Abstract
Europe was a continuing source of headlines during the 1992-97 parliament, whether over the Exchange Rate Mechanism, the Maastricht treaty, BSE, quota-hopping, or the question whether Britain should join the European currency. Throughout, it was a major source of division within the Conservative party at Westminster, most notably at the time of the nine whipless rebels. It thus will certainly have made a major impact on the ungovernability of the Conservative Party, its public image of disunity, and hence on the general disillusion of voters with the government.
But Europe also seemed to have played a more direct role at the polls in 1997 with the votes won by the late Sir James Goldsmith's Referendum Party, which campaigned on the single issue of a referendum to allow the British public to decide whether the United Kingdom should 'be part of a Federal Europe'. David Mellor memorably described Goldsmith's share of the vote in his constituency as 'derisory', but it can perhaps better be viewed as the strongest-ever performance by a British minor party. It could well have contributed to the loss of some Tory seats.
The Referendum Party contested 547 constituencies, standing aside only in seats where the incumbent MP had given explicit support for a referendum. They won a total of 3.0% of the vote in those constituencies where they stood (while the UK Independence Party secured a further 1.1% of the vote in the 194 constituencies where it stood). But the Referendum Party's appeal tailed off noticeably north of the Scottish border where it secured an average of just 1.1% of the vote. As John Curtice and Michael Steed have suggested, this may reflect the rather higher level of support for Europe in Scotland, or perhaps a feeling that the party's anti-Europeanism was a form of English rather than British nationalism (Curtice and Steed, 1997). Its best performances were all in those parts of England where Euroscepticism was at its highest, especially the south of England outside inner London, and in East Anglia where it averaged 3.9%. It tended to do particularly well in constituencies with a large agricultural or elderly population, the former doubtless a reflection of the controversy surrounding European agricultural policy and the European response to BSE.
It was widely assumed at the time that these votes had largely come from the Conservatives and that they might have cost the Tories as many as 19 seats. (In 19 seats the Referendum Party share of the vote was larger than the majority over the Conservatives.) But it was also a rather remarkable phenomenon in its own right. The conventional wisdom is that foreign affairs play little role in domestic voting behaviour and that bread and butter issues tend to be decisive. But at face value it seems that nearly 3% of voters did vote on the basis of a single non-economic issue.
In this paper we ask the following questions:
(1) How widespread was Euroscepticism in the electorate?
(2) Where did the Referendum Party's vote come from and what were its motivations?
Was it a specific vote on the issue of a federal Europe.
Or was it a more general vote of protest by disillusioned Tories on the
right of the party for whom the Liberal Democrats or New Labour
were even less attractive than the Conservatives.
(3) What were the consequences of the Referendum Party for the Conservative share
of the vote and of seats? Would the Conservatives have been better
off, as some urged, in moving to a more Eurosceptic line? Would this
have retained the Referendum Party voters? Or would they have
defected anyway because of their general disillusion?
Data
Our data come from the 1992-97 British Election Panel Study (BEPS). The panel study is particularly valuable for a study of the Referendum Party for a number of reasons. First, it gives contemporary records of vote in 1992 and hence does not suffer the recall bias that affects other sources. Secondly, the panel enables us to look at attitudes and behaviour of eventual Referendum Party voters before the party itself was formed in October 1995. It thus allows us to tackle questions of causation, specifically the question whether the formation of the Referendum Party made any difference, in a way that simply is not possible with a conventional cross-sectional survey.
The BEPS has followed up respondents to the 1992 British Election Survey. Respondents were reinterviewed in 1993, 1994, 1995 and 1996 and for the final time after the general election in 1997. The 1992 survey was carried out by face-to-face interview, as were the 1994 and spring 1995, 1996 and 1997 waves, but the 1993 wave was a short postal one. There were also two short telephone waves in the autumns of 1995 and 1996. We draw primarily on the five face-to-face waves.
In the 1992 BES, 3534 interviews were achieved, a response rate of 73%. Of these original respondents 2,277 participated in the 1994 wave and 1924 in the final wave after the 1997 general election. However, we should note that the original 1992 BES oversampled in Scotland in order to enable a detailed study to be undertaken of Scottish voting behaviour for the Scottish Election Study. Where the data are used as a British sample, as they are in this artcile, a weighting factor must be used to downweight the Scottish oversample to form a representative British sample. The weighted N which we report in this paper for the 1997 wave is therefore 1583. Since we are ultimately interested in how people voted in 1997, we restrict ourselves throughout to people who completed the 1997 wave of interviews. We should note however that the BEPS does not include young voters who entered the electorate after the 1992 general election (since they would have been excluded from the sampling frame for the 1992 BES).
Every effort has been made to maintain the panel, but clearly it has been subject to attrition and inevitably the attrition has been concentrated in certain groups rather than uniform across all groups. Wherever possible, therefore, we check our results against independent random samples (such as the 1997 BES), which have not suffered from this problem of attrition. Checking results against the 1997 BES is also important given the small number of Referendum Party voters in the panel study. In the final wave of the panel we have only 34 of these voters, and we therefore have to be very careful that our results are not due to sampling error. As well as carrying out tests of significance, we therefore also check our results against those found in the 1997 BES, which contains 58 voters for the Referendum Party. Since the BEPS and the BES are completely independent samples, checks of this kind give one much greater confidence in the results.
How Eurosceptic was the electorate?
The attitudes of the British electorate towards the European Union, perhaps like that of the Conservative and Labour parties, has been distinctly ambivalent. On the one had, rather few people actually want to leave the EU, and in this respect there is very little popular support for a radical Euroscepticism, such as that occasionally mooted by Eurosceptics such as Norman Lamont. On the other hand, a large proportion of the electorate is uneasy about the moves towards integration implied by the Maastricht treaty, and (as elsewhere in Europe), doubts about European integration have been increasing since the Treaty was signed. This is shown clearly in Table 1. We asked our respondents:
Do you think that Britain's long-term policy should be ...
to leave the European Community,
to stay in the EC and try to reduce its powers,
to leave things as they are,
to stay in the EC and try to increase its powers,
or, to work for the formation of a single
European Government
As we can see from Table 1, in 1992 the electorate was fairly evenly divided between Eurosceptics and Europhiles, but by 1997 the balance had shifted decisively in a Eurosceptic direction with 66% wanting either to leave outright or to reduce EU powers.
Table 1
Euroscepticism in the electorate (column %)
1992 1994 1997
% agreeing that ...
Britain should leave the 9 10 14
EC
reduce its powers 35 39 52
leave as is 13 14 16
increase its powers 29 21 9
work for single European 11 11 5
govt
Don't know 4 6 4
Replace pound 25 20 16
have both ECU and pound 24 21 26
only have pound 49 58 55
Don't know 3 2 3
Source: BEPS 1992, 1994 and 1997 waves.
Similarly, the electorate was distinctly Eurosceptic, and becoming more so, on the issue of a single currency. We asked our respondents:
And here are three statements about the future of the pound in the
European Community.
Which one comes closest to your view?
Replace the pound by a single currency
Use both the pound and a new European currency in Britain
Keep the pound as the only currency for Britain
As we can see, even in 1992 keeping the pound on its own was the most popular of the three options, and by 1994 opinion had hardened somewhat.
Euroscepticism, therefore, was on the increase over the course of the last Parliament, and in this respect popular sentiment was moving in the direction of the Referendum Party and of the Conservative Party's right wing. Indeed, it seems to have been the only major issue where popular sentiment was moving in a right-wing direction.
To chart the public's changing attitudes towards major issues, and their changing perceptions of the parties' positions, we constructed a number of measures where respondents were asked to place themselves, and the parties, on eleven-point scales running between two contrasting policy options. These measures have been the major tool in BEPS for understanding the changing impact of issues on the electorate. They covered European integration, the redistribution of income, the control of unemployment and inflation, tax cuts versus government spending, and privatization. The question wording of the end-points of these scales was:
Britain should do all it can to unite fully with the European
Community
vs
Britain should do all it can to protect its independence from the
European Community
Government should put up taxes a lot and spend much more on health
and social services
vs
government should cut taxes a lot and spend much less on health and
social services
Getting people back to work should be the government's top priority
vs
keeping prices down should be the government's top priority
Government should nationalise many more private companies
vs
government should sell off many more nationalised industries
Government should make much greater efforts to make people's income
more equal
vs
government should be much less concerned about how equal people's
incomes are
The scales have been coded so that high scores represent right-wing positions.
Table 2 shows how it is only European integration that has seen a substantial popular shift to the right. On every other major issue, public opinion moved away from the Conservatives over the course of the 1992-97 parliament.
Table 2
Respondents' positions on major issues (mean scores)
1992 1994 1995 1996 1997 Change
Europe 5.9 6.5 6.7 6.9 6.7 +0.8
Privatisation 5.9 5.2 5.3 5.3 5.5 -0.4
Inequality 4.9 - 4.3 4.5 4.6 -0.3
Unemployment 3.6 3.8 3.7 3.5 3.7 +0.1
Taxes 4.1 4.5 4.1 3.9 3.8 -0.3
Source: BEPS
The fact that public opinion moved in different directions on the European question and on the standard bread-and-butter issues of unemployment or tax cuts suggests that they may have rather little in common. It may be misleading to think of Euroscepticism as a right-wing issue in the same sense that we think of privatization as a right-wing issue.
This is indeed what we find. Factor analysis shows that public attitudes towards redistribution, privatization, unemployment and tax cuts form a relatively tightly-integrated group of issues where people's attitudes are highly consistent across all four issues. If someone adopts a right-wing stance on, say redistribution, there is a high probability that they will be right-of-centre on tax cuts, privatization and inflation as well. For this reason we can think of these issues as reflecting people's positions on a basic socialist/laissez faire dimension that underpins their specific issue preferences (see Heath, Evans and Martin, 1994).
Attitudes to Europe, however, do not map onto this socialist/laissez faire dimension in a straightforward way. It is not entirely unrelated to it, but there is much less consistency between people's positions on Europe and their left/right values than there is between their positions on the individual economic issues. (For further details see Heath et al., forthcoming).
What this means is that people who share similar economic attitudes on privatization, inflation and so on may be fairly dissimilar in their attitudes to Europe. And this of course may explain why Europe has the potential to divide political parties. Since parties in Britain are fundamentally based on groupings of like-minded people on the socialist/laissez faire dimension, a cross-cutting issue like Europe is quite likely to be divisive - just as nuclear disarmament, another cross-cutting issue, divided the Labour party in the early 1980s.
To be sure, there are other issues, such as questions of abortion, divorce and the death penalty, which cross-cut the left/right dimension even more powerfully. But these are issues which British political parties have historically kept out of party politics. They are ones on which a free vote is typically permitted in the House of Commons, and hence their potential for creating internal party divisions has been rendered harmless. Europe and nuclear disarmament, however, are issues which fall half-way in between the moral questions such as abortion and the death penalty and the economic questions of the left/right domain. Hence it is not so easy to keep them out of party politics. But nor is it easy to keep them within the conventional party framework, as perhaps the Referendum Party demonstrated among the electorate and the whipless rebels demonstrated in the House of Commons.
The source of the Referendum Party vote
It is the autonomy of the European question from the conventional left/right dimension that gives it its potential to divide the parties internally. Is this also a factor in the (relative) success of the Referendum Party in detaching voters from their usual party alignments, or were voters for the Referendum Party simply signalling a general right-wing protest against the Tories?
The Referendum Party was a creation of the late Sir James Goldsmith, who announced his intention to form it on 27 October 1994, although it formally came into being a year later in October 1995. It campaigned on the single issue of a Referendum on Europe. It eventually published the text of its proposed question for a referendum, namely:
Do you want the United Kingdom to be part of a Federal Europe
or
Do you want the United Kingdom to return to an association of sovereign
nations that are part of a common trading market?
Sir James emphasized that his proposed referendum was quite different from those proposed by the Labour and Conservative parties, which dealt only with the question of joining the single European currency.
Voters for the Referendum Party were not a cross-section of the electorate. They were predominantly people who had voted Conservative in 1992 and, to a lesser extent, for the Liberal Democrats. Hardly anyone who had voted Labour in 1992 supported Goldsmith's party. This is shown clearly in Table 3, and although the numbers of Referendum Party voters in the BEPS is very small, we find the same pattern in the 1997 BES.
Table 3
The sources of the Referendum party vote
1992-97 panel 1997 BES
Referendum All Referendum All
Party Party
voters Voters
Vote in
1992
Conservative 64 41 61 35
Labour 9 31 11 34
Liberal Democrat 24 17 19 11
Other 0 3 4 2
Did not vote 3 8 6 18
100% 100% 100% 100%
(N) (33) (1537) (56) (3615)
Sources: BEPS 1992 and 1997 waves; BES 1997.
This evidence on the sources of Referendum Party support does not on its own tell us that the formation of the party in 1995 actually took votes that the Conservatives would otherwise have won in 1997. These voters might have defected anyway from the Conservatives whether or not Sir James Goldsmith had created his party. So what was the motivation of the Referendum Party voters? Was it a vote specifically about Europe, or was it a more generalised expression of discontent among people whose right-wing sympathies made Labour unpalatable? In other words, was the Referendum Party a vehicle for right-wing disillusion with the Tories?
In order to get some leverage on the question whether they would have defected anyway, we compare the attitudes of Referendum Party voters with those of Conservative loyalists and of Conservative defectors. In Table 4 we show their attitudes on the five eleven-point scales described earlier.
Table 4
Positions on the issues (mean scores)
RP Voters Con Con All
loyalists defectors
European 10.2 8.0*** 7.1** 6.7
integration
Taxes and spending 3.5 4.7** 4.0 3.8
Privatisation 5.9 6.7 5.8 5.5
Unemployment 3.7 4.8* 3.8 3.7
Redistribution 5.3 6.4* 5.0 4.5
N (minimum) 32 340 240 1338
Source: BEPS, 1997 wave. Significantly different from RP voters * p > .05. ** p > .01, *** p > .001
What we find is that Referendum Party voters were in fact highly distinctive in their attitudes towards Europe. Europe stands out as the only issue on which these voters were actually more extreme than the Conservative loyalists. On our European scale they placed themselves at an astonishing 10.3. Since the maximum possible score is 11, this shows how Eurosceptic the Referendum Party voters felt themselves to be. Despite the small numbers, the difference between them and the Conservative loyalists is statistically significant. We can check these results from the 1997 BES and this confirms the story of Table 4.
Moreover, the voters for the Referendum Party were distinctively right-wing only on Europe. On all other issues they were remarkably similar to the Tory defectors. This demonstrates once again how Europe cross-cuts the other economic issues. There is little congruence between the attitudes of the Referendum voters to Europe and their attitudes to the standard economic issues. Their extreme Euroscepticism was not matched by an extreme free-market ideology. So in these other respects they were not perhaps natural Tories in 1992. Even back in 1992 their support for the Tories might have been conditional on the Tories' Euroscepticism rather than on their general free-market policies.
We can use the panel to check whether these Eurosceptic views were long-standing. In attitudinal research there is always the problem of establishing causal direction. We can never be sure, from a cross-section survey, whether people have brought their attitudes into line with their voting behaviour or the other way round. The strength of the panel design, therefore, is that we can check what these people felt about Europe long before the Referendum Party had ever been mooted. Because the eleven-point scales were asked only of a half-sample in 1992, we use some of the other Europe items. Table 5 shows the results.
Table 5
1992 attitudes to Europe (column %)
RP Voters Con Con All
loyalists defectors
% agreeing
Britain should leave the 64 57 47 42
EC/try to reduce its
powers
keep the pound as the only 58 55 50 49
currency for Britain
lots of good British 66 40 37 39
traditions will have to be
given up if we stay in the
EC
If we stay in the EC 84 56 48 51
Britain will lose control
over decisions
competition from other EC 26 54 50 50
countries is making
Britain more efficient
N 32 375 169 1540
Source: BEPS, 1992 and 1997 waves.
Table 5 shows clearly that the Referendum Party voters were already distinctively Eurosceptic in 1992 long before Goldsmith had formed his party. The causal priority of their European views is thus clear. It is also interesting that, in 1992, the eventual Referendum Party voters were more distinctive in their beliefs about the EU than in their policy preferences. They were, for example, distinctively pessimistic about the implications for British traditions and decision-making of continued membership of the EU, but they were not, at that stage, markedly different from Conservatives generally in their policy views about the single European currency or withdrawal. However, this changed over the course of the Parliament and by 1997 their views on the policy issues had hardened.
By 1994 they had also clearly turned against the Conservative government. Throughout the panel we measured feelings towards the parties on five-point scales. We have coded these scales so that higher scores indicate more favourable attitudes.
Table 6
Feelings towards the Conservative party (mean scores)
1992 1994 1995 1996 1997 N
Referendum 3.6 2.4 2.2 2.6 2.7 32
Conservative 4.4 3.7 3.6 3.8 4.0 341
loyalists
Conservative 4.1 2.7 2.4 2.5 2.5 146
defectors
All 3.2 2.5 2.4 2.5 2.6 1367
Source: BEPS
Table 6 shows that the subsequent Referendum Party voters had been relatively favourable to the Conservative Party in 1992, more so than the average member of our panel. But by 1994 these people who later voted for Goldsmith's party had already turned decisively against the Tories, their scores falling precipately. By this stage of the parliament they were already very different from the Conservative loyalists and looked remarkably similar to the Conservative defectors. At this point, well before the Referendum Party had been formed, it looked highly probable that few of their eventual voters would support the Conservatives again. What was it that led them to turn against the Tories? Was it simply their record on Europe, or were other factors important too? There were in fact substantial numbers of other voters, especially of Conservative loyalists, who took up equally extreme views on European integration. Why did some of them but not others defect to the Referendum Party? To explore this we can compare the loyalists, defectors and Referendum voters on a range of measures of the Conservative government's record and image. We include in our analysis measures from the 1994 wave of BEPS of the government's record on the economy and the health service (which previous research has shown to be important) together with measures of the government's image. We also include the respondents' distance from the government on major policy issues such as Europe. The results of multivariate analyses are given in Table 7.
Table 7
Logistic regression of loyalists, defectors and Referendum Party voters
Defectors RP voters RPvoters
versus versus versus
loyalists loyalists defectors
Europe -0.01 0.13** 0.10*
Taxes 0.02 -0.10 -0.08
Unemployment -0.12*** -0.12** 0.02
theeconomy 0.01 -0.37 -0.38
the NHS -0.36*** -0.09 0.24
competence -0.39 -0.24 0.35
sectionalism -1.09*** -1.25*** -0.27
Improvement(DF) 90.6(7) 35.7(7) 10.0(7)
N 571 422 213
Source, BEPS 1992, 1994 and 1997 waves
In the first column we report the results of a logistic regression in which we contrast defectors with Conservative loyalists. In the second column we contrast voters for the Referendum Party with the loyalists. And in the third column we contrast Referendum voters with Conservative defectors.
The results are rather clear cut. The eventual defectors could already, in 1994, be distinguished from the loyalists by their positions on unemployment and inflation, their evaluation of the government's record on health, and their perception of the government as 'good for one class' rather than 'good for all classes'. In general, the defectors were people who lay to the left of the Conservative party, towards the centre of the political spectrum; they evaluated the record on the NHS very unfavourably, and they felt that the Conservatives were a sectional party who were not looking after all groups in society equally.
The eventual voters for the Referendum Party also, with one crucial exception, showed the same profile. They too lay towards the centre of the left/right spectrum on unemployment and inflation, and they shared negative views about the government's sectionalism. The only respect in which they differed significantly from the defectors was in their attitudes to Europe, where they lay well to the right, both of the loyalists and of the defectors.
The natural interpretation of these results is that it was the disenchantment with the governments' policies and record that led both the defectors and the Referendum Party voters to abandon the government and that it was their attitudes to Europe which decided whether they abandoned them for a centre-left party or for a Eurosceptic party. To be sure, even from a multivariate analysis such as this, we cannot definitively disentangle the causal processes involved. It is possible that it was the government's record over Europe that led to the general disillusion of the Referendum voters; it is logically possible that if, say, the government had not signed the Maastricht treaty, the Referendum voters would have taken more charitable views of the government's performance in other areas (such as the NHS) too. But while we cannot rule out this alternative interpretation, the similarity of the results for the Referendum Party and for the defectors inclines us to the view that there were similar processes at work.
On our preferred interpretation, then, the formation of the Referendum Party made little or no difference to the Conservatives' share of the vote; it simply affected the distribution of the anti-Conservative vote between the opposition parties. If we repeat the analysis using 1997 rather than 1994 measures, the story is virtually unchanged. The creation of the Referendum Party did not materially affect the patterns displayed in Table 7 (and nor for that matter did three more years of Conservative government).
Some further evidence in line with our interpretation comes from the way our respondents voted in the 1994 European elections.
Table 8
Vote in the 1994 European elections
Conservative Labour Liberal Other didnot
Democrat vote
Loyalists 49 1 4 4 41
Defectors 13 10 19 3 56
Referendum 11 14 13 23 40
All 16 24 12 5 44
Source: BEPS 1992, 1994 and 1997 waves
Again, at this stage, the actual voting behaviour of the eventual Referendum Party supporters showed many similarities with that of the Conservative defectors, especially in their lack of support for the government. They did however display higher turnout in the Euro-elections (no doubt reflecting their concern with Europe), and they also showed at that stage a propensity to support the minor parties rather than the main opposition parties (no doubt because the positions of Labour and the Liberal Democrats on Europe were even more distasteful to them than those of the Conservative government).
So whatever put them off the Tories had already done so by 1994. In other words, the formation of the Referendum Party may well have made no difference to the Tories' loss of votes but simply capitalised upon existing disillusion.
Consequences for the Conservatives
As we noted earlier, there were 19 seats where the number of votes case for the Referendum Party exceeded the size of the opposition lead over the defeated Conservatives. In addition there were another six seats where either the vote for the UK Independence Party alone, or the UKIP in combination iwith the Referendum Party also exceeded the Conservative majority.
Our own data show rather clearly that the Referendum Party primarily gained votes from people who had voted Conservative in 1992 and Table 3 suggests that just just under two-thirds had probably voted Tory in 1992. This surely represents the upper bound for any estimate of how many might have supported the Conservatives again in 1997 if the Referendum Party had not been formed. Depending then on what assumptions are made about how the remaining one-third divided, this means that Eurosceptic candidates can have cost the Conservatives 18 seats at most (and perhaps as few as 13).
But our analysis in the previous section makes it very implausible that anything like two-thirds of the Referendum Party voters would actually have supported the Conservatives again in 1997 in the way they had done in 1992. As we have seen, they were just as disillusioned as other defectors with the Conservative government.
Unfortunately we have too few respondents in the panel to be able to determine how people with similar views behaved in the constituencies where Goldsmith did not put up a candidate. But we can look at what our respondents said their second choice of party would have been in 1997. The answers are shown in Table 9.
Table 9
2nd choice of vote in the 1997 election
Conser- Labour Liberal Other None/
vative NA
Loyalists - 19 45 7 29
Defectors 31 15 26 4 24
Referendum 26 23 25 6 19
All 7 17 45 7 24
Source: BEPS 1992 and 1997 waves (excluding nonvoters)
Table 9 reinforces our story that the Referendum Party voters were very like the Conservative defectors, and it suggests that they would have divided their votes fairly evenly between the main parties. On this evidence, the effect of the Referendum Party was entirely neutral, and so if the upper bound is that Goldsmith cost the Conservatives eighteen seats, the lower bound is that it did not cost them any at all. However, the actual cost may well have been slightly more than this lower bound: the 1997 BES shows somewhat more second-choices for the Conservatives and fewer for Labour. Thus in the BES 36% of the Referendum voters gave their second choices to the Conservatives, 30% to the Liberal Democrats, and 18% to Labour. This suggests a very similar estimate to the one that Curtice and Steed (1997) reached on the basis of their analysis of the aggregate data. They concluded that '... there are just six seats where the presence of an anti-European candidate can be said to have cost the Conservatives the seat, of which two were the result of a UKIP rather than a Referendum Party intervention'.
Conclusions
Voters for the Referendum Party were remarkably Eurosceptic but were unremarkable in most other respects. They show no sign of being right-wing on the economic issues of the left-right dimension and they were not consistently right-wing ideologues. Their vote does not therefore appear to have been either part of a specifically right-wing revolt against the Conservatives or a general diffuse protest vote. The fact that these voters chose the Referendum Party rather than any of the other options open to them was undoubtedly a result of their long-standing and specific concerns about Europe.
At the same time, however, our data from the panel study shows that their disillusion with the Conservative party was also long-standing and was already evident at the European elections of 1994, well before the Referendum Party had been formed. In almost all respects, other than their attitudes to Europe, the eventual Referendum voters were remarkably similar to other defectors from the Conservatives and we therefore conclude that Sir James' Goldsmith's formation of the party probably had little or no impact on the Conservative share of votes or seats, but simply redistributed the anti-Conservative vote away from the centre-left parties.
Our evidence also suggests that, had the Conservatives adopted a more Eurosceptic line in order to appease Sir James Goldsmith, this would have been of little electoral advantage to them. The Conservatives were already well-placed on Europe, their perceived position being closer to that of the average voter than was that of any of the other major parties. A further shift towards Euroscepticism would, if anything, have reduced the Conservatives' overall appeal. It would only have made sense if it would have won back the passionate minority who adopted extreme Eurosceptic views. But on its own a changed policy on Europe might not have been sufficient to win back the voters who defected to the Referendum Party, since their disillusion with the Conservatives was more thorough-going. To become electable once more the Conservatives need to tackle the problems that worried both defectors and Eurosceptics alike.
1. We can check these findings from the 1997 BES. In the BES we find that rather more, 17%, felt that Britain should leave the EU but rather fewer, 43%, felt that Britain should try to reduce the powers of the EU.
2. The 1997 BES shows 16% wishing to replace the pound, 20% wishing to have both, and 60% wishing to have only the pound.
3. The 1997 figures can be checked against the 1997 BES. The results for the BES are comfortingly close to the 1997 BEPS figures: 6.6 for Europe, 5.3 for privatisation, 4.3 for inequality, 3.7 for unemployment and 3.7 for taxes.
4 Factor analysis of the five items used in table 2 yields a single factor, but the loading for Europe, at 0.37, is much lower than those for the other items, which are all around 0.70. The same pattern holds in the 1997 BES.
5. As might be expected, there are some differences in the distribution of 1992 vote between the two sources. The panel shows a lower proportion of 1992 nonvoters since nonvoters are particularly likely to drop out of a political panel study. On the other hand, the 1997 BES shows a higher ratio of 1992 Labour to Liberal Democrats, almost certainly reflecting recall bias.
6. Loyalists are defined as respondents who voted Conservative both in 1992 and in 1997. Defectors are defined as people who voted Conservative in 1992 but then in 1997 either abstained or voted for a party other than the Conservatives or Referendum Party. The Referendum Party voters include all those who reported voting for the RP in 1997, irrespective of how they voted in 1992. Respondents for whom there was missing data on either the 1992 or 1997 vote questions are excluded, but the base includes non-voters.
7. In the 1997 BES the Referendum Party voters scored 10.0 on European integration, 5.4 on redistribution, 5.8 on privatization, 4.0 on taxes and spending, and 3.9 on unemployment.
8. Thus their average distance was 4.2 points to the right of the Tories, but 7.1 to the right of Labour and 7.5 to the right of the LDs (whom they correctly perceived as being the most europhile of the three parties).
Acknowledgements
We wish to thank the ESRC for funding CREST and the British Election Panel Study. Philip Cowley made very helpful editorial suggestions and we are also indebted to our colleagues in CREST, John Curtice, Geoffrey Evans, Lindsay Brook, Pippa Norris, and Mandy Roberts.
References
Curtice, John K, and Steed, Michael (1997) 'An analysis of the voting', in David Butler and Dennis Kavanagh (eds) The British General Election of 1997. London: Macmillan.
Heath, Anthony F., Geoffrey Evans and Jean Martin (1994) `The measurement of core beliefs and values: the development of balanced socialist/laissez faire and libertarian/authoritarian scales', British Journal of Political Science, 24: 115-32.
Heath, Anthony F., Bridget Taylor, Lindsay Brook and Alison Park (forthcoming) 'British national sentiment', British Journal of Political Science.
Return to CREST Page