Examination of Witnesses (Question Numbers
20-39)
DEPARTMENT FOR
TRANSPORT AND
OFFICE OF
RAIL REGULATION
Q20 Austin Mitchell: But he clearly
does not know how to price the risk in terms of giving every customer
a seat in peak-time services. Why isn't that a requirement on
them when they take the contract? Are you so scared that you won't
get bidders that you don't want to enforce—
Robert Devereux: That I won't
get what, sorry?
Austin Mitchell: Are you so scared that
people won't bid for the contract—GNER gave it up, didn't
they?—that you are not prepared to enforce that kind of
term on them?
Robert Devereux: No. Let's just
look at the capacity of the British rail network. It is an extremely
heavily used network. I don't think it is beyond the—it's
not going to be—
Austin Mitchell: Yes, but usage has been
growing for well over a decade and here we are still cattle transporting.
Usage has been growing for over a decade.
Robert Devereux: It has, so we
now have probably the most extensively used passenger rail network
in Western Europe and it would be delightful to be able to lay
on services in which everybody who turns up gets a seat. The cost
of doing that is very, very substantial.
Q21 Austin Mitchell: Yes, but it
is an expectation that customers will travel in reasonable comfort,
isn't it?
Robert Devereux: Yes.
Austin Mitchell: And they are not, particularly
in London—and I'm not bothered about London, particularly.
If people are daft enough to live down here, they deserve everything
that's coming to them, but people on the East Coast Line are much
more important and they are not travelling in reasonable comfort.
Robert Devereux: I'd just as soon
not agree on who is more important than somebody else. We have
tried to establish, as the Chair has just done, that the current
level of crowding is already something we are trying to tackle,
but actually there are so many people who want to come into London
into the next—
Q22 Austin Mitchell: Yes, but the
companies are saying to you that, to quote page 32, paragraph
3.11, they argue that relieving crowding "generates little
new fares revenue". Therefore, they will not do it unless
they get Government support, which you are providing in this extra
subsidy. Now, why are they allowed to get away with that argument?
Robert Devereux: Well, they're
not. If it told you that, consistently, the proposals that the
train operators provide for us in order to achieve the plans that
we have been discussing earlier for these 950 carriages come at
a price that my negotiators are significantly reducing, perhaps
you'd believe that they are not "getting away with it".
Q23 Chair: Can I just ask you to
bring this bit to an end, and then I will move to James. I just
want to ask one question if I may. The Chiltern contract does
look like a more sensible way forward. Are you attempting—it's
almost a yes or no—and I completely understand you will
be fiddled on the price you then have to pay, but are you attempting
to negotiate that sort of contract within all the other franchises?
Robert Devereux: What we have
done is put out a consultation about could that be a sensible
way forward.
Q24 Chair: And when are you taking
a decision?
Robert Devereux: Consultation
closes on 22 October. Ministers will then think about it, but
we have two franchises which we need to get on and let, which
Ministers want to do on a new basis. Those decisions will therefore
be in the autumn.
Q25 Chair: And will you be trying
to negotiate that into them—obviously subject to price and
everything—but would you able to try—
Robert Devereux: Subject to Ministers
making decisions that that package of things is the most effective
way forward, because it does actually address the points that
the NAO have made—
Chair: Okay.
Robert Devereux: —but, yes,
that is what we are trying to do.
Austin Mitchell: There are other passenger
questions that I would like to ask.
Chair: May I come back to you, because
there is a group of people waiting—
Q26 Austin Mitchell: No, let me ask
this part of the question. The allocation was £9 billion
for expanded capacity, but £7.3 billion of that was going
to infrastructure enhancement such as longer platforms on stations
nobody wanted to go to. Only a smaller part of it—much smaller
part of it—was going to buy extra rolling stock. Now, are
you sure you are not being taken for a ride by Network Rail, which
is using this money for its purposes, rather than for relieving
overcrowding?
Robert Devereux: Yes, very quickly,
half of the number you just quoted is the cost of doing Thameslink.
That is an incredibly complicated infrastructure project, involving
lengthening platforms that, I can guarantee you, people will use.
Austin Mitchell: Well, it says better
stations and longer platforms in this document.
Robert Devereux: Longer platforms
are part of that funding.
Chair: Austin, I am going to give other
people a chance to have a go. James.
Q27 James Wharton: I am rather fortunate
when it comes to rail travel. I represent a constituency in the
North East and I have two main operators to choose from—East
Coast and Grand Central—and I declare my interest: Grand
Central stops at a station in my constituency just up the road
from where I live and I find there is very rarely a problem with
overcrowding on their trains. So, I come to this from a regional
perspective of not seeing a big problem when I set off, but when
I get further south I do see that it becomes more and more of
a problem, particularly on the trains that stop at further stations
as they go south. Then in London, and of course in the South trying
to use the trains, overcrowding is certainly a real difficulty.
It seems that one of your primary solutions to this is going to
be adding carriages on to the end of trains to make them ever
longer. At what stage is that no longer going to be possible and
you are going to have to look for new solutions to this problem?
Robert Devereux: It is the case
that in the current control period—the five years to 2014—one
the ways in which we are attacking capacity is by having longer
trains. You should not assume that that is the only thing that
can be done in perpetuity. So, for example, advances in signalling
that would enable us to run trains much closer together and still
do it safely would fundamentally improve the capacity of the network.
Those are things that we are working on plans for, but they are
simply not of a maturity that enables me to solve my 2014 problem
now.
Q28 James Wharton: I understand that
you have highlighted a couple of other options, and demand management
is something that you have actually said is something you do not
want to look at because of the disproportionate impact where you
find people who can financially afford to overcome demand management
through pricing and so on can do so. If you have dismissed that
as a policy area that you want to take forward—
Robert Devereux: I don't.
James Wharton: You haven't? So it is
something that you will be looking at?
Robert Devereux: It does not say
I don't want to do it. What it says is I do not have the technology
at the moment in place for that to be a solution for 2014, but
I am doing quite a lot of work in changing the franchise terms
to ensure that electronic ticketing is a requirement of all new
franchises; I am putting electronic gates into the big London
termini; I have the Oyster Card working on London suburban services.
My expectation is that actually we will indeed use demand management
as another way of actually managing the network in subsequent
years and we would have to tackle the question that, as it happens,
the evidence suggests that—
Q29 Chair: So can I ask another question
arising out of that? Is there a date by which you will know numbers?
Because one of the things that strikes you as you read this Report
is your lack of information on the future. You do this very odd
manual survey, which seems to be very questionable. Your ticket
sales are not a very good way of judging it and clearly just counting.
When will you be able to do that?
Robert Devereux: Okay. As of today,
39% of all the carriages in the country have automatic counts
on them and, since they are on the more busy lines, that probably
represents a much larger percentage of the journeys. So, relative
to the information that you have here, which is that the manual
count itself is probably only picking up 3%, that is a completely
transformational position. The thing that unlocks that is the
ability to process, really, quite a lot of data—as you would
imagine—from every service. The IT project, which the National
Audit Office refers to, is currently suspended as part of the
reviewing all IT projects by the Government. It is in my view
a really critical project to ensure that we have got raw data,
which is governing, as you have been though, £9 billion of
enhancements. That is, in my view, in the urgent box. I think
it is safe to say—
Q30 Chair: If you get the go ahead,
it is in by when?
Robert Devereux: 2011 is what
the National Audit Office says. That would have been written before
the pause that we are currently going through. So, I haven't got
exactly in my head which part of 2011 it is or whether it is still
2011 if we give it the go ahead.
Chair: Ian wanted a quick question, then
I am going to Jo and then Stephen.
Q31 Ian Swales: Yes, just on Mr Wharton's
point about lengthening trains and so on, when one travels abroad,
one is struck by two things. One, that the level of fares in our
country seems breathtaking. I would not be surprised to be told
that we are the most expensive country in the world to travel
by train. But the other thing—and the point of my question—is
the design of trains. I travelled recently from Antwerp to Brussels
on a double-decker train. Are we actually looking creatively at
how to solve this problem and why don't we, as the French and
others do, have the double-decker trains?
Robert Devereux: On the comparison
of fares, I am afraid I have to just point you to choices that
different countries make as to the subsidy. The British taxpayer,
over the next five years, is putting £15.3 billion into the
railway system in order to subsidise people using it. I could
make the fares lower still by making that a bigger number. I do
not currently have a guarantee of getting more than £15 billion.
Chances are I am going to have to work with less than that. So,
it is a choice, I'm afraid, about how Governments work. That is
the difference. On the double-decker stuff, two things are probably
worth saying. One is that double-decker is—what journey
were you making? Antwerp to...?
Ian Swales: Brussels.
Robert Devereux: Yes. On a long
journey you can actually get something useful out of double-decker.
What you typically don't get with the double-decker is fast entrance
and exits. So, the calculation we have to make is: can I actually
make sure that these trains physically go through the various
tunnels and over the various bridges of which we have hundreds
and hundreds; and, secondly, are they the right sort of train,
given that quite often in the British system I need people to
be on and off really promptly and, in a sense, I don't want to
be having people queuing at the top of stairs.
Chair: Jo.
Q32 Joseph Johnson: Thanks. Well,
to use Austin Mitchell's description, I am a daft Londoner and
represent a London constituency of Orpington. Just taking the
conversation in the round: usage is set to double; capacity is
not going to keep up; therefore, overcrowding is going to get
worse. But I wondered if you could tell us where in the country
overcrowding is going to get particularly bad over coming years.
I ask that because the Southeastern franchise, which I use a lot,
is virtually—well, it feels like it is reaching intolerable
levels of overcrowding at some points. This is particularly unfair
in some ways because the Southeastern franchise has been having
RPI +3 fare increases and will do until the end of next year.
My constituents, for example, have not benefitted one jot from
the extra investment that was the justification for that RPI +3
terms of their franchise, in the sense that it was there to pay
for fast trains, and these fast trains go from London Bridge and
they skip Orpington altogether and they wind up somewhere in Tonbridge
or Sevenoaks or wherever, way down the line. So, my poor constituents
are getting the worst of all worlds because they are left taking
the very slow trains that do all the stops and which end up being
incredibly overcrowded because there are tons of people who are
not going to these hub towns, so I wondered if you could possibly
give an indication of whether it is going to get even worse for
people like them?
Robert Devereux: Okay. Well, the
plan that we are embarked on, subject to the spending review securing
budget that we are working to, is targeting an aggregate average
level of crowding about the same it was when the plan started.
We have illustrated that in the White Paper that the Chair talked
about by reference to individual London destinations but produced
a common sense of what level of average crowding we were after.
So, I am not currently trying to differentiate and pick winners
in where crowding might be. Subject to the spending review, I
am trying to make sure that we tackle crowding evenly, as it were.
In practice, because the interventions will be different and your
particular constituency service is going to be slightly different
from something else, there will be changes because at the level
at which I am making strategic choices about a £15 billion
level of investment, I am necessarily having to deal with some
relatively broad aggregates. But it is the case, as you have said,
that we have put very substantial extra capacity into the South
East region—not, sadly, fast trains going through Dartford.
But, trying somehow or other to make sure that some of that benefit
is actually being financed by the users of the service and not
by the taxpayer in Manchester and Birmingham seems to me a plausible
reason for justifying RPI +3. We are taking blocks of the country;
we are not taking individuals. Come the day that I can actually
do quite sophisticated things with ticketing and routes—actually,
a number of permutations start to open up that I simply cannot
do today. Bits of paper as tickets are really quite useless when
it comes to dynamic management of the system. I really seriously
would want to aspire to a world in which I know what is going
on—it comes back to accounts—and I actually have got
the technology and the requirements on the companies to think
really laterally about how they segment their markets. Right at
the moment, you buy a season ticket and it makes very little difference
to you whether you go on Friday, take a fast train, take a slow
train—you have bought it; it's a year; it's gone. It's an
annualised cost and it's just like paying your tax on a car. It
makes a really big difference, as you can begin to see in London,
when people say, "Okay, so if I make one fewer marginal journeys,
I have £1 more than I would otherwise have had." That
simply is not in the British railway system at the moment and
it is absolutely the next critical thing to achieve.
Q33 Stephen Barclay: Just on that,
actually, you just talked about the flaws in data in terms of
identifying future capacity. Is it a specification for all the
new rolling stock that they will have automatic counting, such
as the infra-red sensors?
Robert Devereux: Virtually every
new train—newly built train—comes with that because
everybody can see that that's the case. Within the—
Stephen Barclay: So it is not a specification?
Robert Devereux: I'm sorry?
Stephen Barclay: It is not a specification
from yourselves.
Robert Devereux: I'm not sure
whether I specify that if it is a new train, it should have automatic
passenger count. Could I let you know?
Stephen Barclay: The Report makes it
clear that—
Robert Devereux: The reality is
that it has automatic passenger counts because that's how new
trains—
Q34 Stephen Barclay: Should it not
be a specification? What the Report says is there are problems
with relying on ticket prices. It tends to overestimate journeys
from cities as opposed to rural stations such as my constituency,
and the thing you really want to count is who gets on and off
the train. It strikes me that you are in the process of buying
a whole lot of new trains. Would you not require as a specification
for those new trains they have this facility?
Robert Devereux: I understand.
I am going to make a simple point that I am not sure that the
market for new trains will ever deliver one without one. I will
go back and check—
Chair: And let us know.
Robert Devereux: —whether
that's the case.
Q35 Matthew Hancock: Also, did you
say virtually every train?
Robert Devereux: Well, that is
just because I do not want to say that I absolutely know that
every new train is like it, but can I come back to you? I think
it is the case that new trains that form part of the 900 that
we have plans for will all come with automatic passenger count.
That is the only way in which we have got it up to 39%. Be clear—and
I don't want to mislead you—within that 950 trains for which
we have plans, some of them are reusing stock that is already,
as it were, spare, so they are not of themselves new trains.
Q36 Stephen Barclay: And you are
not going to retrofit them because of cost?
Robert Devereux: Well, the cost
of retrofitting some of that is quite high.
Q37 Chair: High relative to—you
are talking about a £9.2 billion programme. It seemed to
me reading it, I thought, "For goodness' sake; get the proper
data so that you can make proper decisions." What is high?
Robert Devereux: Yes, that is
exactly what I thought when I read that sentence too. It is conceivable
that actually if—I tried to ask roughly what sort of price
are we talking about here, and by the time you multiply that across
all the carriages we have not got, I think you could easily be
talking about many tens of millions of pounds worth of cost.
Chair: That's still relative to £9.2
billion. It may be an appropriate investment.
Robert Devereux: It may, but if,
for example, with 40% of the carriages, which by the way is increasing
as I buy new ones, I get a very substantial coverage of passengers,
I might not get huge value out of that marginal 60%. I think it
is an entirely fair question and actually it is one that we need
to think about, but right now, moving from having 3% of the data
to somewhere well north of 40% will fundamentally transform my
knowledge. The question you should ask me then is, if I were to
retrofit all the rest, would I be getting further and better particulars?
I might do; I am very happy to take it away and think about it.
Q38 Stephen Barclay: Well, a related
issue as well is the Report highlights problems with your modelling
and on page 17 says you "do not test the sensitivity of your
forecast to variations in the relationship between economic growth
and rail demand."
Robert Devereux: Yes.
Stephen Barclay: Why don't you do that?
Robert Devereux: Well, if I'm
honest, I thought paragraphs 1.15 and 1.16 were accurate, but
I am not sure if you are actually running a model that is trying
to get you to some strategic conclusions—they leave you
with the impression that somehow we have not done the right thing.
On the question of the relationship between passengers and GDP,
as a matter of arithmetic, that is exactly the same as making
bigger assumptions about GDP.
Stephen Barclay: It is more than that
though, surely, because I remember Iain Coucher saying to me that
the average time that a commuter was prepared to spend on the
rail element of their commute 10 years ago was around 45 minutes
and it is now an hour and a quarter; I don't know if that is correct—that's
what he said to me. That is very relevant to my constituency because
we are less than 100 miles from London, so it brings us into play.
Yet, the new Thameslink trains are stopping at Cambridge, so we
have a bizarre situation that those new trains stop there and
yet those new trains have the selective door opening that actually
allows them to stop at the rural stations that have short platforms,
which many of the existing trains cannot. But your modelling is
not picking up the increasing demand in the county with the highest
growing population, where there is increasing demand in Cambridgeshire
to use stations such as Littleport, because your modelling is
overestimating the city transport, and at the same time you are
saying, "Ah, but we just look at it from a GDP point of view.
You are not looking at it in the round."
Robert Devereux: No, sorry. I
am making two separate points. If I take a sensitivity analysis
that says, "Here's my base case. What happens if GDP is lower
or higher?" I am effectively doing something to demand and
I can see how that drops through the bottom. Broadly speaking,
if I change the relationship between a given GDP that I started
with and passenger numbers, I am also dropping it through to demand.
So, both of those variables are effectively being tested if I
test a wide range of GDP. That is all I am saying. The second
thing I just want to say is that if I told you that the model
we have, which, as the NAO at least gives me, is based on established
building blocks as well as new, innovative work, has to run all
through the night to give me one answer because the rail system
is so complex and the level of sophistication in it requires that
much computation, you might understand why it is that I have to
be moderately cautious about how many variants I can test. Even
now, the computational requirement to allocate as many people
as possible to the right routes and get some answers is quite
difficult to do. So, I am absolutely up for being encouraged to
do better every year and that is what we are doing, but the sorts
of things that are being talked about here are, in my judgment,
the sorts of simplifications that go with a model that none the
less can actually generate some answers with which you can then
affect policy.
Q39 Nick Smith: On demand modelling,
I wonder if I could just jump in? In my constituency, Blaenau
Gwent in South Wales, we have just got a new train line—absolutely
fantastic—from Ebbw Vale down to Cardiff and within six
months we were breaking all the records for participation on the
railways; it's an absolutely tremendous success. Unfortunately,
when the line was built, only one track was laid and now, of course,
we have realised we need two tracks because that would make a
big difference in terms of being able to get down to Cardiff from
the valleys. Unfortunately, it is way too late. If the track had
been put in at the same time as the first track it would have
been much cheaper. So, now there are greater costs and people
are very frustrated because the shine has been taken off a great
success because of the very poor modelling of demand. What are
you doing about demand at that sort of level?
Robert Devereux: These are comments
by the National Audit Office about the planning for quite a big
White Paper—how would I spend £15 billion/£9 billion
sensibly. That necessarily is not the same as—
Nick Smith: I know it is a different
sort of model, but—
Robert Devereux: —should
Arriva Trains Wales and the Welsh Assembly think in great detail
as to whether it is one track or two tracks in your part of the
world. And so, we are talking about two somewhat different things.
I would aspire to a world in which my model was so sophisticated
it knew whether there should be one track or two tracks in your
world. All I can tell you is that these models are already fantastically
difficult to manage and get to the right answers. As computing
power improves, which is the reality with everything to do with
modelling, we will be able to do better and better. This is a
snapshot of where I am today.
| 
