Bus services after the Spending Review

Written evidence from Professor Peter White (BUS 52)

1. The following evidence is submitted based on extensive experience of research into the role of the local bus industry in Britain over many years, and applying this to current changes affecting the industry. All opinions expressed should be regarded as my personal views, and not those of the University of Westminster or any other organisation. A supplementary note on statistical and financial data will be provided shortly.

Features of the market for local bus travel

2. In contrast to the railways, for example, the bus industry fulfils a distinctive role. The average length of journey is short (about 8 km) and demand is not dominated by the journey to work as might be imagined. Indeed, the most common journey purpose (as revealed by the National Travel Survey, NTS) is shopping, comprising about 30% of all trips. This is followed by journeys to/from work (about 20%), leisure (about 20%) and education (about 17%). A substantial element of personal business travel also found. The effect is that a better spread of demand by time of day is observed than for rail, with relatively high levels between the two traditional ‘peaks’. Indeed, much of the extent to which peak demand exceeds the typical daytime level is associated with school travel, rather than the adult journey to work. The consequence is that many operators run a similar level of service between about 0800 and 1800 on Mondays to Fridays, apart from some peak extra journeys mainly to cater for school travel. However, lower levels of service are typically offered during evenings and Sundays. Saturday daytime service levels are similar to those for Mondays to Fridays between the peaks. Another contrast with rail is that bus users tend to come from somewhat lower income groups. Bus demand is substantially more sensitive to car ownership than is rail. NTS data indicate substantially lower trip rates per person in households owning cars than those in non-car-owning households.

Trends in use and factors affecting demand

3. The broad long-term trend in bus use was until relatively recently one of decline, principally due to rising car ownership. As this rose, potentially fewer bus trips were made. The resultant loss of ridership was then reflected in fare increases and/or service cuts, which in turn caused a further drop in ridership. However, a much more positive outcome has been evident in recent years, notably in London, where bus use has grown by about 60% since 2000 from an already high level per head of population. A broadly stable level of bus use in the rest of Britain has also been observed, with substantial growth in certain areas where external factors (such as the local authority role in parking policy) and a positive management approach within the industry itself are evident, for example, in Brighton & Hove.

4. Another feature of demand for bus services is that demand sensitivity to changes in price and service level is relatively low, at any in the short run (a period of up to about one year). This is expressed through the ‘elasticity of demand’, which may be expressed in a simple form as the ratio of change in demand (in this case, bus passenger trips) to that in some factor causing change. For example, if fares rose by 10% and passenger trips fell by 4%, the elasticity is then equal to

- 4 (percentage change in trips) = -0.4

+10 (percentage change in fares)

The value has a negative sign, as an increase in fare causes a reduction in trips.

5. We can also measure effects of changes in the service level in the same way. A common method is to use the readily-available statistic of bus-kilometres run. For example, if on a fixed route operating for a certain period of the day the service was increased to run 3 times an hour instead of 2 times an hour, then bus-kilometres run would rise by 50%. In this case a similar elasticity is found, but with a positive sign (i.e. +0.4), since an increase in service level will cause a growth in demand (for example, as a result of reduced roadside waiting time, or greater convenience).

6. Extensive research, summarised in the report ‘The Demand for Public Transport: A Practical Guide’ (TRL593) confirms that these values are typical of the short-run outcome in Britain and similar countries.

7. A consequence of these elasticity values is that a fare increase will, despite causing some loss of trips, produce a net growth in revenue. For example, if a bus service carries 100 passengers per hour at a fare of £1 each, revenue produced is £100. Increasing the fare by 10%, to £1.10, would cause demand to fall by 4%, i.e. to 96 trips per hour. However, the new revenue per hour becomes 96 x £1.10, i.e. £105.60, a growth of 5.6%. Hence, the operator is still better off despite some loss of trips made.

8. The same applies to changes in service level. For example, increasing the service offered by 10% would only result in a growth in trips of 4%. If the additional bus-kilometres had the same average cost as those already operated, total cost would rise by 10%, and hence a loss would be made. The outcome is that an operator can offset a loss of trips due to some external factor (such as rising car ownership, or effects of the recession) by a combination of fare increases and service cuts such that financial viability is maintained, even after allowing for a further loss of ridership that occurs.

9. It also follows that ridership can be increased by raising service levels, and/or reducing fares, but that additional financial support may be needed in such cases - this is an element in the strong growth in London in recent years (discussed further below).

10. A further consequence of the low short-run price elasticity is that an operator can increase price levels to improve financial performance, even with a loss of ridership. In theory, excessive price rises (for example, to give profits above ‘normal’ levels) can be deterred by competition in the deregulated regions outside London. However, such competition appears to have limited effects in practice (albeit some exceptions are found).

Factors affecting costs

11. Although it is convenient to present unit costs of bus operation in terms of an average cost per bus-kilometre (see below) it is important to bear in mind that about 60% or more of costs are those of employing staff. These are paid on a basis of time, not distance as such. Some other costs, such as the capital cost of vehicles - reflected as an annual depreciation charge, or a leasing charge - likewise vary by time. Indeed, fuel costs are the one substantial element which varies directly with distance, corresponding to about 10% of total costs at current levels (after the effects of BSOG).

12. In many respects, the best way of reducing bus costs is to raise speeds, since the same vehicle and driver time can be used to produce more bus-kilometres (or alternatively the same level of service can be provided with fewer drivers and crews). These may come about through reduction in dwell times at stops (notably through fare collection methods avoiding cash transactions) and speeding up between bus stops, notably by provision of bus priorities and/or measures to reduce traffic congestion in general such as the congestion charging in central London.

13. Costs also vary by vehicle size, so that for example a 20-seat minibus will have substantially lower costs than an 80-seater double-decker. However, while some elements clearly vary in this way (vehicle capital costs, fuel consumption, etc.) the extent to which driver costs vary depends on whether an agreement exists to pay wage rates which differ by vehicle size. It is also the case that while total costs per kilometre of running smaller vehicles are lower, unit cost per seat-km is usually higher.

14. Until the mid 1980s, real cost per bus-kilometre tended to rise, as wage levels increased (in many cases, needed simply to attract sufficient staff as wages and working conditions in other industries improved). Lower speeds due to traffic congestion, and increased boarding time at stops (following conversion to one-person operation while retaining cash fares) aggravated these effects.

15. A striking feature from the mid-1980s was a sharp reduction in unit operating costs, by about 45% in real terms by 2000, both within London and in the deregulated areas elsewhere. One aspect of this was lower wage rates, as previous national agreements were replaced by local negotiation. Productivity was increased by sharp reductions in engineering and administrative staff. Utilisation of buses and drivers was improved by running more intensive services, improving the "between peaks" frequency at low marginal cost. However, since 2000 real costs have generally increased. For example, DfT statistics show a growth of about 23% in real terms between 2004-05 and 2009-10 outside London (nonetheless costs are still about one third lower in real terms than in the mid 1980s). Further reductions in engineering and administrative staff may be difficult to attain. One obvious element is that wage and working conditions have had to be improved to recruit sufficient staff or the right quality (notably in London), at any rate until the onset of the current recession In some cases, traffic congestion has slowed services down, causing costs to rise for reasons mentioned above. Fuel costs have fluctuated sharply, one effect of BSOG being that price to the bus operator is much more sensitive to world market prices than for fuel taxed at the full rate.

. Effects of changes in BSOG.

16. The Bus Service Operators Grant (BSOG) is a rebate to operators of registered local bus services of 80% of the duty that would otherwise be payable on fuel used. The effect is to approximately halve the net cost of fuel to the operator (varying by world market prices). Given that fuel is currently about 10% of total operating costs, abolition of BSOG would increase costs by about 10%. This is of similar magnitude to overall profits made by the industry. By reducing total costs, BSOG enables many services to be worth running commercially, and reduces the total cost of services which are run on contracts (for example to local authorities or Transport for London).

17. It can be argued that BSOG is a relatively crude means of assisting the industry, albeit simple to administer. It is certainly the case that fuel efficiency (kilometres per litre) in bus operations has deteriorated substantially in the last 10-15 years.. However, it does not follow that BSOG is a cause of this outcome (even at lower unit fuel costs, bus operators are unlikely to use fuel wastefully if they can avoid it). Contributory factors include the lower fuel efficiency of buses meeting the stricter environmental standards for pollutants (currently ‘Euro 5’) than older vehicles of the same size they replaced, and some increases in vehicle weight (for example, as a result of wheelchair ramps being fitted, or air conditioning). The requirement to ensure new vehicles meet Euro 5 standards is compulsory, and thus outside the control of the operator. Hence as fleet renewal takes place (in part to meet other statutory requirements, notably accessibility) some increase in fuel consumption may be inevitable. It should also be borne in mind that the overall average for bus operations in Britain as a whole is in effect a weighted average for the whole industry. As bus-kilometres have grown substantially in London (see below) this will itself influence the weighted average, since operations in London typically involve a higher proportion of larger vehicles (typically double-deckers) and operate under congested conditions.

18. Having said this, BSOG clearly has some limitations. It is paid to any operator, irrespective of whether an existing service is already provided. It can be argued that payments related to volume of passengers carried would represent a more effective use of public expenditure, by incentivising ridership growth – see, for example, reports from the Commission for Integrated Transport (CfIT). Such changes would, however, depend on good quality of data being available –for example, from smartcard systems. There would also be a danger if only ‘trips’ were measured (i.e. each time a passenger boards a bus). This could be to the disadvantage of rural areas, where passenger densities are lower (but average trip length may well be greater). Some improvements in data quality are desirable before such a shift occurs.

19. The decision of the government to reduce BSOG, but in a phased manner, should enable the industry to adjust to these changed conditions without undue difficulty, although nonetheless some increase in total operating cost is likely to occur, with potential effects on service viability.

Concessionary fares

20. A notable feature of the bus system in Britain is the relatively generous provision of concessionary travel to certain categories in the population, notably free travel after the morning peak to those aged 60 upward. Elsewhere in Europe, while overall levels in financial support are often more generous, and adult fare levels lower in consequence, completely free travel on this scale is unusual, a more common pattern being to offer a reduced price pass to such categories of user.

21. The present system of nationwide concessionary travel in England can be traced to the introduction of common half-fare minimum entitlement in 2001, followed by free travel within the area of residence from 2006, and nationwide in 2008. It has resulted in substantial growth in bus demand compared with what would otherwise be expected, notably in rural regions. There is evidence – for example, from a case study in the Salisbury area – that a large element of the growth in usage is due to those already aged 60 and over taking up the concessionary pass entitlement for the first time, and making additional bus trips as a result. Such ‘new’ users tend to be more affluent than traditional bus users. As a result, one might question whether social equity objectives are being attained – conversely, there has been a greater effect on car use as a result for attracting such user groups, thus assisting other transport policy objectives.

22. In principle, bus operators are compensated on a ‘no better off, no worse off’ basis, i.e. the revenue that would have been received had the equivalent category of passengers paid the fares set commercially by the operator can be estimated, and compensation paid accordingly. In addition, some capacity costs may be incurred – for example, on a rural service which normally requires a minibus, a larger vehicle may now be needed. In the past it was generally assumed that such costs were low, using empty capacity outside the peak periods. This, however, may be less true now – for example, given that free travel applies in the afternoon peak, extra capacity may be needed as at such times as a result.

23. The compensation payment depends on the ability to estimate the volume of travel that would have been made had the user category in question been required to pay the normal fare. This is in effect a special case of the elasticity concept described earlier. A particular difficulty arises in respect of estimating what would happen where a fare is charged in place of the free travel now available. Extensive research has been conducted to attempt to determine these effects, and hence the net compensation payable to operators (however, relatively little research has been undertaken on evaluating the benefits to the recipients of the concession, and hence whether such a policy is socially justified).

24. A very large annual compensation payment in the order of £1,000m per annum for Britain as a whole, is now made in respect of concessionary travel. Current changes in the compensation principles used are likely to reduce this, although operators may be affected adversely as a result, if no corresponding changes in costs can be attained (which are unlikely, since passenger demand is affected by the free travel entitlement, not the level of compensation between public authorities and the operators).

School travel

25. Concessionary travel is a major example of ‘mandatory’ commitments placed on local authorities, and hence of unavoidable expenditure. The other major case is school travel, for which Local Education Authorities (LEAs, typically county councils or unitaries) are obliged to provide free travel to eligible pupils (broadly speaking, those aged 16 or under living more than 3 miles from the nearest appropriate school, with some variations introduced in recent years). This can be provided by local authorities running their own fleets (relatively rare) or more commonly; contracting in vehicles from bus and coach operators. Such travel can also be provided by local authorities purchasing season tickets on behalf of pupils, for travel on scheduled public services (typically registered local bus routes, but also including rail, for example). As in the case of concessionary travel, this also incurs expenditure in the order of £1,000 million per annum.

26. Scope exists for integrating such school services with the public networks, especially in rural areas, so that the same vehicles carry both school pupils and the general public – this may the only means of providing a journey-to-work service for adults, where demand would not justify a separate operation for this purpose. Even where such direct integration does not occur, vehicles required for school services at peak periods may then become available for operation of services at other times of day – for example, for shopping – at relatively low marginal cost.

The role of smart cards.

27. As mentioned earlier, one means of speeding up bus services is to reduce dwell time at stops. A feature of the conversion to one-person-operation (opo) from the early 1970s onward (i.e. removing the conductor) to reduce operating costs was that ticketing systems with a high proportion of cash fare payment were often retained. This incurs boarding times in the order of 6 seconds per passenger, compared with about 2 seconds or less where a pass is displayed or smartcard automatically validated. As well as slowing down average bus speeds, high boarding times per passenger tends to produce more irregularity in service operation, due to total dwell times varying with volumes of passenger boarding. London offers an example where very low boarding times per passenger are observed (especially for the articulated single-deckers with simultaneous boarding through all three doorways, in effect a marginal boarding time of less than one second per passenger)

28. The government is giving encouragement to greater use of smart cards as part of its reform of BSOG. An irony at present is that nationwide passes for the English concessionary fare system have generally been issued as smartcards, but in many cases lack of necessary equipment on-vehicle for validating such cards (or ‘back office’ facilities for analysing data) means they do not function as such. Hence, the potential benefits of stricter validation of card use, and data collection to estimate usage more accurately, have not been attained. A further benefit of smartcard systems is that, as well as enabling data to be obtained on travelcard and concessionary pass use, they also enable new products such stored value ticketing (’Pay as You Go’ in London) to be introduced, making cashless travel attractive to the less frequent users.

29. It must be stressed, however, that such boarding time benefits only come about where the cash transaction and issue of a paper ticket has been eliminated. This is relatively easy in a system such as London where a common fare applies over the whole network. It is much more difficult in networks where complex graduated fares are still found, and in some cases paper tickets are still issued even for free concessionary travel, in order to provide a record of fares otherwise payable (and hence the compensation described above).

Local authority expenditure and tendered services

30. In addition to the mandatory elements described above, a substantial discretionary element exists, notably in respect of provision of tendered services. In the deregulated areas outside London, about 80% of bus-kilometres are operated ’commercially’, i.e. they are registered by the operator, who also determines fares to be charged. There is no specific public expenditure contribution to the costs of each individual service thus registered, although their viability is clearly improved by the role of BSOG (see above). Public expenditure is also incurred through compensation to the operator for concessionary travel.

31. The remaining 20% of bus-km is operated with support from local authorities, generally secured through a system of competitive tendering. Such services typically comprise routes or periods of the day/week where low loadings are found (for example, some routes in rural areas, and very often operation of certain parts of the day or week on routes which are otherwise served commercially such as evenings and Sundays). There are also cases where tendered operation occurs at periods when large numbers of ‘non-statutory’ school trips are made (i.e. those below the distances for which free travel entitlement applies). Although a fare is paid, this often does not cover realistic costs for extra peak capacity.

32. Such tendered operations thus enable provision of a range of services that would otherwise not be available, and benefit in particular low-income users from non-car owning households for whom travel might otherwise not be possible (for example, in rural areas or on Sundays).

33. The current cutbacks in central government grants to local authorities potentially create the threat of serious cuts to such service provision, and already some authorities have indicated cuts in such expenditure, in some cases exceeding the general rate of local authority expenditure reduction. A particular danger exists in local bus operation that, as authorities are required to maintain the mandatory requirement to fund free concessionary travel, the budget for tendered service provision is squeezed still further. This creates the perverse outcome that users have retained the right to free travel, but may no longer have service on which it can be used. I believe that a strong case exists for reviewing the present level of free concessionary provision, both in terms of its overall affordability, and whether within the public transport sector, it would be more sensible to shift some of this expenditure to service support (for example, by reverting to a half-fare entitlement, rather than completely free travel).

34. Much attention has been paid to the development of demand-responsive services in rural areas in recent years. While enabling very low density patterns of demand to be served, it should be noted that these generally incur fairly high public expenditure per passenger trip made, in some cases exceeding the upper limit which authorities set for support to conventional ‘fixed route’ services (yet both ultimately serve the same purpose). It is also noteworthy that some well-run ‘conventional’ bus services in rural areas (i.e. offering a fixed route schedule) have substantially increased demand in recent years, such as Norfolk Green (in the Kings Lynn area), and Western Greyhound (in Cornwall). Contributory factors include more attractive frequencies, better quality of service, innovative marketing, etc.

The London case

34. The very strong growth in bus use in London in recent years, especially in comparison to trends in other conurbations, raises the question of whether the regulatory system applying in London should be applied to other areas. In addition to the recent growth, it is also worth bearing in mind that ridership in London remained broadly stable from the mid 1980s to late 1990s, compared with sharp decline in aggregate in the deregulated areas. The fall in bus operating costs (see above) enabled operations to be run with almost zero directly support from public funds by1999/2000 (apart from BSOG and concessionary travel compensation). However, given the nature of the labour market in London some increase in wage rates was probably inevitable simply to sustain operations at that level. Subsequently, following a policy decision by the first elected Mayor of London, expenditure to support bus operations in London has greatly increased, enabling much higher levels of service, lower real fares and other improvements.

35. I have analysed the components of growth in bus use in London since 2000, and find that the most obvious explanatory factors are the increased level of service, and lower real fares. The direct impact of the congestion charge, while causing a substantial shift from car to public transport, only forms a small element of the growth in bus usage. Other factors include:

- Some growth in population in London

- A marked expansion in service levels in evenings, Sundays and overnight, providing a much more comprehensive service

- Comprehensive passenger information, not only through real-time systems, but also printed material at stops, etc.

- Simplified fares and use of smartcards (Oyster) enabling very low boarding times per passenger, improving service quality

It is also noteworthy that about three quarters of the growth in bus use has occurred outside zone 1, associated with substantial expansion of service levels in outer zones. Another striking feature of the London case in that car ownership per head has changed very little since the mid 1990s, in contrast to strong growth elsewhere. This does not directly produce a growth in bus use as such, but represents the absence of a negative factor found elsewhere.

36. The London experience suggests that while some factors may be unique to London, certain elements could be transferable elsewhere, notably the benefits of comprehensive network planning, off-vehicle ticketing and comprehensive information. The ‘quality contract’ concept introduced in the Transport Act 2000, and made more feasible in the Local Transport Act 2008, could enable such systems to be introduced in other conurbations. A rather fruitless argument has taken place for many years about whether the trends in London are due to policies which could be followed elsewhere, or due to unique local factors (and more recently, high public expenditure levels). It would be very helpful to test the quality contract concept in one or two conurbations outside London, monitoring the outcomes very thoroughly, to see what the outcomes are in practice.

Network planning

37. Any form of bus network planning, whether undertaken by a public body such as Transport for London, or a commercial bus operator, is in effect seeking to establish the best compromise between individual person trips (defined by origin and destination, timing, etc.) and provision of a public transport service (which relies on consolidating demand in order to produce adequate loadings on the services). Evidence such as service level elasticities (described above) enable effects of service changes on demand to be assessed. Clearly, a good set of data on existing demand is needed, for which systems such as smartcards may augment traditional survey methods.

38. Passengers effectively participate in this process by responding to the service changes produced, which can be monitored, both with respect to overall social objectives, and the financial ones of operating commercially-registered services. It may also be appropriate consult users on specific service change proposals, to determine their feasibility in practice (there is, however, the danger that a small number of adversely affected users may be very vocal, while a larger number who might benefit have less impact on this process).

39. The London approach offers an example where a centralised planning system is used. This enables estimates to be made of the likely effects on demand on changes in bus service patterns. These methods in turn incorporate values derived from passenger behaviour to assess the effects of interchange, changes in waiting time, service reliability, etc. This provides a good method of making incremental changes, which are beneficial to users. However, it does have the limitation that more radical changes may not be considered. It also involves a much greater element of formal consultation than in deregulated areas, but this does carry the risk that where new services are proposed in areas not currently bus served, articulate car users may object, possibly delaying introduction of services which maybe very beneficial to bus users (such as elderly people unable to walk longer distances to existing services).

40. In contrast, the planning found in deregulated areas is largely initiated commercially by operators, with little requirement for formal consultation. This enables innovations to be introduced quickly (a notice period of 56 days), but can also produce some instability in the network. User reactions can be assessed through ridership and revenue changes. One problem arising is the relationship between commercial and tendered services. The latter may tend to be a ‘gap filling’ exercise in which links not provided commercially by operators are maintained by local authority support – given the fact that existing commercial services (for example, an evening operation) can be withdrawn with only 56 days’ notice, local authorities may have little scope to evaluate whether such services should be continued. A more comprehensive approach to commercial and tendered services viewed together might provide a more effective network than the present system.

41. In addition to consultation on specific network changes, users can be consulted on their ratings of existing service attributes (as already found through the quarterly survey of bus users’ attitudes), and the relative importance they attach to different aspects of services (such as frequency, reliability, information, etc.). Recent market research by Passenger Focus has been very useful in this respect.

January 2011