Examination of Witness (Questions 360-380)|
Professor Andrew Evans
14 FEBRUARY 2006
Q360 Lord Macdonald of Tradeston:
Professor Evans, if you have an evidence-based framework for the
amount of money spent on safety in transport and that can be offered
to politicians and officials in the public sector, to what extent
do you think those politicians should allow pressures, either
from the public, from the media or from other politicians to influence
investment patterns? Do you think they should stand by the evidence-based
framework, backed by experts perhaps like yourself?
Professor Evans: In short, yes, they should.
We should remember that willingness-to-pay is public preferences,
they are elicited in a fairly controlled manner rather than in
the somewhat chaotic manner they might be elicited in the newspapers
after an accident, but essentially willingness-to-pay means that
you are trying to base your valuations on the preferences of the
public, so in that sense the public preferences are counted and
they are indeed the basis on which we are doing it. As far as
the media are concerned, in my view measuring column inches is
not a very good basis for making policies, though it may be difficult
to avoid that, because the media, particularly newspapers, it
seems to me, publish what is newsworthy and, in particular, newsworthy
events are events that happen quite rarely and therefore get undue
attention. I say unduefrom the newspapers' point of view
it is perfectly sensible attention, but from the point of view
of policy-making you have to take the rarity of the events into
Q361 Lord Macdonald of Tradeston:
But you will understand the pressure that elected politicians
are under when they are accused of not listening to what the public
wants and its preference in terms of investment in the area of
safety. Can you think, therefore, of a better mechanism that would
help them defend their evidence-based framework? Should they be
more prominent, more familiar, more accessible before the accident
happens so that people can rely on it more?
Professor Evans: I cannot think of any other
way of doing it. In some ways the trouble is that people are not
interested in this sort of material until an accident happens,
but the more the bodies involved can be explicit about what their
criteria are, the more defensible the position is after an event.
Q362 Lord Skidelsky: Could I ask
a supplementary? Your written evidence suggests that the subjective
value people attach rises with real income. Why should that be?
I could think of a lot of general reasons, but I wonder what you
Professor Evans: I would expect that, simply
because the valuation of many things goes up with real income,
people can afford to spend more on safety.
Q363 Lord Skidelsky: They can afford
to spend more on safety, but this is the value they attach to
a life going up. Why should the willingness-to-pay go up with
real income? I can see why the figure would go up with nominal
income, but I wonder why with real income?
Professor Evans: I am not clear what you are
getting at, I am afraid.
Q364 Lord Skidelsky: Why do richer
people value their lives more than poorer people?
Professor Evans: Because they can afford to
Q365 Lord Skidelsky: Proportionately
Professor Evans: Yes, I would say proportionately
more. We are not talking about lives now, we are talking about
spending on reducing risk by small amounts, and I would say if
you have got more to spend you can reduce your risk pretty effectively
if you can afford it.
Q366 Lord Sheppard of Didgemere:
Can we stay on the subject of costs and what some of the costs
mean? Your written evidence implies that the Department for Transport
values the prevention of a statistical road fatality at some £1.4
million. What actually does that mean? Does it mean that all road
safety improvements that involve a cost per fatality prevented
of less than £1.4 million are actually undertaken, and if
not why not et cetera et cetera?
Professor Evans: We do not have comprehensive,
systematic information about the costs of adopted road safety
measures. We have quite a lot of rather piecemeal kind of evidence,
of which I quoted a little bit in my evidence, so we cannot say
definitely the answer to that question, but there seems pretty
conclusive evidence that there are many safety measures which
could be implemented at a cost of less than £1.4 million
which are not undertaken, so that is the short answer. Why are
things not undertaken? I would say, in the short run, budget constraints
with the safety authorities, particularly local authorities; secondly,
also in the short run, staffing constraints, because a lot of
road safety measures are quite staff-intensive and even if you
had a lot more money you could not use that in the short run because
the staffing levels are determined by the sorts of monies that
have been available at present. With some road safety measures
you run into public acceptability issues. For example, the evidence
seems to be that by having more enforcement cameras we could reduce
road risks at a cost of less than £1.4 million per prevented
fatality. But there is a lot of discussion about whether the public
will accept more safety cameras. Those three, therefore, are the
general reasons. The last question, should not something be done,
I think there is a good case on those grounds to allocate more
resources to road safety, and that has been the case and continues
to be the case for some time. This is an interesting question,
whether it should be up to the point where the costs of preventing
fatalities equal the value of preventing fatalities because of
public sector budget constraints. In general, we do not spend
up to the level where costs equal benefits because we have budget
constraints and also because there is opportunity cost to public
funds. It costs more than a pound to raise a pound of public money.
Q367 Lord Sheldon: Surely, if you
were to spend the £1.4 million for the 3,000 people who die
every year, how much would you reduce the number of accidents
by? The trouble is, you do not know just where to spend this money.
It is a bit difficult to see how these figures are worked out.
Professor Evans: I do not know the answer, but
I think we would reduce it pretty substantially but not to zeroobviously
not to zero. I cannot give you a figure off the top of my head
but one could reduce it over time by a substantial proportion.
Q368 Lord Sheldon: Of course, the
more money you spend, the more likely you are to reduce accidents.
Professor Evans: Indeed.
Q369 Lord Sheldon: But knowing in
advance where that money should be spent is to know in advance
where those accidents are going to occur, which is much more difficult.
It is very much a theoretical exercise, is it not?
Professor Evans: One is reliant essentially
on past experience as to where accidents happen and what the effects
are. I am thinking in particular, when I say the costs are below
the value, of local road safety engineering measures, and there
is quite a lot of experience of them. It is that experience of
the past that gives us an estimate of what the benefits will be.
Safety enforcement cameras would be another example; we have got
quite a bit of experience and they are pretty good value for money.
Q370 Lord Skidelsky: Professor Evans,
your written evidence says that the cost of adopted road safety
measures is less than the VPF, whereas the cost of rail safety
measures is greater. Does that suggest an inconsistency in applying
the principles of risk assessment across different transport sectors
and, to go on from that, does it suggest that too much is being
spent on rail safety?
Professor Evans: I think it certainly implies
an inconsistency, although I could qualify that by saying the
inconsistency is less than it was because it is only in the last
decade or soa bit more than a decadethat the railways
have appealed to a value of preventing fatalities at all, so they
are at least anchored together now in a way that they were not.
We are in a position where on the whole in road safety the value
of preventing fatalities is treated as a maximum, we would not
do any safety measures that cost more than that, whereas on the
rail it tends to be regarded as a minimumif we can save
a life by that we must do it and we might do a bit more than that.
There is an inconsistency, therefore, in general attitude there.
As I have said, we could usefully spend more resources on road
safety, but it is another question whether it should come from
rail safety measures, which I am less sure about.
Q371 Lord Skidelsky: Do you think
the inconsistency is a cultural one, the attitude to the motor
car as opposed to the railway service?
Professor Evans: I think I probably would, yes.
If you raise the question later I shall come back to it, but there
is a long tradition on the railways, particularly in certain sorts
of accidents, of taking the view that if you can do something
you should do something.
Q372 Lord Vallance of Tummel: Looking
again at your written evidence, Figure 1, we can see over a 30
year period there is a steep decline in accidental fatalities
and then, perhaps rather perversely, over the last ten years,
when there has been an unprecedented interest in health and safety,
the trend has flattened out. What do you think of the public policy
implications of this and has the law of diminishing returns set
in, or should we be looking at more radical options, even though
they might impinge on individual freedomssuch as banning
motorcycles or whatever?
Professor Evans: I do not think there are any
obvious policy implications of that drop. I do not think you can
say that because we have not had a reduction in the last decade
that somehow we have fallen behind and therefore we should do
more than we otherwise would. On your last point, as I have tried
to indicate before, I do not think we are at the stage of diminishing
returns with regard to safety measures, certainly on the roads
and possibly on the railways as well. What is efficient is always
changing with technology as wellthings that are now thinkable
were not thinkable 20 years ago. I would not favour the sorts
of draconian bans that have been mentioned, largely because for
road safety it is necessary to get the public on side because,
in the end, we do not wish to criminalise many otherwise law-abiding
citizens, and therefore what the public will accept is quite a
limitation in road safety. Banning using the car to drive to work,
for example, if the public transport was a good alternative, is
not a way forward. If you want to reduce road useand there
may sometimes be a case for that and safety may be part of that
case, but only part of itthen pricing seems to be the obvious
mechanism, which is very much on the agenda. One of the reasons
why cost-benefit appraisal is used a lot in transport is that
you do rely on valuations of the users to quite a high degreein
other words, you do not try and second-guess the users, but to
get the users to act efficiently you need the right price.
Q373 Lord Vallance of Tummel: How
would that impact on motorcycles, which are perhaps the highest
risk form of transport? They are at the moment exempt from such
Professor Evans: From the safety point of view
that was a questionable decision and it is certainly true that
motorcycles have risks per kilometre that are far higher than
any other mode of transport. If we had not invented them it is
questionable whether we would, but we have and they are very useful
for some people, so again I do not think I would ban them.
Q374 Lord Roper: As far as diminishing
returns on rail are concerned, your written evidence suggests
that the cost for each prevented fatality associated with the
introduction of the Train Protection and Warning System (TPWS)
is significantly higher than the value placed on a prevented fatality
in normal safety investment assessments. Does this imply that
the decision to introduce TPWS is inconsistent with the normal
rail safety assessment guidelines, or could you say something
about the other benefits of introducing TPWS and reducing accidents
which presumably justify it?
Professor Evans: One needs to distinguish between
TPWS as implemented and TPWS as originally proposed. I drew attention
to that distinction by quoting the House of Commons evidence that
they had received from Mr Armitt, the Chief Executive of Network
Rail, where the original proposal for the Train Protection and
Warning System had a cost of preventing fatalities that was almost
certainly less than £3 million, when you take the other benefits
of preventing train collisions into account. At the time also,
I should say, the railways valuation of a prevented fatality was
higher than the £1.4 million it now is, so at the time with
the lower initial cost the Train Protection and Warning System
was within shouting distance, shall we say, of the official valuation;
in that sense it was not wildly out of line. I would also say
that TPWS is a very clear case. Its characteristics areagain
as I have tried to indicate in my written evidencethat
there was a known and persistent risk. The data on the kind of
accidents that would be prevented by TPWS are remarkably clearthey
were risks that went on and on, not very frequent but very persistent.
Until Automatic Train Protection came along there was no technical
solution to it, but then suddenly we had a technical solution
and in the case of TPWS quite a cost-effective one as well, and
there was also a clear railway responsibility there, because if
ever there was a class of accident that was not the victim's fault,
it was that one. The situation is you have a known risk, you can
do something about it, it is the railways' responsibility. Back
to your question about accountability, there is a very strong
tradition in that situation of saying you should do it and that
was a very strong argument. In that situation, therefore, given
the original costs, it seems a good buy and so that would be a
case where I think I would have exercised political judgmentif
it had been my responsibilityin favour of it. That is an
example of a situation where, as it were, you should have discretion,
you should not regard the values as gospel. The further expenditure
for the last five fatalities which I mentioned, I do not know
how that happened and obviously the case for that is a lot weaker.
Q375 Lord Sheldon: The rail industry
was opposed to Automatic Train Protection but it welcomed the
Train Protection and Warning System. Was it because it was cheaper?
Professor Evans: The tone of the description
of resolute resistance is slightly in error. It is not true to
say that the British Railways Board, which was the key body at
the time, was resolutely resistant. In fact, the history is that
the British Railways Board committed itself to installing Automatic
Train Protectionit did not resist itin November
1988, really as soon as it became a practicable possibility, and
essentially for the same grounds as later TPWS was approved. It
was a known risk, you had a technical solution, we are a responsible
operator, therefore we should install the safety measure. That
was essentially the line of argument, but there is no reference
to costs in that argument as I have just said it. It was in those
circumstances that the British Railways Board ordered equipment
for the Chiltern and Great Western lines in order to develop it
to a working level for the whole network. So in the late Eighties
and earlier Nineties they were very keen on it. What made them
change their minds, as they did in the end, was that (a) they
got more reliable estimates of cost than they previously had,
and it turned out to be pretty high, and (b) they had more reliable
estimates than they previously had of how many lives it could
be expected to save, and when they divided the latter into the
former they got a very high figure which they felt they could
not justify. They felt you could spend better safety resources,
even within the railways, and save more lives in other ways. One
of the British Railways Board's last acts before it handed over
to Railtrack in March 1994 was to advise the Secretary of State
against network-wide installation of ATP. They were convinced
by the argument, but it was still against their nature in a way
as a responsible railway operator. Just to complete the story,
when the Secretary of State received that recommendation he referred
it to the Health and Safety Executive who were the safety regulators,
and they in turn said that they considered it to be not reasonably
practicable by their normal criteria. I have forgotten the exact
wording, but it was something like that. The Secretary of State
concurred with that decision, so it was in the end a collective
decision not to do it, essentially on the grounds of cost and
with TPWS being developed as a cheaper alternative. They did decide
against it, therefore, but in a sense reluctantly. Part of the
reason for that is that the train drivers were driving very safely.
It is because they were so good at obeying red signals that you
got so little return on devices to help them. On the question
of the change of heart about TPWS, I said there is a much stronger
case because the cost is loweror at least in the original
form it wasbut I think once it is mandated it is not in
the interest of anybody in the railway community to say we do
not want it, because it does save them from accidents that they
really do not want. Likewise, the public will also welcome it
provided that the taxpayer is paying for it and it does not affect
Q376 Chairman: The last question
we have got is, in your view, how should the safety levels on
different transport modes be compared: risk per passenger-mile,
risk per passenger-hour or risk per journey, the three options
we have got down here? How would you do the comparison?
Professor Evans: The answer to that is it depends
on the purpose for which you are doing the comparison, and I will
try and give some illustrations. I have used all three of those
measures for different purposes and I might try and use some illustrations.
I gave an example of the access risk in my written evidence of
rail journeys and I tried to show that in fact the access risk
of rail journeys is bigger than the risk on the train journey
itself, and for that purpose I was using per journey as the divisor
and I think it is appropriate because there are two accesses,
one at each end, and that is independent of the length of the
rail journey. Within that calculation I used walking kilometres
and driving kilometres as measuring the risk in the access process,
so I combined journeys and kilometres or miles covered in I hope
a sensible way. For some other purposes I think risks per hour
are very interesting, and one of the purposes for which they are
very interesting is that you can calculate passenger risk per
hour as the Department for Transport does and as I have done.
One of the uses of that measure is that you can compare travelling
per hour with doing other things per hour, which you cannot on
any other measure, but I did not produce any results of doing
that in my written evidence because I have not got anything that
I regard as sufficiently up-to-date. When I last did it, one of
the interesting findings was that, for most people who do not
do exotic things like risky sports, travel is the most risky thing
you do, especially travel as a pedestrian. We all do it, and there
are very few activities that have a higher risk per hour. That
seems interesting to me and it is another reason for devoting
resources to road safety.
Chairman: Thank you very much. Lord Skidelsky.
Q377 Lord Skidelsky: Two quick ones,
if we have got time, My Lord Chairman. I just want to go back
to the willingness-to-pay methodology. Would you say that the
value attached to preventing fatal accidents as measured by willingness-to-pay
is less for road transport than for rail transport?
Professor Evans: The only piece of evidence
that I am aware of, and it is very specific evidence, has actually
been prepared by your Specialist Adviser, who did a studyhe
can correct me if I am wrong after I have left the roomfor
the Health and Safety Executive and, essentially, in a more careful
way than I can describe tried to ask people that very question:
"are you willing to pay more to prevent an accident as a
railway passenger than you are as a road user?" And the answer
to that was "no, we are not". They got the same value.
They even repeated it, very interestingly. The first study was
done in 1998, which was about halfway between the Southall and
the Ladbroke Grove railway accidents, and they repeated the study
after Ladbroke Grove. They found that people were willing to pay
slightly more when Ladbroke Grove was in people's minds, but negligibly
more, not much more. The answer was the same even then.
Q378 Lord Skidelsky: So the culture
should not be constrained on safety measures?
Professor Evans: If you rely on willingness-to-pay
I agree with you.
Q379 Lord Skidelsky: Just one last
question, to go back to the assumption that as societies become
wealthier they become more risk averse, which seems to be the
common view. If you believe, as the Prime Minister seems to, that
we are in danger of becoming excessively risk aversethat
would be the tenor of his big speech last yearhow does
the Government act to counter this natural tendency towards increasing
risk aversion, if it believes that that is not a healthy development?
By reducing the volume of safety regulation as the natural risk
aversion increases? I know it is not a question one can answer
quickly, but do you think there is a danger that we are becoming
too risk averse?
Professor Evans: I think there is a possible
danger that we are putting too much weight in some areas and not
enough in other areas. One thing that I had not thought about
that struck me as I prepared the graph in Figure 1, which is the
graph of trends in accidental death, I suspect that most of the
deaths in that graph are actually in private arenas rather than
public arenasthe two big groups are roads, which is a mixture
of public and private, and at home which is almost entirely private.
I suspect that a lot of the large "Other" group is also
largely private, sporting accidents and that kind of thing. Therefore,
in a way, a lot of what this risk regulation is about is attacking
a rather narrow field of total deaths.
Q380 Lord Sheldon: I am just looking
at table 4 in your report and I am looking at the access modes.
It seems that walking is the most dangerous of all the access
Professor Evans: Absolutely, apart from motorcycling,
which is not in that table. Motorcycling is even worse, but I
have omitted it from the table because it is not much used in
access to rail journeys. One of the surprising things about this
piece of work we did is the amount of walking that is associated
with rail journeys, something I had not guessed. My estimate is
that 5 per cent of all walking is to and from mainline railway
stations, and probably another 5 per cent is to and from underground
stations, because they are about the same number of journeys.
So that accounts for a lot of walking, and that is why it is important
in that table, it is about a kilometre for every rail journey.
Motorcycling has got higher risks, but there is a great deal less
of it associated with rail journeys.
Chairman: Thank you very much indeed.
You have answered a lot of questions in jolly quick time.
We are most grateful to you for that and we are grateful
to you for coming. You have been very helpful to us and you have
been spot-on some of these areas we are most concerned about.
Thank you very much indeed.