Scotland-wide Older and Disabled Persons Concessionary Bus Scheme - Further Reimbursement Research
5 Additional Costs
5.1 Additional operating costs of increased concessionary demand
5.1.1 The two reimbursement components discussed in the previous Chapters primarily affect the calculation of reimbursement for revenue forgone. Bus operators should also be reimbursed for additional costs incurred as a consequence of the concession, particularly with regard to additional operating costs associated with carrying generated concessionary passengers.
5.2 Base additional cost rates
5.2.1 The current study brief did not call for a clean-sheet analysis of additional costs, since this was the larger part of the ITS 2009 Study for Transport Scotland. The emphasis of the current work is therefore on changes over time, using the ITS results as a base.
5.2.2 The key conclusion from the ITS work was that reimbursement of bus operators for additional costs could be summarised in terms of a payment per generated passenger. ITS established a range of values. Those used as the basis of the negotiations between Scottish Government and CPT which led to the National Scheme arrangements that came into effect on 1 April 2010 were:
- £0.065 per generated passenger for marginal operating costs; and
- £0.279 per generated passenger for marginal capacity costs and other operator responses.
5.2.3 Both of these appear to be defined in terms of 2009-10 prices, although there are some questions about the price base which are discussed below. However, our starting point is the ITS conclusion that reimbursement payments should be £0.344 per generated passenger in 2009-10 prices.
5.2.4 In the current Study, CPT has made us aware that the ITS final report did not address a number of concerns and questions raised at the time. There is not scope within the timescale and resources of the current study to unpick the detailed analysis undertaken by ITS, but we have further examined some of these and other issues, which would lead to the possibility of the values quoted above being underestimated.
5.3 Relative labour costs
5.3.1 The ITS conclusions are largely (but not exclusively) based on econometric analysis of marginal cost rates, using STATS100 data from bus operators between 1999 and 2006. Most of this data is from operators in GB outside Scotland, and so although the analysis could be carried out separately for Scottish operators, the smaller volume of data for Scotland gives results which are less statistically robust. However, ITS was also able to identify from the GB econometric results, statistically sound "fixed" effects from which it was concluded that Scottish costs were 13.9% lower than the those for GB (outside London). The preferred econometric model marginal cost rates were therefore based on GB (outside London) values, which were regarded as being statistically robust, less 13.9% to reflect the ITS conclusion about comparative costs in Scotland and elsewhere.
5.3.2 It is understandable that the plausibility of the 13.9% adjustment should have been questioned, since it can be demonstrated that wage rates for bus and coach drivers in Scotland are now higher than in the rest of GB outside London. But other evidence on average costs shows a similar picture to the ITS conclusion, namely that operating costs in 2006-7 were lower in Scotland than in GB outside London at that time (14.0% less per vehicle km, and 8.4% less per passenger journey). This apparent conflict of evidence can be reconciled by noting that the increase in bus operating costs between 2006 and the present day was greater in Scotland than in the rest of GB outside London, as shown, for example, by the CPT cost indices.
5.3.3 In fact the ITS conclusion on marginal cost rates, and on the appropriate level of reimbursement rate per generated passenger, were based on an amalgam of sources, including data supplied by CPT (although ITS modified the assumptions made by CPT). ITS used different methods to update their figures, including a GB-based rate (which would probably understate the increases actually experienced in Scotland), and updating with the CPT cost index. The value of £0.279 quoted above (identified in paragraph 5.2.39 of the ITS report) is actually based on operator supplied data, updated by the CPT cost index.
5.3.4 While it is not self-evident that the ITS conclusions should be modified so as to reverse the 13.9% reduction, it is not at all clear that the ITS conclusions would have been the same if later data had been available for Scottish operators on which to base more detailed analysis. It is probable, given the differences in wage rates between Scotland and the rest of GB which can now be demonstrated, that a higher value could be justified than that adopted by ITS. There are almost certainly further areas of debate that could be had about the ITS analysis.
5.3.5 There are technical difficulties in coming to a definitive view, and a pragmatic approach is that the 13.9% reduction in cost rates applied by ITS should be removed. This is equivalent to increasing the ITS rate by 16.14%, leading to a base value for additional cost per generated passenger of £0.344 * 1.1614 = £0.400 in 2009-10 prices.
5.4 Base costs
5.4.1 There remain a number of other areas of uncertainty with regard to the basis of for the specific additional cost values recommended by ITS. In particular, the cost indices that ITS have used in various part of the analysis to bring costs to a common 2009-10 price base, seem likely to have underestimated the differences in cost inflation incurred by Scottish bus operators relative to those in Great Britain (outside London).
5.4.2 CPT has attempted to simulate the ITS calculations, using a different cost inflator, and reflecting the fact that some of the alternative cost rate estimates are derived directly from Scottish operator data in 2009. The overall conclusion reached was that the combined net additional cost per generated passenger (including both marginal operating and marginal capacity costs) should be increased by 7% relative to the figure quoted by ITS.
5.4.3 Making these adjustments is far from straightforward, and there remains considerable scope for uncertainty with regard to both the interpretation of how ITS reached the quoted conclusions on additional cost rates, and how it might be corrected. But it seems likely that the ITS analysis may not have fully recognised the differences in changes in costs rates between the bus industry in Scotland and that elsewhere in Great Britain, and that as a consequence some adjustment is merited. Given the difficulty of further diagnosis of the ITS methodology, it is therefore proposed that the net additional cost rate per generated passenger is increased by 7.0% from £0.400 to £0.428 in 2009-10 prices.
5.5 Impact of differences in average journey length
5.5.1 Data provided by CPT, referenced previously in Chapter 3, shows that in 2011, the average single cash fare that would be paid for concessionary journeys was some 19% higher than the average adult single cash fare actually paid by adult non-concessionary passengers. This implies that on average, concessionary passengers make longer journeys than non-concessionary passengers buying cash single tickets. There is evidence from a number of sources that journey lengths by concessionary passengers increased substantially following the introduction of free travel in October 2002. CPT has suggested that concessionary journey lengths have continued to increase beyond that date, and in particular from 2006-7 (the period on which ITS conclusions are predominantly based), and consequently proposed that additional cost rates should be further increased to a cumulative value of £0.459 in 209-10 prices.
5.5.2 The most consistent data on bus journey lengths comes from the Scottish Household Survey (SHS), using the age of respondent (60 or above, and below 60) as a proxy measure to distinguish between journeys by concessionary passholders and others. However, the modest sample sizes for individual years, coupled with the probable influence of erratic winter weather patterns, do not give as clear a picture of overall trends as would be desirable. Bespoke analysis of SHS data by Transport Scotland gives the annual data summarised in Table 5.1, in which data has been summarised by calendar year and by financial year.
|Calendar Year||60 or older||Less than 60||Financial year||60 or older||Less than 60|
5.5.3 With a typical sample size in later years of less than 800 or so, there is little merit in attempting a more detailed disaggregation by time, but plotting the annual data at the mid-point of each year (i.e. 1st July for the calendar year and 1st September for the financial year) gives a broad sense of the underlying trends, as shown in Figure 5.1 below.
Figure 5.1 Average bus journey lengths
5.5.4 It is clear that over the period covered, the average bus journey lengths of passengers aged 60 or more have increased to about the same length of journeys as those of bus passengers aged less than 60. However, in recent years the data shows large fluctuations, with in particular a significant fall in the average trip length of older passengers in 2009-10 which presumably coincides with a period of extreme winter weather. However, it is difficult to conclude that there is strong evidence of a consistent upward trend in journey lengths from 2006-7, nor indeed that concessionary journey lengths are currently significantly different to those of non-concessionary journeys.
5.5.5 Consequently, we do not believe there is any justification for further uplifting of the ITS additional cost rates on the grounds of increased average concessionary trip lengths. However, it is acknowledged that there is uncertainty about future trends in concessionary journey lengths. We recommend that Transport Scotland should seek to monitor the extent to which the trend in journey lengths by concessionary passengers differs significantly from that of non-concessionary passengers. If this is the case in the future, it would present some grounds for review of additional cost rates.
5.5.6 It should be noted that even if not increased on grounds of concessionary journey length, the additional cost rate of £0.428 per generated passenger that is our preferred value is substantially greater than equivalent values that we are aware of on an anecdotal basis from applications of the DfT Reimbursement Calculator in England. In our experience it is rare to obtain values that exceed £0.25 per generated passenger, and often additional capacity costs are regarded as zero, so that the only reimbursement is for marginal operating costs at about £0.06 to £0.07 per generated passenger.
5.5.7 However, for meaningful application the DfT Calculator requires an array of local values that reflect local network characteristics, and has certain characteristics that lead to counter-intuitive results. Consequently, although the recommended additional cost rate of £0.428 per generated passenger is at the high end of the potential range of values, we have no particular reason for believing that this does not properly reflect the data and analysis carried out by ITS, with the exceptions of the modifications identified above. In any case, it is worth noting that even with additional cost rates of this order of magnitude, the overall contribution of additional costs to total reimbursement is significantly less than that of revenue forgone.
5.6 Updating additional cost parameters over time
5.6.1 The most appropriate way of allowing for additional costs in reimbursement calculations is to apply a marginal cost rate per generated passenger to an estimate of the volume of generated passengers, as derived from the Reimbursement Factor. The discussion above leads to the view that the best basis for the additional cost rate per generated passenger is a value of £0.428 in 2009-10 prices, based on the ITS conclusions but increased to allow for the likelihood that differences in cost rates between Scotland and the rest of Great Britain were not sufficiently taken into account.
5.6.2 The question then arises as to how this value should be adjusted for later years, given the impact of general inflation, as well as of price factors (such as fuel and labour costs) that are specific to the bus industry.
5.6.3 Different indices are available, including two indices of average costs per vehicle kilometre and average cost per passenger journey which are published as National Statistics by Transport Scotland. However, these can be volatile, partly because of the dependence on estimates of total vehicle kilometres and passenger journeys, and are subject to some publication lags. The preferred option is to use the bus industry's own cost index, which is prepared and published by the CPT. This takes the form of a report on the percentage change in operating costs over the previous year, based on returns submitted to an independent third party which collates the data from individual operators. Updates are published every six months (reflecting the annual change to end‑December and then end-June) with a lag of about three months.
5.6.4 The index values calculated from the published CPT index for the years from 2009-10 are summarised in Table 5.2. These have been constructed by taking the average of the December and June increases to give an estimated increase in the year ending in the March between these dates, thus smoothing short term fluctuations. For 2012-13, the increase in the year ending June 2012 has been assumed to apply to the year ending December 2012.
|Additional Cost index||2009-10||2010-11||2011-12||2012-13|
5.7 Conclusions on Additional Costs
5.7.1 Bus operators are entitled to reimbursement for additional costs incurred as a result of providing bus travel at free fare. The preferred approach is to calculate this through the application of a marginal cost rate per generated passenger.
5.7.2 We have reviewed the ITS 2009 Study for Transport Scotland and concluded that on the whole there is no strong reason to dispute the methods used by ITS to derive estimates of marginal cost rates.
5.7.3 However, an unresolved issue from the 2009 study was an ITS view, based partly on econometric analysis, that relevant Scottish bus operator's labour costs are lower than those in England. It is not possible to unpick the ITS analysis that led to this view, but there is sufficient room for doubt to lead us to the conclusion that it is not unreasonable to recast the ITS values without this reduction. We have also looked at the internal calculations within the ITS work, and concluded that these seem likely to have inadequately accounted for faster growth in costs in the bus industry in Scotland from 2006-7 compared to the rest of Great Britain outside London.
5.7.4 CPT has also proposed that the cost rate should be increased to reflect an assumption that on average concessionary journeys are longer than those of non-concessionary passengers. In our view, there is little strong evidence to support this assumption, and we do not believe that an adjustment on these grounds is justified.
5.7.5 This leads to an estimated marginal cost rate per generated passenger of £0.428. We note that this is much higher than equivalent values currently in use in England. A compromise rate, which is the average of our preferred value and that proposed by CPT, is £0.443 per generated passenger in 2009-10 prices.
5.7.6 This value will need to be updated from time-to-time to reflect general price inflation and influences on cost levels that are specific to the Scottish bus industry. In our view an index based on the cost index published by the CPT reporting on changes in costs for Scottish bus operators provides the most-robust basis for this updating of the cost per generated passenger.