the capital costs provide to the STAG study already had an element of optimism bias / risk adjustment included.  The approach and figures used are in line with the guidance of the day. The original 44% figure is not reflected in the current STAG though, i.e. the optimism bias values have been updated since the original study.  Outturn construction costs for the new station were £3.5m (2009 prices) or £3.0m in 2003 prices, so outturn costs were actually very close to the costs used in the appraisal, if optimism bias is taken into account.

5.48 The STAG also reported annual operating costs of £60,000, comprising station leasing charges, maintenance costs and train calling costs.  We have no new information with respect to these charges so have assumed the £60,000 is correct in the analysis which follows.

BCR Analysis

5.49 Table 6 below shows the Present Value of Benefits , Present Value of Costs (PVC), Net Present Value (NPV) and Benefit Cost Ratio (BCR) figures as reported in the STAG (Table 8-2), and as recalculated here on the basis of the analysis described above.  

5.50 In the ‘core’ test, we have assumed that the timetable disbenefits associated with the new station identified in the STAG are valid.  The exclusion of these disbenefits (as suggested in the outturn analysis) is explored in a sensitivity test.  In all cases, net rather than gross additional rail revenue has been used in these calculations.

5.51 To retain consistency with the STAG, figures here are reported in 2003 prices, and a 30 year appraisal period from the opening year of 2009 has been assumed for the core test.  The wholly new trips have been attributed with half the average benefits where relevant.

5.52 Sensitivity tests are also included here as follows:

  • the appraisal has been reconfigured with a 60 year appraisal period, as recommended by the latest STAG guidance, including the current 3.5% / 3.0% discount rates;
  • the STAG also assumed no growth in benefits over time.  The ‘60 year growth scenario’ below includes an assumed population growth of 0.6% per annum based on GROS projections for Aberdeenshire; and
  • it was noted above that the timetable related disbenefits to through passengers may not have transpired.  This test excludes these disbenefits, i.e. it is assumed that the only costs are associated with construction and maintenance.
Table 6. Benefit Cost Ratio, STAG and Outturn (2003 prices)
STAG report (£m) Core Outturn 30 Year (£m) Outturn 60 year (£m) Outturn 60 year (pop growth) (£m) Outturn 60 year (no timetable disbenefit) (£m)
5.0 8.3 11.4 13.1 15.8
PVC 3.3 3.3 3.6 3.6 3.6
NPV 1.7 5.0 7.8 9.5 12.2
BCR 1.5 2.5 3.2 3.6 4.4

5.53 On a like-for-like basis, the outturn BCR of 2.5 is a significant improvement on the STAG Report value of 1.5.  The BCR is further improved to 3.2 when a 60 year appraisal period is considered, and then improved again to 3.6 if an allowance is made for population growth.  If it is assumed that the timetable related disbenefits to through passengers did not materialise, then a BCR of 4.4 is achieved.

5.54 It can therefore be concluded that the outturn benefit cost ratio of the new station is significantly higher than that reported in the STAG.

Travel Time Savings

5.55 In many transport appraisals, monetised travel time savings typically comprise the majority of the economic benefits.  However, the Laurencekirk STAG report did not report any benefits associated with changes in travel time.  

5.56 To recap, Figure 19 below shows the previous or hypothetical journeys undertaken by long term and new residents of Laurencekirk respectively.  This underlines that the majority of rail trips have transferred from car driver / car passenger or are mixed mode / Park and Ride from Laurencekirk as opposed to another previously used station.

Figure 19. Details of Equivalent Journeys Undertaken prior to Laurencekirk Opening

Figure 19.	Details of Equivalent Journeys Undertaken prior to Laurencekirk Opening

5.57 In the appraisal of public transport improvements, economic benefits materialising through travel time savings are typically derived from:

  • reduced journey times for existing public transport users, e.g. through faster trains or buses;
  • modal shift between bus and rail, e.g. from bus to rail; and
  • modal shift from car to public transport - in practice though, even complex multi-modal models tend to forecast a low modal shift between car and public transport, and this therefore tends to be a lower order impact.

5.58 In the case of Laurencekirk, the opening of the new station brings no benefits to existing rail users (indeed existing rail journeys using Park and Ride are likely to increase in length / duration as users switch from Stonehaven / Montrose to Laurencekirk), and there is a negligible switch from bus to rail as seen above.  Instead there is a mix of:

  • switchers from car end-to-end to car / rail or walk / rail;
  • a change in Park and Ride location from Stonehaven / Montrose to Laurencekirk; and 
  • a significant number of newly generated trips.

5.59 This makes the appraisal more complex in terms of travel time savings.  

5.60 Travel times by car from Laurencekirk to Aberdeen city centre are typically reported as being 40-45 minutes in the peak, figures also found in Aberdeen Subarea Model (ASAM) transport model results.  However, there may be substantial upwards variability around this average figure caused by congestion at key locations such as Bridge of Dee and Anderson Drive.  We do not have definitive survey data with respect to this, but for example, Google Maps suggests there is typically more than one kilometre of northbound stationary queuing traffic on the A90 at the Bridge of Dee in the AM peak.  In practice therefore, those requiring to arrive at a specific time in e.g. Aberdeen or Dyce may require to allow significantly more time for their journey, and this is not readily captured in conventional appraisal.  

5.61 The average time by train from Laurencekirk to Aberdeen is 33 minutes, but to this must be added station access time (walk / drive), wait time, and onward travel to the final destination from Aberdeen railway station.  In appraisal, walk and wait times are also weighted to reflect the fact that people tend to seek to minimise these aspects of any journey, adding further to the perceived journey time.

5.62 In pure time terms, for many of the journeys recorded in the survey, switching from car to train therefore actually increases journey times.  When a fare is added, this also means that the generalised cost of travel via Laurencekirk by rail can be greater than the generalised cost of travelling end-to-end by car.  However, this ‘revealed preference’ travel behaviour does indicate that these individuals must be gaining a benefit, otherwise they would not have switched mode.  In practice, there are a range other factors which account for people choosing rail over car, which may not be captured in a conventional generalised cost such as:

  • car parking cost and availability at destination;
  • household car ownership;
  • ability to work or relax on the train; and
  • relative reliability of train over car on a congested network – rail offers relative certainty of travel time.

5.63 Here, the data obtained from the user survey has been used to estimate travel times for:

  • journeys undertaken in the present day using the new station; and
  • the equivalent journeys undertaken either before the station was reopened or hypothetically had the station not reopened.

5.64 These estimated travel times include:

  • the walk time to station (derived from home postcode);
  • drive / bus time to station (derived from home postcode);
  • transfer penalty from car to train;
  • time on train;
  • walk time from train to final destination;
  • end to end car journey time (estimated from ASAM model, or Google Maps for long distance trips); and
  • assumed walk time from parking to end destination.

5.65 It has been assumed that those walking or making short drives to Laurencekirk would time their journeys such that wait time would be minimal.  Note that if a typical MOIRA derived wait time (derived with respect to what is a relatively infrequent service) was applied, this would make the end-to-end journey time even more uncompetitive with the car option.

5.66 These journey times have been monetised using WebTAG values of time for commute, business and leisure.  Figure 20 below shows the results of this analysis for the pre and post Laurencekirk trips, shown as an annual aggregate total. 

Figure 20. Total Annual Monetised Journey Times by Purpose – Pre & Post Laurencekirk

Figure 20.	Total Annual Monetised Journey Times by Purpose – Pre and Post Laurencekirk

5.67 It can therefore be seen that the new station has had a broadly neutral impact on travel times, using the assumptions outlined here.  Note that these findings are sensitive to assumptions surrounding car travel times, parking costs, and train wait times, so should only be seen as indicative.  

5.68 However this analysis does serve to illustrate that for new stations located some distance from major settlements (i.e. where there is very little existing bus use), travel time savings are perhaps unlikely to form a major component of the economic benefits.  In other words, very few people are using the mode from which the greatest time saving would be derived, i.e. the bus, presumably due to the prohibitive journey times and / or service frequency associated with this mode.  

5.69 The impact of the new station on before and after journey times has been broadly neutral.

Ideas for Further Research

5.70 There are many examples around Scotland where new stations or station re-openings are being promoted and the data collected via the Laurencekirk user survey provides an invaluable insight into how new stations affect travel patterns.

5.71 New station openings, particularly in rural areas, also present particular challenges for transport modelling.  Successfully modelling a new station such as Laurencekirk would require a high level of detailed base year data, beyond that which would normally be achieved in a strategic model.  In addition, the survey has revealed behavioural changes which a conventional model would not necessarily be able to reproduce.

5.72 The Laurencekirk dataset could therefore be used as a test bed to undertake research into how strategic multi-modal modelling would have to be improved in terms of data, spatial detail and calibration to ‘match’ this outturn behavioural change.  The Aberdeen Sub Area Model (ASAM) covers this area and would be a suitable model to undertake this research.  The findings could then be applied in other models across the country which are being or could be used to appraise the impacts of new station proposals.  

5.73 The Laurencekirk survey respondents were asked for contact details and their willingness to take part in further research.  To assist in understanding these issues further and calibrating the model, the survey respondents could be approached with a further, more detailed, survey examining their before and after travel choices (including for example parking cost / availability, perceived journey times, household car availability etc), or this could also be undertaken via focus groups.  

Wider Economic Benefits

5.74 This section provides a summary of the findings of the analysis looking at the Wider Economic Benefits (WEBs) associated with the reopening of the station at Laurencekirk.  Wider Economic Benefits (WEBs) are a relatively new feature of economic appraisal within the transport sector and they were not formally considered within STAG at the time the Laurencekirk appraisal was undertaken.  The aim would be to investigate whether there are additional impacts that were not covered in the original appraisal that would have a significant bearing on the results if they were captured today.

5.75 WEBs contribute to the positive impact of transport on productivity and GDP and are caused by the existence of market imperfections in transport using industries.

5.76 The analysis focused on two areas[12]:

  • Agglomeration Impacts; and
  • Wider benefits arising from improved labour supply.

5.77 Given the limited impact the reopening of the station will have on WEBs, a proportionate and qualitative exercise was carried out in line with STAG.  The key aim was to understand whether a project of this size and nature is likely to have wider economic impacts not captured in the conventional Transport Economic Efficiency appraisal, and therefore feed the findings back to Transport Scotland to inform future appraisal and evaluations of similar schemes.  While the existence and scale of the impacts was of interest, the precise monetary value of the impacts was of secondary importance for the evaluation and these have not therefore been quantified.

5.78 In order to provide the required data for the WEBs analysis, we undertook an extensive survey of local businesses, with a view to understanding the impact of the station reopening.  A summary of the business survey is provided here. 

5.79 Interviews were carried out with 20 businesses, details of which were provided by Aberdeenshire Council.  The businesses covered a wide range of sectors and accounted for 7% of all businesses in Laurencekirk and the surrounding area area.  

5.80 Interviews were undertaken with a senior member of the business, typically the business owner, manager or director with the discussion focussing on the following areas:

  • business performance;
  • business turnover and profitability;
  • business access to key services;
  • access to customers;
  • journey time to and from key markets;
  • business competition;
  • staff recruitment; 
  • staff retention;
  • business travel; and
  • supply chain linkages.

5.81 The businesses interviewed were also asked about the impact of the reopening of the station on the wider local economy and the likely consequences had the station not reopened (i.e. the counterfactual position).

5.82 A key question in the surveys was whether the business had been located in Laurencekirk prior to the reopening of the station.  Businesses that set up in the town after the station was reopened were asked the extent to which the station impacted on their decision to locate in the town.  Of the 20 businesses contacted, 15 were already operating in Laurencekirk prior to the station reopening, while five located to the area after the reopening.

5.83 In the interests of completeness, interviews were also carried out with the Laurencekirk Business Club and Mearns Academy, one of the largest employers in the town.

Agglomeration Impacts

5.84 Transport can alter the accessibility of firms in an area to other firms and workers.  Agglomeration benefits arise because firms derive productivity benefits from being close to one another and from being located closer to larger labour markets.  If transport investment brings firms closer together and closer to their workforce this may generate an increase in labour productivity above and beyond that which would be expected from the direct user benefits alone as measured in conventional Transport Economic Efficiency analysis.

5.85 The results from the business survey suggest that, while the station has had a positive impact in terms of access to suppliers, markets and customers, agglomeration benefits have been limited and are likely a reflection of the relatively small number of firms located in Laurencekirk.

Labour Supply Impacts

5.86 Transport costs can affect the overall costs and benefits to an individual from working:

  • more people can choose to work if the costs of commuting (time and / or fare) fall;
  • people may choose to work more hours if their commuting time falls; and
  • improved transport can open up new opportunities and lead to better matching of labour supply with demand, leading to higher productivity.

5.87 The findings from the Accessibility analysis reported in section 4.2 showed that the station reopening had resulted in a reduction in public transport journey times to a number of areas of employment.  This was particularly so for Aberdeen, where journey times by public transport to the city centre had fallen from 115 minutes to 65 minutes.  In practice, due to the mix of car / Park and Ride travel behaviour observed in the surveys, the impact on observed travel times is less significant.

5.88 A number of questions were included in the user survey to understand whether the station reopening had contributed to people entering the workforce or moving job, whether people could get to work quicker, or whether they had used the time to work more hours.

5.89 Respondents were asked whether they had moved jobs since the station re-opened in 2009.  Of those responding to this question, the majority (65%, n=129) indicated that they had not moved jobs, 24% (n=47) stated that they had moved jobs and 12% (n=23) indicated they were not in employment.

5.90 For those that had moved jobs, Figure 21 below indicates how significant the re-opening of Laurencekirk Station was in participants’ decision to change employment. 

Figure 21. Significance of Station Reopening on Decision to Change Jobs

Figure 21.	Significance of Station Reopening on Decision to Change Jobs

5.91 As shown, the majority of those who moved jobs stated that the reopening of the station was a factor in their decision, with 11% (n=5) stating it was the main factor, 51% (n=24) indicating that it was one of a number of important factors and 9% (n=4) stating it was a fairly minor factor.  

5.92 Of those who moved jobs, 30% (n=14) stated that the re-opening of the station was not a factor.  Various reasons were given for this including participants’ old and new job both being in Laurencekirk, working abroad or driving as part of their role.  

5.93 Respondents were also asked whether the re-opening of the station encouraged them to enter the workforce or work more hours because the commute to work was now quicker.  As shown in Figure 22, of those who responded to this question, 70% (n=33) stated that it had not had an impact, 15% (n=7) said it had allowed them to enter the workforce, 9% (n=4) said it allowed them to work longer hours and 6% (n=3) said it had allowed them to both enter the workforce and work longer hours. 

Figure 22. Impact on Workplace Activity

Figure 22.	  Impact on Workplace Activity

Conclusions

5.94 One of the key tasks was to explore why the forecast number of passengers using Laurencekirk station were less than half of the outturn passenger numbers.  However, in undertaking the review of the original 2004 STAG report it became apparent that it does not contain enough detailed information, and sufficient clarity on the assumptions made, to determine the precise reasons why the forecasts were lower than outturn passenger numbers.

5.95 While the approach does appear to be generally consistent with the proportional approach recommended in STAG, the high-level nature makes it difficult to interrogate thoroughly.  

5.96 It is an important lesson learned that, if future outcome evaluations are to explore in detail why outturn numbers diverge from forecasts then the key underlying cause and effect assumptions must be recorded and clearly laid out in the STAG Report.

5.97 Because the approach was high-level and not totally consistent with the expected methodological approach, it was not possible to simply re-run the analysis using the outturn data.  In addition, the approach recommended in the current appraisal guidance is different from that which existed in 2004.  Consequently the cost benefit analysis was re-done in line with current guidance.

5.98 Using the outturn data, the revised BCR ranged from 2.5 to 4.4 depending on the scenario and assumptions used.  This compares with the BCR in the original report of 1.5.  Given the costs used are the same in both approaches, the increase in the BCR is generated by a higher level of benefits.  

5.99 Current transport appraisal guidance includes advice on techniques to capture impacts that have traditionally not been captured in conventional appraisal.  This includes Wider Economic Benefits (WEBs).  Analysis was therefore carried out to understand the impact of these Wider Economic Benefits from the re-opening of Laurencekirk railway station.  The analysis revealed that agglomeration benefits are limited.  It also revealed that while there may have been labour supply impacts, as the new station encouraged people to move to more productive jobs or those not employed to enter the workforce, the impacts will be limited for projects similar to Laurencekirk, i.e. a small station with relatively few users.  In light of these findings, in the spirit of the proportional approach to appraisal emphasised in STAG, it is recommended that WEBs impacts are not required to be analysed as part of future appraisals of projects of this nature.  It should still be considered for large and / or urban rail projects.