House of Lords - The Economics of High Speed 2

The Economics of High Speed 2 - Economic Affairs Committee Contents

CHAPTER 3: DEMAND AND FARES MODELLING

79. The Strategic Case stated that "the modelling set out in the economic case for HS2, published alongside this document, shows that there is a long term demand for additional capacity on our north-south railways."[89] This Chapter explains that modelling and compares the long-term demand forecasts to recent trends. It also examines why the modelling assumes that fares on HS2 will be the same as on the existing railway and looks at demand forecasts for HS1 and the Train à Grande Vitesse (TGV) in France. The next Chapter considers the extent to which additional capacity is required.

How demand is modelled

80. Travel demand forecasts are prepared using an 'elasticity based model'. This approach seeks to determine a statistical relationship between observed demand for transport and factors that affect the demand. The relationship between these factors and demand are determined from previous experience and research.

81. The model described in the Economic Case for HS2, developed by HS2 Ltd on behalf of the Government and published in October 2013, splits Great Britain into 235 "zones". Using travel statistics, it models the demand for journeys between each zone by road, rail and air (for the purposes of the model the base year is 2010). Using the historic rate of demand growth for each mode of transport, and other factors such as GDP growth and population changes[90], the model predicts future demand for each mode of transport between each zone in two scenarios: without HS2 and with HS2.[91]

82. The Department for Transport requires demand forecasts to be capped at some point in the future as "it is not reasonable to expect rail demand to grow indefinitely".[92] The cap is reached when the model predicts that the number of long-distance rail journeys over 100 miles nationwide reaches 290,146 a day.[93] Professor Mackie, Emeritus Professor, Institute for Transport Studies, told us that "it is a bit of an arbitrary assumption when you cut growth off. However, it is better to have a cap than not to have a cap, because there are limits to forecasting".[94]

83. The rest of this chapter looks at the forecasts in the 'without HS2 scenario' and the 'with HS2 scenario'. These forecasts are central to the economic case for HS2: the forecast increase in long-term demand for rail travel without HS2 supports the argument that extra railway capacity is needed; the forecast number of passengers that will use HS2 provides the basis for estimating the economic benefits of the new railway.

Forecast demand—'without HS2 scenario'

84. The model predicts that without HS2, the number of long-distance rail journeys over 100 miles will reach 290,146 trips a day (the 'demand cap') in 2036. This equates to an average annual increase of 2.2 per cent across all the zone-to-zone pairings in the model that are over 100 miles.

85. The 2.2 per cent figure is an average; some zone-to-zone forecasts are higher, some are lower. The Economic Case does not include information on individual zone-to-zone forecasts but a report by Atkins[95], describing the development of the most recent version of the demand model, lists some of the relevant ones for HS2:[96]

Table 6: Average number of weekday rail trips[97] and growth between London and city council areas without HS2, as predicted by the demand modelling

Zone-to-zone movements	2010 demand	2036 demand	Average annual increase, %[98]
Birmingham-Central London	7,000	13,700	2.6
Manchester-Central London	6,600	13,500	2.8
Leeds-Central London	4,200	8,800	2.9
Liverpool-Central London	2,600	4,800	2.4
Newcastle-Central London	2,300	4,200	2.3
Edinburgh-Central London	2,100	4,500	3.0
Glasgow-Central London	1,100	2,200	2.7

Source: Atkins Model Development Report: PFMv3.0-PFMv4.3, September 2014, page 95

THE DEMAND FORECASTS ARE "CONSERVATIVE"—THE GOVERNMENT

86. The Economic Case compared the average annual growth to recent trends: "The very strong growth in demand for journeys on long-distance rail operators' services since 1994 … has equated to an average year-on-year growth rate over the past 18 years of 4.9%". It continued that "the assumed rate of growth of demand … at an average of 2.2% per annum from 2010 to 2036 … is lower than the recent trend."[99]

87. Mr Prout of the Department for Transport said that if you compare the demand forecasts to the historic trend over the last 20 years, "they are reasonable. They are possibly an under-estimate … The population will go on increasing at 5 million every 10 years, but we cap demand in 2036."[100] He described the forecasts as a "conservative estimate".[101] Ms Munro of HS2 Ltd explained that:

102

88. We note that the increase described by Ms Munro equates to a 38 per cent rise in the average number of long-distance rail trips per person per year. This is a substantial increase by any standard. The Secretary of State said "history and recent precedent shows that it [growth in demand] has been far in excess of … 2.5 per cent." He described the forecasts as "a very conservative estimate about future growth."[103] This is not borne out by Figure 3.

THE DEMAND FORECASTS APPEAR BELIEVABLE—ACADEMIC WITNESSES

89. Professor Tony Venables, BP Professor of Economics, University of Oxford, thought the forecast demand "does not seem to me to be grossly inflated at all."[104] Professor Chris Nash of the Institute for Transport Studies said that the "best estimate is that the sort of growth that we have had in the last 20 years will continue."[105] Professor Mackie said that he did not find the input assumptions as "completely outside the plausible ballpark."[106]

90. Professor Dan Graham, Professor of Statistical Modelling, Imperial College London, believed that "you could construct an argument to say it is too high, you could construct one to say it is too low."[107] Professor Glaister told us that the forecast "does not look unreasonable, but whether it is the best estimate, I could not say."[108]

RECENT GROWTH IN LONG-DISTANCE RAIL TRAVEL—NATIONAL TRENDS

91. The Economic Case stated that the average year-on-year growth rate from 1994 to 2012 for journeys on long-distance rail services was 4.9 per cent. It said that long-distance rail has grown "particularly rapidly and consistently since 2004."[109]

92. The Office of Rail Regulation's website provides statistics for long-distance rail journeys going back to 2002/03. These statistics show that the number of journeys on franchised long-distance train operators increased from 77.2 million journeys per year in 2002/03 to 129 million journeys per year in 2013/14:[110]

Figure 2: Passenger journeys on franchised long-distance rail operators, millions

Source: Office of Rail Regulation Data Portal, Passenger journeys by sector: Table 12.6 [accessed March 2015]

93. This equates to an average annual increase of 4.8 per cent. The following graph shows the year-on-year percentage increase in long-distance rail journeys:

Figure 3: Year-on-year percentage increase in journeys on franchised long-distance rail operators

Source: Office of Rail Regulation Data Portal, Passenger journeys by sector: Table 12.6 [accessed March 2015]

94. Mr Rukin from Stop HS2 told us that "growth in long-distance travel has completely bottomed out over the last four years. It has been on a consistent decline. It is now under 1 per cent."[111] The graph above shows that long-distance journeys increased by 1 per cent from 2012/13 to 2013/14.

95. The latest quarterly figures show an upturn in growth for franchised long-distance rail journeys: in Quarter 3 of 2014/15, there was a 5.1 per cent increase on the same quarter last year.[112] In Quarter 1 and Quarter 2 of 2014/15, there was a 1.5 per cent increase and a 3.5 per cent increase in journeys compared to the respective quarters for 2013/14. The Office of Rail Regulation statistical release attributed the increases primarily to higher sales of advance and off-peak ticket travels, "key drivers for journeys in this sector" which signify "increase in the leisure travel segment, with people making the most of the travel incentives offered by the operators".[113]

RECENT GROWTH IN LONG-DISTANCE RAIL TRAVEL—REGIONAL TRENDS

96. The Department for Transport does not publish information on the number of journeys between stations. The Office of Rail Regulation however does publish statistics on journeys between regions and London:

Figure 4: Annual rail journeys to/from London, thousands (including weekends)

Source: Office of Rail Regulation, Regional rail journeys: London: Table 15.4 [accessed March 2015]

97. Table 7 below compares the average annual growth between regions and London since 1995/96 against the forecast average annual growth between selected cities and London (as shown in Table 6 above). Although not a direct comparison, the selected cities provide the majority of the demand for travel to and from London from within their region. Also, the growth since 1995/96 includes weekday and weekend travel; the predicted growth in the model is only for weekday travel.

Table 7: Predicted average annual increase in journeys to/from London versus observed average annual increase in journeys to/from London since 1995/96

City council areas[114]	Predicted average annual increase in weekday trips by the demand model for HS2, %	Average annual increase in journeys between the region and London since 1995/96, %[115]
Birmingham	2.6	6
Liverpool	2.4	4.8
Manchester	2.8
Leeds	2.9	5
Newcastle	2.3	4
Edinburgh	3.0	3.5
Glasgow	2.7

Source: Atkins Model Development Report: PFMv3.0-PFMv4.3, September 2014, page 95; Office of Rail Regulation, Regional rail journeys: London: Table 15.4 [accessed March 2015]

98. To compare actual weekday growth with predicted weekday growth, the Chairman wrote to the Secretary of State to ask for data that showed the split between growth in weekday and weekend travel since 1995/96. The Secretary of State replied that the data is only available for the past 400 days. Data from the National Travel Survey shows that the weekend share of long-distance journeys has increased in recent years: from an average 76 per cent/24 per cent split for weekday/weekend travel between 2006 and 2009 to an average 74 per cent/26 per cent split between 2009 and 2013.[116]

99. Richard Scott, Director of Corporate Affairs, Virgin Trains told us that demand had been "increasing consistently" in the 17 years that they had run the franchise for long-distance trains on the West Coast Main Line (which serve Birmingham, Glasgow, Liverpool and Manchester). He said that Virgin were forecasting 60 per cent growth on their services to 2026.[117]

100. In written evidence, Chris Stokes, former executive director at the Strategic Rail Authority and former deputy director for British Rail Network Southeast, stood by his challenge to the growth assumptions made in a December 2012 article in Modern Railways.[118] In the article, he offered a "cautious hypothesis" on recent growth on the West Coast Main Line:

· "Recent high growth has been driven by a step change following completion of the [West Coast Main Line] upgrade in December 2008";

· "There has been a one-off modal shift, especially from air to rail in the Manchester-London market";

· "There has been significant growth in off-peak and weekend travel, but the business market is saturated";

· "Rail has a high mode share to central London, so future growth is dependent on growth in total travel demand, not mode shift".[119]

DEVELOPMENTS IN TECHNOLOGY AND WORKING PRACTICES

101. The demand forecasts were criticised for not taking into account developments in technology and working practices that may reduce demand for rail travel. Dr Wellings said that "the possible impact of disruptive technology is an enormous risk."[120] Councillor Tett of the 51M alliance thought the forecasts contained "little or no assumption about the growth of new technology".[121] Other witnesses felt that technological developments would be complementary to rail travel. The effects of high-speed broadband and driverless cars were the two main developments mentioned by witnesses.

High-speed broadband

102. Tonge & Breedon HS2 Action Group thought that as "electronic systems of work and work behaviour will continue to evolve so that the concept of physical travel … will be out-dated."[122] Dr Wellings told us that "teleworking and remote meetings" could reduce demand for rail.[123]

103. Professor Nash however thought that "there is no evidence of improved communications reducing demand for travel" adding that "if anything demand is growing faster, not slower."[124] Professor Vickerman said that the evidence was "generally that travel and other communications are complements rather than substitutes", citing the ease with which travel can now be booked online.[125] Lord Adonis similarly saw high speed broadband as "complementary" to rail travel: "it is not an alternative. The evidence is that superfast broadband does indeed generate more local working and home-working, but it does not stifle growth in demand for travel."[126]

104. Mr Scott from Virgin Trains believed demand would continue to increase: "it may be dented if we have a superfast broadband network everywhere, but people will still need to get to work, will still need to do business deals face-to-face and will still need to go to football matches."[127] Mike Blackburn representing the North-West Business Leadership Team said that:

128

Driverless cars

105. Several witnesses said that the effect that the development of driverless cars would have on rail demand had not been considered by the Government. Bruce Weston from HS2 Action Alliance told us that "if people found that they could work and do what they wanted as the car whisked them along to wherever they wanted to go, the trends that we have been looking at are going to see a reversal."[129] Mr Stokes thought this was a possibility but also saw how driverless cars could complement rail:

130

106. The Society of Motor Manufacturers and Traders (SMMT) told us that the Automotive Council was targeting mainstream introduction of autonomous vehicles from 2020, noting "autonomous features are already a readily available aspect of many new vehicles on the UK's roads today."[131]

107. The SMMT said that it was "too early to be able to predict the impact of increased connectivity and the introduction of autonomous vehicles on road or public transport". They noted how changing driving trends and models of vehicle ownership and "the concept of purchasing mobility rather than the traditional concept of purchasing vehicles demonstrates the growing interconnectedness of transport, reflecting the needs of travellers and drivers." It was uncertain whether these trends would reduce the number of vehicles on the road, but it was "clear that opportunities will be created by the introduction of connected vehicles across all sectors."[132]

108. When we asked whether driverless cars had been taken account of in the Government's forecasts, the Secretary of State replied that "what we have factored in is what we have seen happen … but exactly what the transport picture will be in 25 years' time is anyone's guess."[133]

109. Partial information on current railway usage, as well as uncertainty about future technological developments in automotive transport and working habits, makes it difficult to assess the plausibility of the Department's forecasts of future demand for long-distance rail travel.

Forecast demand—'with HS2 scenario'

110. The model predicts the effect HS2 will have on demand for road, rail and air between the 235 zones. As well as predicting the number of people who will switch to HS2 from existing modes of transport, the model also predicts the number of new journeys generated by the existence of HS2. The comparison between the 'without HS2 scenario' and the 'with HS2 scenario' allows the benefits of the project to be assessed. The calculation of the expected benefits of HS2 are considered in Chapter 8.

111. Neither the Economic Case nor the supporting documentation published contains the number of passengers the model assumes will use HS2. It is also not clear how the model predicts which passengers will switch from existing transport modes to HS2 (see Box 3 for an example). We do know that the model assumes that fares on HS2 are the same as on the existing railway; it is unable to model the effect on demand for HS2 if there are differential fares.

112. The rest of this chapter looks at whether demand forecasts for other high speed railways proved to be accurate and considers the assumption about fares on HS2.

Box 3: Access to HS2 stations in the demand model

Some of the proposed HS2 stations, for example Birmingham and Sheffield, are not next to the existing city centre station. Some witnesses thought that the model had not taken proper account of this possible time penalty and therefore the predicted number of passengers who will switch to HS2 was too high. Councillor Tett said that "people who travel do not live in the city centre of Manchester … Most of those journey-time savings, on which they predicate so much of the demand, are completely nullified when you look at where people actually live". Dr Nigel Shepperson, a private individual, wrote that the "business case is flawed because it is not based on real door-to-door journeys … existence of local stations [from the Birmingham area] with direct services to London are likely to outweigh any benefit of a high speed line."[134]

Professor Glaister thought that a "good attempt" was made at modelling this time penalty in the Economic Case.[135] The model allows for a passenger being less likely to choose HS2 over an alternative if the former involves an additional walk or transfer. However, there is relatively little experience of modelling this effect for long-distance passengers. Professor Vickerman said "it is a very tricky issue … because you are dealing with a very large spread of areas over which people are being funnelled."[136]

Demand forecasts for other high speed railways

HIGH SPEED 1

113. High Speed 1 (HS1) describes the 68 miles of high speed railway between London and the Channel Tunnel. This was built in two stages: construction of the first section (Channel Tunnel to Fawkham Junction[137]) started in 1998 and was completed in 2003; construction of the second section (Fawkham Junction to St Pancras) started in 2001 and was completed in 2007. It is used by international and domestic services.

114. International rail services to the continent began in 1994 with Eurostar services using existing rail lines between the Channel Tunnel and London Waterloo. The Eurostar used the first section of HS1 from 2003 and the full line from 2007, the London terminus switching from Waterloo station to St Pancras station.

115. The National Audit Office compared actual demand for international rail services against demand forecasts produced by the Department for Transport and London & Continental Railways (who were awarded the contract to build HS1). The National Audit Office report found that:

138

Figure 5: Forecast and actual international passenger demand on HS1

Source: National Audit Office, The completion and sale of HS1, Figure 2

116. Some witnesses told us that the comparison between forecast and actual demand for international travel show that forecasts could be unreliable.[139] The Department for Transport and London & Continental Railways attributed lower than expected demand to unforeseen developments such as competition from low-cost airlines.[140]

TRAIN &AGRAVE; GRANDE VITESSE (TGV)

117. The demand forecasts for the TGV in France were mentioned by some witnesses. Professor Graham said that when lines were assessed after opening, "what they found is that demand forecasts were always too high".[141] The Cour des Comptes in France[142] published an assessment of the TGV in October 2013. Looking at six lines that had been open for at least 20 years, it found that traffic was on average 24 per cent lower than predicted, with one line (Paris to Lyon) having higher traffic than forecast and five lines having lower traffic (of which one, LGV Nord, reached only half the forecast level of traffic).[143]

Level of fares on HS2

118. A report by SKM Colin Buchanan, which audited the demand model on behalf of HS2 Ltd, explained that the model was not able to test the effect of differential fares on demand:

144

119. Mr Prout told us the assumption that HS2 charges the same fares as the existing network, "is to maximise use of the railway, not maximise revenues."[145] Ms Munro explained that "you want people to use that capacity and that is the philosophy that underlies our assumption that you would not charge premium fares."[146]

120. Sir David Higgins, Chair of HS2 Ltd acknowledged that there were arguments for providing facilities such as conferencing and dining cars at higher prices, but that "when you build a train service that will have 18 train paths an hour[147], and each train can have 1,000 people in it. You want to fill it."[148]

121. Professor Glaister described the setting of fares on HS2 as a "dilemma":

149

122. Lord Adonis said that the view of his advisers when he was Secretary of State for Transport was "that the railway should not be built as a premium-cost railway, but equally I took the view that the pricing policy and strategy on the line was going to be a matter for a later day."[150]

CHARGING A HIGHER FARE TO TRAVEL ON HS2

123. Some witnesses questioned the logic behind the decision that fares should be the same. Mr Rukin of Stop HS2 told us that "the ridiculous thing is that they are saying that people will be willing to pay for quicker journeys but base it on the idea that the costs of tickets will be exactly the same as the current ones."[151]

124. Mr Prout told us that the Department had undertaken some rough modelling of the effect on demand of increasing fares. He said that the modelling found that different fare levels "came out roughly even" in terms of revenue: "If you increase the fares, you will reduce the demand, you will increase the payment per person." When asked if the operator would be allowed to charge a premium fare, Mr Prout replied that "the franchise conditions have not been determined yet."[152]

ALLOWING OPERATORS ON THE EXISTING NETWORK TO COMPETE ON PRICE WITH HS2

125. The franchising model will determine whether HS2 will be able to charge higher fares. It will also determine whether train operators running services on the existing network will be able to compete with HS2 for passengers by offering lower fares.

126. The Strategic Case acknowledged that "the introduction of HS2 services will require a major change to the structure and scope of rail franchises compared with those currently operating."[153] Some witnesses, including Cheryl Gillan MP, assumed that HS2 would operate in a franchise in competition with the East Coast and West Coast Main Lines: "I would assume that HS2 … should operate as a franchise, and unless the Government is willing to fix the price on HS2 at the same level as the [West and East Coast Main Lines] (which would seem unlikely) that there will be price competition."[154]

127. 20 Miles More however argued that a franchise model that allowed competition between HS2 and the West and East Coast Main Lines could lead to a reduction in services and higher fares in some areas. They suggested an alternative franchise model where "the [HS2] operator would be responsible for delivering a predetermined level of service and fares with payment on the basis of key performance indicators."[155]

128. The Government should undertake further modelling of the effect of charging premium fares for HS2 services and the effect of competition from other operators on demand forecasts for the new high speed railway. The results of this work would increase understanding of the implications for the funding of HS2.

Chapter 3: Conclusions and recommendations

129. Partial information on current railway usage, as well as uncertainty about future technological developments in automotive transport and working habits, makes it difficult to assess the plausibility of the Department's forecasts of future demand for long-distance rail travel. (Paragraph 109)

130. The Government should undertake further modelling of the effect of charging premium fares for HS2 services and the effect of competition from other operators on demand forecasts for the new high speed railway. The results of this work would increase understanding of the implications for the funding of HS2. (Paragraph 128)

89 Strategic Case, p 74 Back

90 The model considers 14 factors: population growth, employment growth, GDP growth, National Rail fares, London Underground fares, car ownership, car journey times, car fuel prices, bus and coach fares, bus and coach journey times, bus and coach frequency, domestic air fares, domestic air frequency and domestic air passengers. Back

91 The factors are adjusted for each zone-to-zone forecast; for example the forecast for journeys between London and Manchester will take account of local conditions such as expected population increases in both areas. Back

92 HS2 Ltd, The Economic Case for HS2, October 2013, p 32: http://assets.hs2.org.uk/sites/default/ files/inserts/S%26A%201_Economic%20case_0.pdf [accessed February 2015] Back

93 HS2 Ltd, The Economic Case for HS2 PFMv4.3: Assumptions Report, October 2013, p 8: http://webarchive.nationalarchives.gov.uk/20141027142236/http://assets.hs2.org.uk/sites/default/files/inserts/SA%2020_PFM%20assumptions%20report_V3_0.pdf [accessed February 2015] Back

94 Q4 Back

95 Atkins were appointed to develop a "demand forecasting framework" for HS2 in 2009. Since then, the model has been updated a number of times, the latest version (version 4.3) being developed for use in the 2013 Strategic Case. Back

96 The Secretary of State provided the same zone-to-zone movements to the Chairman in a letter on 14 December 2014. We have however used the figures from the Atkins report here as these are the inputs used in the latest demand model; the figures given by the Secretary of State (which showed higher zone-to-zone movements in the base year and forecast year) were inputs from an earlier version of the model in 2012. Back

97 The total number of trips in both directions. Back

98 The average annual increase implied by the 2036 figures is our calculation. Back

99 Economic Case, p 31 Back

100 Q67 Back