Written evidence from Dr Paul Hoad (HSR
198)
SUMMARY
1. This paper has been written by Dr Paul Hoad
who has 16 years' experience in Transport Modelling and Planning.
Dr Hoad has worked on a wide range of transport projects, in the
UK, East Asia and the Gulf, including the economic assessment
of transport schemes.
2. This note considers a number of issues relating
to the transport modelling work undertaken for the HS2 study and
calculation of economic benefits. Documents provided at the public
exhibitions plus further documents and Spreadsheets available
from the HS2 website have been the basis for this review.
3. The general approach to the assessing the
scheme has a number of problems including the following:
— The
choice of Euston as a London Terminus is poor as the connections
to other rail services will be irrelevant.
— The
costs do not include improvements to London Underground even though
HS2 recognise the need for such work.
— The
choice of Curzon Street as the Birmingham Terminus is poor owing
to the poor connectivity and was abandoned as a passenger terminal
150 years ago as it was located too far from the city centre.
— The
HS2 model has only been used to produce forecasts for 2021 and
2043, despite the fact that the opening year is 2026 with subsequent
improvement in 2033. The ability to understand the requirement
of the system and design accordingly is therefore in question.
4. From the available reports it is evident that
the model used HS2 to determine the transport benefits has serious
weaknesses. This evidence includes warnings from HS2's consultants
about the limitations in the model which do not appear to have
been addressed.
5. Weaknesses in the model include the following:
— The
benefits claimed may merely be due to random noise in the model.
— There
has been lack of observed data upon which to build the model.
— The
increase in overall rail patronage forecast is likely to build
up much more slowly than anticipated.
6. Furthermore, due to the availability of Spreadsheets
used by HS2 to derive the economic benefits of the proposals,
it has been possible to carry out a detailed review of the calculations
involved. From this it has not only been possible to identify
a number of errors in the methodology but also to correct and
repeat the calculations used by HS2 and produce revised Benefit
Cost Ratios. These errors include:
— An
over-estimation of the number of working days in a year.
— The
cost of paying for tickets has not been included when calculating
the transport benefits even though the resulting Revenue is recognised
as being of significant size.
— Artificial
time penalties applied to replicate travel behaviour have been
included in the economic benefits as if they were real times.
— The
Value of Time for Rail Business Passengers is too high.
— The
average occupancy of cars has been overestimated thereby over-estimating
the Value of Time for cars.
7. In each case the resultant Benefit/Cost Ratios
show a decrease, demonstrating that the benefits have been over-estimated.
In particular correcting the failure to include the cost of ticket
purchase seriously weakens the economic argument for HS2. Taken
together it is demonstrated that there is no net economic benefit
from the system.
8. This paper therefore questions the validity
of the model used to support the proposed High Speed Rail, and
demonstrates that the Benefits derived from its construction will
be outweighed by its Cost, at best giving only 70p back for every
£1 spent.
1. INTRODUCTION
1.1.1 This document has been written by Dr Paul
Hoad, who has 16 years' experience in Transport Modelling and
Planning. Dr Hoad has worked on a wide range of transport projects,
in the UK, East Asia and the Gulf, including the economic assessment
of transport schemes. The evidence set out in this report is
based upon the scrutiny of documents provided by the High Speed
Rail Public Consultation, some of which were available at the
public exhibitions and some of which were available via the HS2/DfT
webpages, with particular reference to modelling issues. Particular
scrutiny has been paid to Microsoft Excel Spreadsheets which detail
the Cost Benefit analysis undertaken by the High Speed Rail consultants
and which are available on the internet.
1.1.2 This work has been undertaken solely by
Dr Hoad, with the purpose of checking that the proposals meet
the necessary requirements to ensure value for money, and has
not been funded by any third party. This note has been written
for submission to the House of Commons Transport Select Committee,
and formatted in accordance with the committee's requirements,
including the need to confine the report to six pages.
1.1.3 This note comprises two main sections;
firstly there are comments on the general methodology of the investigation
into the viability of High Speed Rail; and secondly a more detailed
examination of the calculations used to determine the economic
benefit, making use of the Spreadsheets actually used by HS2 in
their analysis.
2. GENERAL ISSUES
WITH HS2 METHODOLOGY
2.1.1 The following issues have been identified
in HS2's methodology to determine the value for money of their
proposals, and which undermine the validity of the conclusions.
2.1.2 There has been a failure to include the
significant costs associated with improving Euston Underground
Station even though HS2 have recognised that the service would
add 5,500 extra passengers to the Tube in the three hour morning
peak (Para 5.4 of the HS2 Consultation document) equivalent to
six completely full Tube trains and that the Tube services would
already be heavily crowded.
2.1.3 There is a lack of information on where
the "Wider Economic Impacts" would occur and how the
forecasts have taken into account constraints such as Green Belt
designation, or whether this would result in unwanted urbanisation
of rural areas. Similarly, the HS2 model forecasts a significant
increase in trip making, but no information is given about where
these extra trips would come from, again meaning that the model
results could be based upon increased development in unreasonable
areas.
2.1.4 The transport model used by HS2 has been
used to produce forecasts for only 2021 and 2043, yet the proposed
opening date is 2026. Therefore the ability to design for the
opening years of the new system is weak and the approach is contrary
to DfT Guidance on the need to provide forecasts for the opening
year (WebTAG Unit 3.5.4 Para 4.2.3).
2.1.5 There is a lack of information on the viability
of the West Coast Mainline once passengers switch to HS2, which
even with reduced services to reduce costs may no longer be economic
to run (particularly in the years immediately following the opening
of HS2.
2.1.6 The documentation for the HS2 model indicates
that there is poor convergence between the supply (the available
network and services) and the demand (ie the trips) contrary to
DfT Guidance (WebTAG Unit 3.10.4, Section 1.5), meaning in real
terms that the model is subject to a high degree of randomness
and therefore any benefits may simply be random noise rather than
real impacts.
2.1.7 The documentation for the HS2 model indicates
that it is recognised that there is a lack of observed data upon
which to base it, making its reliability as a forecast tool open
to question.
2.1.8 The Highway component of the HS2 model
is unreliable due to its inclusion of fixed traffic flow values,
making it unsuitable for forecasting future flows and the subsequent
derivation of economic benefits.
2.1.9 The HS2 model does not include the representation
of local road congestion, which can be a key factor in choosing
between road and rail travel.
2.1.10 There is no evidence of First and Standard
Class Passengers being dealt with separately in the model, and
hence its ability to properly deal with capacity issues is seriously
in question as it will not be able to reflect how Standard Class
can be full yet First Class can be almost empty.
2.1.11 Although HS2 should be commended for identifying
that the increase in trip numbers forecast by the model will not
appear instantaneously from opening, the assumption of 80%, 90%
and 95% of this extra demand for the first, second and third years
respectively is still likely to be optimistic, over estimating
both the travel benefits and the additional revenue gained.
2.1.12 In a report dated February 2010, Atkins
advised HS2 "When the project proceeds to a detailed design
and public scrutiny stage, it is expected that additional independent
validation data would be collected to allow further and more detailed
scrutiny of the model and inform detailed design decisions."
However this advice does not appear to be followed as the development
of the HS2 model is detailed in a report also dated February 2010.
Despite this, the public consultation exercise was still carried
out.
2.1.13 The justification for High Speed Rail
set out in the consultation is based strongly on the argument
that the UK should have HSR2 simply because other countries have
invested in it. However there is no evidence presented to demonstrate
the economic case of these systems, indicating that this is simply
a matter of jealousy of other countries rather than rational argument,
and that therefore the study carried out is a matter of attempting
to prove a decision that has already been taken.
2.1.14 The choice of Curzon Street as the site
for the Birmingham terminus is poor due to its lack of interchange
with conventional rail and bus services, and was abandoned 150
years ago as a passenger station owing to its distance from the
city centre.
2.1.15 The choice of Euston as the site for the
London terminus is poor as it connects to the Underground via
a station that is already overcrowded. There will be little valid
connections with conventional rail services as few people are
likely to use the HS2 line from Birmingham to Euston then travel
back northwards on the West Coast Main Line.
3. ANALYSIS OF
HS2 APPRAISAL SPREADSHEETS
3.1 Overview of the HS2 Appraisal Spreadsheets
3.1.1 Twenty four Microsoft Excel Spreadsheets
used to assess the economic benefits of the High Speed Rail are
available on the internet via "http://www.hs2.org.uk/supporting-documents-temp".
These Spreadsheets are for scenarios examined. The files of primary
interest are listed below.
| File Name | Scenario Modelled
|
| Central Case MM67b Day 1 (revised 060711).xls
| The initial London-Birmingham network |
| Central Case MM88 Y Network (revised 060711).xls
| The London-Birmingham network plus extensions to Manchester and Leeds
|
3.1.2 The final figures in these Spreadsheets are very close
to, but not exactly the same as, the economic figures provided
in the DfT/HS2 document "Economic Case for HS2, The Y Network
and London-West Midlands", February 2011 provided at the
HS2 exhibitions. The minor differences would presumably be due
to the implied revisions made on 6th July. (NB The identities
of the files are also confirmed in document "http://hs2.org.uk/assets/x/78218"
from HS2, in a Freedom of Information response, which makes reference
to these file names, although without the revised date suffix.)
3.1.3 These two Spreadsheets have been reviewed and a number
of items have been identified for which there has probably been
an error made in the calculations. Corrections have therefore
been applied to the Spreadsheets providing updated results for
the economic analysis. These issues are set out below in the
form of tests carried out by the author, examining the effect
of each correction individually, including the effect on the Benefit
Cost Ratio (BCR) of each as derived from the corrected Spreadsheet.
In addition a final test is detailed where the corrections were
applied together to assess their combined impact.
3.1.4 In accordance with the data presented in the economic
report, two BCR values are given, both with and without the Wider
Economic Impacts (WEIs).
3.2 Test 1: Too High Annualisation Factors
3.2.1 Annualisation factors are values used to convert the
forecast Daily values taken from the model into total Yearly values
(and ultimately the 60 year assessment period). The values used
in the Spreadsheets are set out below, although it should be noted
that values under "HSL" do not appear to have been used,
with Rail covering both High Speed Rail and conventional rail.
| Journey Purpose | Rail
| HSL | Air |
Highway |
| Business | 273 | 273
| 291 | 365 |
| Commute | 279 | 279
| n/a | 365 |
| Other | 353 | 385
| 395 | 365 |
3.2.2 The factors used for Business and Commute are too high.
These trips are weekday trips, and not likely to occur during
the weekend. As set out in the guidance for DfT's TUBA (Transport
Users Benefit Appraisal) Software, which would normally be used
to undertake the economic assessment and which the Spreadsheet
should replicate, there are 253 working days in the year (365
days, minus 52 weekend days and eight Public Holidays).
3.2.3 Business and Commute Annualisation factors were therefore
corrected by setting them to a value of 253. The factors for
Other trips were kept the same. Although they appear very high,
these are not unreasonable if the need to consider time saving
during the unmodelled weekend is to be included. However, there
is an additional question of whether time savings during the weekend
can be considered as valid for inclusion in the economic benefits
(it is not unusual for economic assessments to exclude the weekend
period completely). Therefore a further test was carried out
setting all factors to 253 to derive the transport benefits, but
setting the total revenue to equal that calculated when only the
Business and Commute factors were corrected. This latter element
is important as the Annualisation factors are used to derive the
total annual revenue, and even if Weekend time savings are assumed
not to contribute to the economic benefit, the passengers are
still paying the fares. The results of the testing can be seen
below:
| Test | Network
| with/without WEIs | Uncorrected
BCR
| Corrected
BCR | Change in
BCR
|
Business and
Commute Factors
set to 252
| London—Birmingham | BCR without WEIs
| 1.61 | 1.45 | -10.1%
|
| | BCR with WEIs
| 2.01 | 1.82 | -9.4%
|
| Y Network | BCR without WEIs
| 2.20 | 1.96 | -11.0%
|
| | BCR with WEIs
| 2.59 | 2.35 | -9.2%
|
Business,
Commute
and Other
Factors set to 252
| London—Birmingham | BCR without WEIs
| 1.61 | 1.34 | -16.9%
|
| | BCR with WEIs
| 2.01 | 1.71 | -14.8%
|
| Y Network | BCR without WEIs
| 2.20 | 1.83 | -17.0%
|
| | BCR with WEIs
| 2.59 | 2.20 | -15.1%
|
3.2.4 From the results it is clearly evident that the correction
of the Annualisation factors does have a significant effect upon
the BCR values. This impact is in greater proportion than the
actual change in the factors, owing to the fact that not only
are the forecast annual transport benefits reduced but also the
forecast annual revenue decreases thereby increasing the net cost
to government.
3.3 Test 2: Absence of User Charges
3.3.1 Examination of the HS2 Spreadsheets has found that the
User Charges included in the analysis are in likelihood incorrect
which has resulted in the Transport Benefits being seriously overestimated.
User Charges cover financial payments paid by travellers and includes
such elements as road tolls and, more importantly for HS2, rail
(and air) fares. From Tables 2 and 10 of the "Economic Case
for HS2" report it is clear that significant additional revenue
is expected to be raised. Yet despite this there are no corresponding
values in the User Charges to show travellers having to make these
payments. Such charges effectively offset part of the benefits
gained and can have a material effect on the economic assessment.
3.3.2 An example of this is the two existing alternative rail
services between Birmingham and London: the New Street - Euston
service and the Snow Hill - Marylebone service. The latter is
the slower of the two but is cheaper. For someone currently travelling
on the Snow Hill - line, it would be possible to transfer to New
Street and gain a time saving but will suffer the loss of the
increased fare, and so the choice is a trade-off between the time
saved (and the Value Of Time of the traveller) and the additional
financial cost.
3.3.3 Reviewing the Spreadsheets has shown that there are
calculations of User Charges for Rail passengers but the values
being dealt with are so small as to be zero when rounded to the
nearest million pounds in accordance with the other larger values
considered. For example, for the "Day 1" Birmingham
- London scenario, the total Revenue over 60 years (in 2002 prices
and discounted to 2002) is £7,753,695,520 but the corresponding
total User Charge is a mere £1,657. For the "Y"
network the values are £15,444,335,550 and £859 respectively.
The User Charges are therefore more than a million times smaller
than the revenues despite the fact that the Revenues must derive
from the User Charges. In addition to this, reviewing all the
available Spreadsheets shows Revenue values varying quite widely
but the User Charges staying relatively fixed. For example, the
forecast 2021 annual revenues vary between £107 million and
£451 million, whilst the User charges vary between £33
and £97.
3.3.4 The cause of this error is suggested by Item 8 of the
minutes of the HS2 Analytical Challenge Panel for Thursday 23
April 2009. (NB On the HS2 website this is incorrectly identified
as 2011.) In this item it is stated that the revenue modelling
should be included in the benefits modelling. It is therefore
probable that the calculations were updated to add the Revenue
into the calculations but the corresponding need to include the
User Charges was overlooked.
3.3.5 In order to correct this, a simple assumption has been
made to correct the Spreadsheets, setting the User Charges equal
in size to the Revenues although they are set as negative values
as they are a cost to travellers. In the original Spreadsheets
the User Charges were added to the Transport Benefits, implying
the money was being paid out by HS2 to passengers, not being paid
by passengers. Having applied this change to the Spreadsheet
a further correction was found to be necessary in the "Y
Network" calculations as the cost of the User Charge was
being applied twice.
3.3.6 It should be noted though that the User Charges should
be higher than the Revenue. Travellers accessing the stations
by taxi will be paying out taxi fares which will not be recouped
as Revenue. Car parking costs will also be incurred which may
not be recouped by the government. Even if HS2 operates its own
car parks in order to capture these payments, this will still
miss out those users of the conventional rail system who use car
parks not controlled by Network Rail, and consideration would
also need to be given to the additional costs of operating such
car parks. It can therefore be seen that the setting the User
Charges to equal the Revenue is a conservative assumption.
3.3.7 The results of the testing can be seen below:
| Network | with/without WEIs
| Uncorrected
BCR | Corrected
BCR
| Change in
BCR |
| London—Birmingham | BCR without WEIs
| 1.61 | 0.56 | -64.9%
|
| BCR with WEIs | 2.01
| 0.91 | -54.6% |
| Y Network | BCR without WEIs
| 2.20 | 0.95 | -57.1%
|
| BCR with WEIs | 2.59
| 1.36 | -47.4% |
3.3.8 As can be seen, there are very significant decreases
in the BCR owing to the considerable size of the Revenues and
consequently also the User Charges. Although the magnitude of
the Revenues provides a major benefit of paying for roughly half
the cost of Capital and Operating Costs, they also have the disadvantage
of decreasing the net transport benefits. Although many people
will have the benefit of shorter journey times at the same fare
level, new travellers and those transferring from road (who are
the ones providing the additional revenue) will be gaining less
than was originally calculated.
3.3.9 For the initial London-Birmingham network the total
benefits are less than the actual cost, even with the WEIs. As
result it is not a case of considering the relative value for
money but a case of not being value for money at all. For the
enlarged "Y Network" the scheme appears more marginal,
with the WEIs providing sufficient additional benefits to push
the scheme over the 1.00 level. However the original testing
this scenario by HS2 assumes that the full network opens all together,
rather than having only the initial (loss making) London-Birmingham
for the first seven years. In addition, as was reported in the
"Economic Case for HS2", delaying implementing the scheme
actually increases the BCR, by roughly 5% a year. By ignoring
the first seven years (2026 to 2033) when only the London-Birmingham
section is in operation the BCR of the "Y Network" would
have therefore been increased by 40%. Correcting for this is
therefore likely to reduce the BCR with WEIs from 1.36 to below
1.00 and hence in all scenarios the scheme would be an economic
loss.
3.3.10 The calculation of economic benefits includes the effect
upon tax revenues. Additional money that is paid out by travellers
would have otherwise been spent elsewhere and would have been
taxed to some degree (whereas rail fares are not). The calculations
in the Spreadsheets do include this aspect but this is linked
to the Revenue input whereas it should be linked to the User Charge.
Although this means that the error of having too low a value
has been avoided, given that User Charges are likely to be higher
than the Revenue, this would still need to be corrected for.
DfT's TUBA software uses the User Charge as an input from which
the Revenues (public and private) are then derived, rather than
assuming separate inputs and calculations to derive User Charges
and Revenue, and the Spreadsheets should follow the same processes.
3.4 Test 3: Boarding and Crowding Penalties
3.4.1 In order to determine route and mode (ie road, rail,
air) choice, the HS2 model not only includes actual travel time
but also additional time penalties that affect travel behaviour.
These include Boarding Penalties and Crowding Penalties for Rail
travellers. The former would be a perceived penalty of using
transport service and would be applied for each service is boarded.
This reflects the inevitable preference travellers have for private
car as opposed to public transport and also militates against
a series of connecting services. As the actual travel time, the
time taken waiting for a service and the time taken to access
and exit the public transport system are covered under separate
items, the Boarding Penalty would not be a real time delay merely
a perceived penalty required to replicate passenger behaviour.
However these values have been included in the economic assessment,
meaning that incorrect additional benefits have been claimed for
the proposals.
3.4.2 It is also notable that the analysis claims that there
is an actual overall positive benefit from the Boarding Penalties.
Given that there is an overall increase in Rail usage, and the
HS2 will not reduce the need for intermediate interchanges, it
would be expected for there to be an increase in the number of
Boardings. This in turn should give a negative rather than positive
benefit (if these were to be included) yet a positive benefit
is claimed.
3.4.3 Similarly, Crowding Penalties are applied to reflect
the physical constraints of rail capacity and reflect the decreasing
attraction of rail travel as the level of usage increases and
carriages become more crowded. This should be seen as a perceived
penalty not an actual time penalty and hence it should have been
excluded from the economic assessment. Whilst it might be argued
that the effect of crowding will be to degrade the opportunity
to carry out productive work on a train, this would be to ignore
the fact that the Value of Time for Business users on rail is
already much higher than for car users, so to add further penalties
would be double counting. The level of Crowding has however been
included in the economic assessment, meaning that incorrect additional
benefits have been claimed for the proposals.
3.4.4 The Spreadsheets have therefore been corrected by applying
zero values to both the Boarding and Crowding time penalties.
The results of this test are shown below.
| Network | with/without WEIs
| Uncorrected
BCR | Corrected
BCR
| Change in
BCR |
| London—Birmingham | BCR without WEIs
| 1.61 | 1.37 | -15.1%
|
| BCR with WEIs | 2.01
| 1.70 | -12.4% |
| Y Network | BCR without WEIs
| 2.20 | 1.91 | -13.2%
|
| BCR with WEIs | 2.59
| 2.30 | -11.2% |
3.4.5 As can be seen, there are significant decreases in the
BCR values of between 10% and 15%. The bulk of this reduction
is in fact due to the removal of the Crowding Penalties, although
both elements contribute to the lower values.
3.5 Test 4: High Value of Time for Rail and Air Business
Passengers
3.5.1 As has been argued by other parties, the Value of Time
(VOT) for Rail and Air Business passengers is much higher than
for car Business users. The VOT is effectively the penalty paid
for being unproductive during travel, and it can be seen that
someone travelling by train or air is likely to be more productive
than someone in a car. For Business trips the existing car occupancy
is 1.2 (from DfT's WebTAG Guidance), or in other words for every
six cars, five have only the driver and only one has an extra
passenger. Even with high levels of crowding, rail and air passengers
will still be able to get more work done than a driver.
3.5.2 In order to test the impact of the high VOT for rail/air
business users, the values have been adjusted downwards. Two
alternatives are reported below, firstly setting the air/rail
VOT to the same value as for road (a reduction of 31%) then applying
a 50% reduction to reflect an overall lower VOT. A reduction of
50% is to some extent speculative but such a decrease was discussed
in the Analytical Challenge Panel of 30 June 2009. (N.B. Minutes
of these meetings are available at http://www.hs2.org.uk/acp although
there are some errors in these links, including mislabelling the
year of some meetings.)
3.5.3 The results of these tests are shown below:
Rail/Air Business
VOT | Network
| with/without WEIs | Uncorrected
BCR
| Corrected
BCR | Change in
BCR
|
= Road Business
VOT | London—Birmingham
| BCR without WEIs | 1.61 |
1.35 | -16.0% |
| | BCR with WEIs
| 2.01 | 1.73 | -14.1%
|
| Y Network | BCR without WEIs
| 2.20 | 1.85 | -16.1%
|
| | BCR with WEIs
| 2.59 | 2.23 | -13.8%
|
| reduced by 50% | London—Birmingham
| BCR without WEIs | 1.61 |
1.19 | -25.7% |
| | BCR with WEIs
| 2.01 | 1.55 | -22.7%
|
| Y Network | BCR without WEIs
| 2.20 | 1.63 | -25.9%
|
| | BCR with WEIs
| 2.59 | 2.01 | -22.3%
|
3.5.4 As can be seen, this test produced significant decreases
in the BCR values, reducing them by about 15% where there is equality
between Rail/Air and Road VOT and by 25% when the Rail/Air VOT
is decreased by just under another 20%.
3.5.5 It is recognised that changing the VOT would require
the model to be updated accordingly and rerun, rather than simply
revising the Spreadsheets. A lower rail VOT would be expected
to mean a greater attraction for rail services but this is potentially
misleading. Changing the VOT would require the base year model
to be revised not simply the forecasts. The process whereby the
model splits trips between road and rail travel would therefore
be recalibrated to the new values but to the same observed values
as used before, damping out the impact in the future. Therefore
rerunning the model is unlikely to gain any further benefits to
offset the decrease in benefits shown above.
3.6 Test 5: Too High Car Occupancy
3.6.1 Car occupancy is important to the economic assessment
as it is necessary to convert between person trips used in a Rail/Air
model and vehicle trips use in a Road (or Highway) model. In
the Spreadsheet it is stated an average occupancy of 1.6 has been
assumed with a comment stating "Ideally update HW [Highway]
to vary by year/time period as occupancy varies - but impact relatively
small". There is a calculation of vehicle occupancy for
the years up to 2031, but these are incorrect and show too low
a decline, and anyway have not been used in the final calculations.
3.6.2 The level of car occupancy is declining slowly but steadily
and DfT guidance is for this to continue up until 2036. The guidance
gives a decrease in occupancy of about half a per cent per annum,
but over 36 years (the baseline data being from 2000) compounded
year on year this does have an effect. Indeed the fact that it
is included on DfT Guidance (WebTAG Unit 3.5.6) shows that it
should be taken account of.
3.6.3 The economic benefits of the proposals are derived for
a sixty year period, starting in 2026 for the "Day 1"
network and 2033 for the "Y Network". All but a very
small proportion of period will therefore have the 2036 level
of car occupancy. It was therefore possible to derive a correction
factor, comparing the calculated car occupancies for 2036 against
the assumed average of 1.6. The correction factors were as set
out below:
| Journey Purpose | Calculated 2036 Occupancy
| Correction Factor |
| Business | 1.18 | 0.737
|
| Commute | 1.12 | 0.701
|
| Other | 1.58 | 0.989
|
3.7 The results of these tests are shown below:
| Network | with/without WEIs
| Uncorrected
BCR | Corrected
BCR
| Change in
BCR |
| London—Birmingham | BCR without WEIs
| 1.61 | 1.58 | -1.8%
|
| BCR with WEIs | 2.01
| 1.98 | -1.6% |
| Y Network | BCR without WEIs
| 2.20 | 2.16 | -1.9%
|
| BCR with WEIs | 2.59
| 2.57 | -0.8% |
3.7.1 As can be seen, although the correction factors are
significant, the overall effect is fairly small, at less than
a 2% reduction. This is because the Highway benefits are a relatively
small component of the total travel benefits (and a smaller proportion
when the WEIs are included). However, as the Highway model has
issues which may need correcting, it is important that the occupancies
are included as it could have a greater impact with a revised
model.
3.7.2 It should be noted that a change in occupancy would
affect the VOTs assumed in the Highway mode, and hence would require
the network model to be rerun for a definitive test to be applied,
although as with changing Rail VOTs the effect is likely to be
small. However such an update would be necessary in order to
ensure the accuracy of the model.
3.8 Test 6: Combined Test
3.8.1 This test combines the individual tests as follows:
As Test 1: Annualisation Factors for Business and Commute travellers
set to 253.
As Test 2: User Charges added equal to the Revenue, but as negative
benefits.
As Test 3:.Boarding and Crowding Penalties removed from Economic
Benefits.
As Test 4:.Business Value of Time for Rail/Air Business travellers
set to equal Car travellers'.
As Test 5: Future Year Car Values of Time corrected for Occupancy.
The results of this test are as below:
| Network | with/without WEIs
| Uncorrected
BCR | Corrected
BCR
| Change in
BCR |
| London—Birmingham | BCR without WEIs
| 1.61 | 0.00 | -99.7%
|
| BCR with WEIs | 2.01
| 0.30 | -85.0% |
| Y Network | BCR without WEIs
| 2.20 | 0.20 | -91.0%
|
| BCR with WEIs | 2.59
| 0.72 | -72.2% |
3.8.2 As might be expected from the individual tests the BCR
values are significantly lower than the 1.00 break-even point,
the worst case producing no net benefit and the best case giving
only a 72% return on the investment. Therefore the proposals
would not provide any real value for money.
3.8.3 It should be noted that the less stringent of the two
options for Test 1 was applied, as was the case for Test 4 hence
there is the potential for the BCR values to be further diminished.
4. CONCLUSIONS
4.1 It is evident that the model used by HS2/DfT has significant
weaknesses and would need to be improved before it can be considered
reliable. Therefore if the project is to continue in any way,
the focus should first be on improving the model, after which
retesting of the proposals can be undertaken. This would not
in itself be likely to identify any additional benefits but should
provide real confidence in the results. As it has been stated
that delaying the implementation of the scheme would lead to a
higher Benefit Cost Ratio (about 5% extra per year of delay) it
would not be unreasonable to take the time needed to build a robust
model, costing a small fraction of the overall benefit claimed
for the delay. Even if the High Speed Rail proposals are still
found to be impractical, this would lead to the creation of a
more reliable national rail model which could be used to make
better use of conventional rail.
4.2 investigation of the method of calculating the benefits
of the proposed scheme has shown a number of errors, which when
corrected demonstrates that the Benefit to Cost Ratio is much
lower than presented to the public, and that this shows not just
that the scheme is of lower value for money than other potential
investments but that it does not even provide value for money,
having a greater cost than it benefits.
4.3 In conclusion therefore is can be said that based on the
evidence there is no justification to proceed with the High Speed
Rail proposals as the scheme will cost the country more than the
benefits it will produce.
11 August 2011
GLOSSARY OF
TECHNICAL TERMS
| Annualisation Factors | Factors to convert Daily Values derived from the HS2 network into yearly values. The values should reflect what proportion of the total yearly journeys is represented in the actual model. Annualisation factors will vary by Journey Purpose due to the different proportions of travel in periods that are not modelled.
|
| Benefit Cost Ratio/BCR | The sum of the benefits (comprising elements such as changes in travel time, vehicle operating costs) divided by the sum of the costs (comprising elements such as the capital cost of construction and yearly operating costs). A value of 1.00 therefore indicates that the Benefits equal the Costs and the higher the value is the greater the overall return on investment.
|
| Convergence | Most transport models follow an iterative approach in deriving forecasts in order to take account of the interaction between travel demand (ie the number of people travelling) and the transport supply (ie the capacity and connectivity of the network). For Highway trips, increased usage causes congestion imposing additional time penalties; for Public Transport increased usage leads to Crowding which penalises travellers. In a model such as HS2 have used, the supply element is fixed by the user and it is a case of the model deriving the level of demand (ie the number of trips between each origin and destination) as well as the routeing once that demand is set. This is a complex process, where an averaging method or other algorithm is required to allow each iteration to make smaller and smaller changes than the previous one, so that the result converges to a final answer. Poor convergence implies that there are large differences between iterations and it can be difficult to achieve a good level of convergence, particularly in deriving the overall demand. Poor convergence can lead to large differences being observed between scenarios which are solely caused by the fluctuations between iterations but can be mistaken for actual scheme benefits. Achieving convergence within a model can be time consuming and there is the risk of sacrificing convergence in order to produce model results within a short time-frame.
|
| Journey Purpose | The subdivision of travellers into different categories, to reflect different behaviour. In the context of the HS2 model the Journey Purposes are: Business (otherwise referred to as Work) trips which are trips taken during work time and funded by the employer and have a high Value of Time when compared with; Commute trips which are the movements between home and work during the travellers own time; and Other (otherwise referred to as Other Non-work) trips which are the remaining trips made by travellers in their own time.
|
| Mode of Travel | The means of travel, such as Air Rail or Car (or Road).
|
| TUBA | The DfT's Transport Users Benefit Appraisal software, developed to provide a standardised method of calculating economic benefits from transport improvements.
|
| User Charges | Financial costs paid by travellers in the course of their journey, such as rail fares, and tolls. This does not include the cost of fuel etc. for cars which are considered under the category of Vehicle Operating Costs.
|
| Value of Time (VOT) | The opportunity cost of the time that a traveller spends on his/her journey, in effect the value to be gained by the economy for each unit of time saved.
|
| WebTAG | The DfT's website for guidance on the conduct of Transport Studies (http://www.dft.gov.uk/webtag/index.php)
|
| Wider Economic Impacts (WEIs) | Benefits indirectly caused by changes in transport and travel costs, and not included in the Transport Benefit calculations
|
| 
