Examination of Witnesses (Question Numbers
109-119)
PROFESSOR PATRICK
DUNLEAVY AND
PROFESSOR JUSTIN
FISHER
27 JULY 2010
Q109 Chair: Welcome, Professor Fisher
and Professor Dunleavy. Sir Peter Soulsby and Mr Williams will
begin the questioning. I just wondered if there were any general
things you wanted to say to start us off?
Professor Fisher: Yes.
Q110 Chair: We have received your
evidence, thank you very much.
Professor Fisher: Firstly, just
to say that I am agnostic on the AV system but I would emphasise
that it represents a relatively minor shift from first-past-the-post,
contrary to some of the evidence that was presented last week.
Secondly, in terms of a referendum, I think it is worth bearing
in mind that we are on almost entirely new ground for holding
a referendum following the Political Parties, Elections and Referendums
Act (PPERA). There has only been one referendum held under that
Act in 2004 in the North East, so we can learn something from
past experiences in this country but we need to be wary of the
relatively new legislation. Thirdly, to emphasise a point that
has been picked up in the first part of this session, to look
at the Government's reform process more broadly and that is that
there are knock-on effects for other pieces of legislation which
need to be borne in mind. I am thinking in this case of fixed
parliaments, which is not something in this Bill but I think it
is something that is worth raising to flag up. Finally, whilst
I take Nick Boles' point about the manifestoes at the last election,
it does seem to me that if it is deemed sufficiently important
to have a referendum on AV, it is arguably sufficiently important
to have a referendum on reducing the number of MPs by such a large
amount given that this would be the largest reduction since the
elimination of Irish MPs in the House of Commons in the early
1920s. I have no particular view on this but it does strike me
that this is an issue which has not attracted the sorts of discussion
that we might expect.
Q111 Chair: Professor Dunleavy?
Professor Dunleavy: Thank you.
I am a strong advocate of making the minimum necessary change
to increase the proportion of MPs, in fact to get all MPs to have
local majority support. I do think the measure is highly desirable
for the reasons that are set out in my evidence. I also think
if it was not to be approved there will be a continuing problem
of fewer and fewer MPs having local majority support, and that
is very unlikely to go away. However, I think the Government has
perhaps skipped a stage, a rather crucial stage, and it comes
out of the need to do a coalition agreement rather quickly which
does mean that we have a proposal for the referendum which in
my view is ambiguous because it seems to be offering voters either
a whole class of electoral systems or one system. It is not clear
to me which of these is being proposed. Alternative Vote means
that you are electing a single office holder but you are using
an instant runoff form and that you are looking at multiple preferences.
As my evidence sets out, there are three or four existing versions
of that system you might want to use and the Government has in
mind the particular version that is currently used in Australia
but that was not the version that was used for many years in Australia
and it has certain advantages and disadvantages. Because there
has not been a little commission or a royal commission or an investigation,
except of course by your Committee, I think there has been a bit
of a stage missed out. Election systems are not cast in stone,
they are not implemented in the same way everywhere. When you
take a new election system into a country you always tweak it.
Sometimes you tweak it deliberately and sometimes you do it inadvertently
because you have made a mistake. For example, we have a very distinct
kind of additional member system in this country which is very
specific to this country and is not found in any other use of
additional member systems, so the British one has more local MPs
and fewer top-up MPs. That came out of the Scottish constitutional
convention and was used in London and Wales as well. There is
a whole set of tweaking and very detailed decisions. Voters need
to know in great detail what exactly it is that the government
means when it says, "Do you want to replace first-past-the-post
by the Alternative Vote". There is this ambiguity between
basically two versions of the Alternative Vote, one of which allows
people who are placed third, fourth or fifth in the initial ranking
of votes to win office and another one, the London version, which
really creates a kind of runoff between the top two.
Q112 Chair: Could I ask you, just
to start you off, whether you feel it is helpful or unhelpful
to link in one Bill the issue of electoral systems and the issue
of the number of Members of Parliament?
Professor Dunleavy: If you were
going to introduce AV-plus it would be helpful to be reducing
the number of MPs because you would be redistricting anyway because
you would need to create top-up MPs. I do not see any clear connection
between AV as a class of system very closely related to first-past-the-post
so I do not see any virtue in linking it but I do not see any
particular disadvantage either.
Professor Fisher: I think it is
a risky strategy. I do not see any particular problem with linking
them, but there is a danger if one half of the Bill gets into
difficulty then the whole Bill may fall. In that sense it is risky
for the Government but I do not see a problem in linking them
together in one Bill.
Q113 Sir Peter Soulsby: Professor
Fisher, in your introduction you talked about a change to AV as
a relatively minor issue. Professor Dunleavy, you talked about
it as an overwhelming public interest case. Those are rather different
ways of describing it. How significant do each of you feel it
will be were we to adopt it in terms of the outcomes it would
achieve and public perception?
Professor Fisher: The only evidence
that you can really use is evidence based on survey work. If you
have simulations based on aggregate data there are an enormous
number of assumptions which are not terribly helpful. The work
that was done by Patrick in the 1990s and more recently by the
British Election Study shows firstly that people's first preferences
tend to be fairly similar, particularly for Conservative and Labour,
but the outcome tends to amplify the national mood. If you take
the 1992 election, when there was a simulation done then, rather
than being a small Conservative majority, there was a Conservative
minority. If you run the simulation in 1997 the Labour majority
would have been rather larger than it was. If we are looking at
the effects, it is fair to say that it probably amplifies very
slightly what you get under first-past-the-post. The general principle
of electing one person to serve in a constituency seems to me
to be not a huge departure from where we are currently.
Professor Dunleavy: I think my
view is not totally distinct from Justin's. We did a lot of work
in simulating when Labour was changing the electoral system in
Scotland, Wales and Northern Ireland. Basically the changes in
behaviour which took place were much more extensive than those
that we had envisaged. In particular I think if you are introducing
a numbering of preferences—one, two, three, four, five—I
would expect that support for smaller partiers will go up. I would
expect that, depending on how you design the system, in the London
AV you are only given the first and second preference so that
restricts your expression of preferences. If you go one, two,
three, four, five at least a substantial proportion of voters
may begin to cast multiple preferences, so more smaller parties
may tend to get first preference votes and the first preference
support of the winner will tend to decline. When we were advising
Nick Raynsford on the London mayoral system, one of the reasons
why he and the government at that time went for the restricted
expression of preferences to first and second was to try and avoid
having large numbers of candidates. On the whole I think that
has been quite successful, the number of candidates for London
Mayor went up to ten but it has not gone up continuously. I would
think there would be a slight danger that the AV system might
conduce to a fragmentation of the votes further. If everybody
expresses a complete preference set so that they number all the
candidates you still have MPs with majority support but if you
get a fragmentation of the votes and then people expressing only
a few preferences then the winning candidate may not have majority
support. I think there are a lot of dynamic things that
you have to take into account. You have to think what will the
ballot paper look like and I have included in my evidence a couple
of versions of the AV ballot paper. You have to think how will
the candidates and voters behave when they are confronted by a
different ballot paper with this different task that you are asking
of them. You have to think what is the trend of political party
activity in the UK and it is overwhelmingly towards increased
fragmentation. For example, there were no constituencies at all
in 2005 or 2010 with two party contests and there were almost
none with three party contests. We are heading in a very different
direction. There are a large number of other parties bubbling
under that might be encouraged or fostered by particular choices
that are made on the voting system.
Q114 Sir Peter Soulsby: Why do you
think the AV system is so rare? Is there any example of it having
led to increased public satisfaction with the outcomes?
Professor Dunleavy: AV has been
used in Australia more or less since its foundation. Initially
it was used in a version where people had to number all the parties
but as the number of parties increased the Australians moved to
a system where you do not have to number all the choices. That
is quite a big difference. It is not used virtually anywhere else
in the world. It has been used in some Canadian provinces.
Professor Fisher: It is not something
that is unique to AV; it is something that you find with preferential
systems. STV, for example, is only used in a relatively small
number of cases. One of the issues with that is firstly that AV
is not proportional and a number of democracies have opted for
a more proportional system. Secondly, there can be a danger with
preferential systems, particularly if, as was the case in Australia,
you have to rank every candidate. You can experience something
called donkey voting where people simply vote for the candidates
in the order that they appear on the ballot paper. It is the same
principle that is suggested if you go in the Yellow Pages
you will have lots of Aardvark plumbers but very few Zoo plumbers,
if you see what I mean. There are some dangers with that. I do
not think they are insurmountable. Through administrative techniques
you can override the effects, such as reordering the ballot paper
in different districts and so on. The general trend has been not
to move from a plurality system to a majoritarian system but often
to move towards something that is more proportional.
Q115 Sir Peter Soulsby: Some of the
proponents of AV see it as a step towards a proportional system
of some sort. Do you yourselves see it as an end state or a step
toward something different?
Professor Dunleavy: In my evidence
in part three I pointed out that general elections are very often
held on the same day as other elections, that in particular the
Government will be announcing plans in January, I understand,
for the new constitution of the House of Lords. We do not know
what timetable that is on but that is likely to require the House
of Lords election takes place on the same day as the general election.
It is common for general elections and local elections to take
place on the same day and we have had instances of general elections
and European Parliament elections on the same day. There would
be a risk also of Scottish Parliament, Welsh National Assembly
and London elections possibly coinciding. When you are making
a voting system decision like this you do need to take into account
how consistent or inconsistent is the new voting system with the
ones you have already, particularly for voters. There are small
problems for administrators. In particular most of our existing
voting systems use X voting and moving to numerical voting will
create problems in having the elections on the same day and would
also tend to mean that if you move to numerical voting for Westminster
perhaps there will be more pressure for STV for the House of Lords
as one of the three possible big systems coming there. Perhaps
there will be more pressure for STV for local elections because
that way you would be able to recreate consistency. I certainly
think if AV in the Australian form was introduced you might need
to make some consequential changes in how the London Mayor is
elected and other English mayors. People have complained already
about a bewildering variety of electoral systems in use in the
UK. I do not think that really matters very much to voters but
I do think the thing that matters is if they are being asked to
do X voting and numerical voting on the same day. We have had
one case of this in Scotland and it was a very searing problem
for Scottish voters. The Government should be careful and should
indicate how this is going to become possible for the future.
Professor Fisher: I have to say
I do not really buy this Trojan horse argument. It strikes me
that particularly because the decision on whether to adopt this
will be based on a referendum it would be very difficult for a
subsequent government to hold another referendum or simply change
the decision. One of the advantages of a referendum is that it
embeds the will of the people. Patrick may be right with other
levels of government but it strikes me that for Westminster elections
the idea that a party propping up a coalition could in effect
demand a further change is perhaps unlikely.
Q116 Stephen Williams: Before I ask
the question on AV which I do have, Professor Fisher said something
interesting in his opening remarks that perhaps we could have
done with in the previous session. He said the reduction of the
House of Commons by 50 was the biggest change since the Irish
or the Southern Irish left in 1922. The difference between then
and now is that responsibility left as well. Westminster is no
longer responsible for Dublin, Limerick, Cork, et cetera. Can
either of you think of another instance in the democratic world
where an assembly has had its numbers reduced without some transfer
of sovereignty or devolved power somewhere else?
Professor Fisher: I do not know
of any.
Professor Dunleavy: The UK House
of Commons, to put it in perspective, is one of the largest legislators
in the world. Apart from the Chinese and Supreme Soviet there
are not many that are bigger than the UK. UK constituencies are
quite small. It is not massively over provided with MPs but it
is certainly not short of MPs at the moment compared with other
countries.
Professor Fisher: I think that
is an important point. There may well be international examples.
I am not aware of them but it is important that the idea of what
works in a particular country is retained. My point about whether
or not there should be a referendum on this, if it is deemed sufficiently
important to have one on AV, is that whereas there has been a
reasonable amount of discussion about the potential impact of
a certain electoral system, it strikes me there has been almost
no discussion about the potential impact of a reduction of MPs
even to 600 as opposed to 585 which stems, it seems to me, from
a rather populist response to the expenses crisis. Given that
we know constituents use their MPs more than they ever did, given
that we know that people value local representation, I am not
sure that the argument has been put to people that, effectively,
increasing constituency size by an average of 10% is necessarily
what the voters want. I do take the point that it was in the manifesto,
and that seems to me a perfectly legitimate argument, but I go
back to the point: if it is deemed legitimate to have a referendum
on AV then it strikes me that this change is of a similar magnitude,
and therefore might be worthy of further consideration.
Q117 Chair: May I just ask a question
of political theory? By what right does a government, an executive,
reduce the numbers, powers or composition of a legislature in
the UK?
Professor Dunleavy: The UK is
in a rather unique situation because we have a concept of Parliamentary
sovereignty instead of a concept of constitutional sovereignty.
So I do not think that this is inconsistent. The size of the House
of Commons has gone up and down in the past; this is the most
dramatic change, but I do not see it as a constitutional change.
Q118 Chair: Let me be clear: you
see the exercise of executive power as an exercise of Parliamentary
sovereignty?
Professor Dunleavy: I think the
justification that has been given is a mandate doctrine justification.
It is a stretching, certainly, but I would not say it was quantitatively
a constitutional change.
Q119 Chair: Do you regard it as legitimate?
Professor Dunleavy: It is not
something I am terribly keen on, but—
Professor Fisher: Let us be clear:
there is no existing practice about when one has a referendum.
It is not set in stone that it is about constitutional change;
it is often for purposes of convenience of government. So I go
back to my point about AV; there is no particular need to have
a referendum on AV—one could do without one—and indeed
I think I am right in saying that had Labour won the election
it would have been introduced anyway. My point is simply that
if you are embarking on this road we are differentiating between
two different fairly significant changes in the way in which people
are represented. Whereas there is an acceptance that you should
have a referendum on one, it strikes me that the other issue has
not been discussed in such depth.
| 
