The British General Election of 1997 
by David Butler and Dennis Kavanagh.
Macmillan, 343 pp., £17.50, November 1997, 0 333 64776 9
Show More
Labour's Landslide 
by Andrew Geddes and Jonathan Tonge.
Manchester, 211 pp., £40, December 1997, 0 7190 5159 2
Show More
Britain Votes 1997 
edited by Pippa Norris and Neil Gavin.
Oxford, 253 pp., £12.99, January 1998, 9780199223220
Show More
Collapse of Stout Party: The Decline and Fall of the Tories 
by Julian Crtitchley and Morrison Halcrow.
Gollancz, 288 pp., £20, November 1997, 0 575 06277 0
Show More
Les Election Legislatives, 25 Mai-1er Juin 1997: Le president desavoue 
Le Monde, 146 pp., frs 45, June 1998Show More
Show More

No one this time last year would have predicted a victory for the Left in France, yet it is in a sense far easier to explain Jospin’s triumph than Blair’s. President Chirac, elected in 1995 on a promise that he would reduce unemployment, had actually done the opposite; and, faced with the need of his deeply unpopular prime minister, Alain Juppé, to squeeze the economy yet further to meet the Maastricht criteria, called a snap election simply because he feared things would get tougher by the time Parliament’s mandate ran out in 1998. He was eager, too, to take advantage of the fact that the Socialist Party was still in a mess after the devastating unravelling of the Mitterrand Presidency and its own catastrophic defeat in the 1993 election, which left the Party with 56 seats in a chamber of 577 Deputies.

French voters, however, were indignant at being asked to perform ‘au bon plaisir du président’, as Le Monde put it – something which Jospin grasped more quickly than most. ‘What we are going to do,’ he said succinctly, ‘is to take our democratic revenge.’ In the course of the five-week campaign the election simply ran away from Chirac and the PS swept to power. It won 246 seats and the rest of the Left a further 74. To most commentators, it seems, all this came as a great surprise, yet there was nothing very remarkable in the voters’ sharp reaction to record unemployment, broken promises and Presidential arrogance, particularly in a country where the most deeply felt political sentiment has always been the wish to ‘sortir les sortants’ – to throw out whoever’s in power.

The British situation was exactly the opposite. Parliament had run its full term, the Prime Minister was generally liked, the economy had been growing strongly for five years, inflation was low and unemployment had dropped steadily throughout the Parliamentary term. Interest rates were low, income tax had been cut, the balance of payments was in surplus and the currency was strong. In sum, the Government’s boast that Britain’s economy was in better shape than any other in the EU looked reasonable enough. It was a boast no other government in the post-war period had been able to make and it would have been taken as axiomatic that any government that could make such a claim would be the safest of shoo-ins. Yet Labour, which had taken a lead in the polls in late 1992, never lost it thereafter and always looked like the massive winners they ultimately proved to be. Given the conventional wisdom about the centrality of the ‘feel-good’ factor, it was an astonishing result.

It was in 1945, that other great Labour landslide year, that the first Nuffield election study appeared, with the hand of a young David Butler already apparent. It was, sadly, a jejune and disappointing volume which explains very little of the last great realigning election of modern times. Since then Butler has taken an increasingly dominant role in the enterprise and the books have improved accordingly. By the late Fifties the Nuffield studies had become part of the election ritual and David Butler himself a national monument. Extraordinarily, he has now been doing this job for over fifty years. Yet by the Sixties it had already become fashionable to sneer at the sheer abstruse weight of Butler’s empiricist style. For most students of electoral behaviour the analysis of the result is the main point of the whole thing, but Butler has been happy to farm this section out to others and has even coined a new word – psephology – to describe his own preoccupation with the election-as-event. The term has often met with derision elsewhere. ‘Ah, Professor Stokes, you’re a psephologist, aren’t you?’ the Queen once said to the great American political sociologist, Donald Stokes, who collaborated with Butler in the Sixties and Seventies. To which Stokes, who was fond of telling the story, replied: ‘Absolutely not, ma’am.’

Impressed by the Nuffield studies, French political scientists based at Sciences Po (the Ecole Nationale des Sciences Politiques) and Cevipof (Centre d’Etude de la Vie Politique Française) decided in the mid-Sixties that they would do the job properly, with more thorough and more sociological accounts of de Gaulle’s great electoral contests – not just the Parliamentary and Presidential elections but the referenda too. Their collective expertise was impressive, but they were no match for Butler in terms of speed. Their chercheurs were still labouring away on the 1967 election when de Gaulle called the climactic contest of 1968. (Butler, by contrast, produced two large and separate volumes on the twin elections of 1974.) Unable to cope, the chercheurs threw their hands in the air and the entire project collapsed. Since then there has not been a single book-length academic study of a French election. It is now evident that Butler’s (and, since 1974, Dennis Kavanagh’s) work is irreplaceable: each volume includes things a political historian needs to know and can find nowhere else. The lack of a stronger sociological and analytic perspective remains a vexation – the French are undoubtedly better at this – and the sense of a partially wasted opportunity haunts every Nuffield study. But in the end there’s no denying that Butler and Kavanagh have outrun and outlasted all the clever men of Sciences Po and Cevipof.

The French academy’s loss was the media’s gain. One reason it was never possible to reassemble French politicologues for an election book was that French newspapers and broadcasters have shown an enthusiasm for sophisticated sociological analysis of electoral behaviour which their counterparts in the UK would have claimed was far above the heads of most of their readers. Every major French newspaper, polling agency, political party and radio or television station has its own expert political sociologist.

The politicologues quickly realised that they could earn far more writing in newspapers than an academic election study could ever offer them and with that discovery all hope of a French equivalent of the Nuffield series died. Newspaper readers clearly appreciate their work and as a result the level of media coverage and popular understanding of such matters in France has been markedly higher than it is in Britain. Typically, the best analysis I have found in Britain has been in one or two large post-election articles by Ivor Crewe in the Guardian, but in France such pieces were legion before as well as after elections, not only in Le Monde, Le Matin and Le Figaro but in the mass-circulation weeklies such as L’Express, L’Evénement du Jeudi, Le Nouvel Observateur and so on. Several of the major papers began producing their own election books.

The best of these has generally been the volume produced by Le Monde. The practice established in the first such production, covering the 1973 Parliamentary elections, was to include the best of the newspaper’s own coverage over the campaign period, together with a full analysis of voting trends and results and – most usefully – a constituency-by-constituency description of the battle, often in hilarious detail. Unfortunately these often juicy details have now largely been dropped, though the analysis of constituency results is still fuller than it is in Britain. In fact, Le Monde now brings out no fewer than three companion volumes, the Guides du pouvoir: in 1997 a 128-page set of biographies of all 577 Deputies, a further 128-page analysis of the make-up of the various cabinets ministériels (including civil servants and advisers as well as politicians); and an 800-page discussion of all the major French political institutions, together with photos and biographies of the ‘4000 principal public personalities’.

Any election represents a huge assertion of social behaviour and the great question is bound to remain: what did it all mean? There is no real consensus among political sociologists as to how we should go about answering this. Originally commentators were inclined to envisage a collective electoral mind (from which to deduce that the electorate was saying this or that). The problem was that there is no such thing – an election is the aggregation of millions of acts of social behaviour. This was brought home to one and all by the work of André Siegfried, the father of modern electoral analysis. Siegfried mapped French election results and compared them with known distributions of other social phenomena, and the results were startling. The map of France showing which regions supported and opposed the 1789 Revolution was also the map which best indicated the distribution of support between Right and Left (under a succession of changing party labels) more than a hundred and fifty years later – and this in turn was largely a map of differential regional patterns of church attendance. Voters, it became clear, made their choice according to region, class and religion. And if one found such groups moving in different directions, how did one aggregate that into a single national electoral mind?

Siegfried’s work was, however, vulnerable to the so-called ecological fallacy. If you showed that the heaviest distribution of redheads was in Normandy and that Normandy had swung most strongly to the right, it might seem obvious that redheads were more right-wing than, let’s say, blonds, but of course one cannot make such an inference: for all you knew there had been a sufficiently large rightward swing among non-redheads in Normandy to mask a leftward swing among redheads. In a word, the unit of analysis was too large.

This weakness has been overcome with the emergence of modern survey research and opinion polling, which has revolutionised our understanding of electoral behaviour. But it, too, has weaknesses. It is expensive and subject to sampling error; and its data are only as good as the questionnaire used – and a lot still turns on how they are analysed. For example, in 1992 Clinton’s pollsters found that the great majority of those who had voted for Ross Perot said they would otherwise have been just as likely to support Clinton as Bush. Heartened by the thought that this meant there was a real pro-Democrat majority, Clinton pushed ahead with a liberal Democrat agenda until he was stopped in his tracks by the disastrous 1994 mid-term elections. His new pollsters looked again at the 1992 data and concluded that the bulk of Perot voters had been defecting Republicans so fed up with Bush’s tax increase that half of them were considering the ultimate treachery of voting Democrat. Even then they hadn’t done so and most had since homed back to the GOP. All the mid-term elections had done was to confirm that the electorate had the same ‘natural’ Republican majority in 1994 as in 1992 – it was this crucial new analysis of the data that provided the motivation for Clinton’s sharp shift to the centre and his ultimate victory in 1996.

The other problem with survey data is that they are very far from being the same thing as the election result: in fact, they have sometimes failed even in gross terms to predict that result. Hence the knobbly British tradition of constituency studies, which attempt to see how things worked out on the ground. The Geddes and Tonge volume makes a gesture in this direction with one-page summaries of local fights interleaving its various more general essays, but the key data are to be found in the exit polls it reproduces. These show that Labour owed its victory to the fact that it held 90 per cent of its 1992 vote (compared to 71 per cent for the Tories and 66 per cent for the LibDems), while taking 55 per cent of previous non-voters and 57 per cent of new voters. Much of the euphoria and the feeling of major social change that the Blair victory – however briefly – inspired was due to the fact that a new generation was taking power. For Labour’s vote was terribly young. Among the 18–29s the swing to Labour was over 20 per cent while the Party managed to lose 2 per cent among the over-65s. It was also middle-class: the pro-Labour swing was 16 per cent among home-owners, but only 3 per cent among council tenants; 16.5 per cent among non-manuals, but only 9 per cent among manual workers. Blair’s victory was in every sense a yuppie triumph.

The Butler-Kavanagh volume also depends heavily on constituency analysis, with every constituency result set beside its income levels, unemployment rate, population density and the managerial proportion within it. This doesn’t really work: readers are not going to do the necessary regression analysis to make sense of it. John Curtice and Michael Steed also produce an immensely careful and laborious analysis of constituency results, their main conclusion being to confirm that the economy was not the master issue which moved the electorate. The Norris and Gavin volume contains a great deal more equally detailed but essentially short-term analysis. There is, for example, a tremendous fascination in these election books with the question of how well the polls did in predicting the result. This has as much bearing on the actual meaning of the election as yesterday’s weather forecast on the phenomenon of global warming. Besides, if you spend too much time ploughing through such leaden and unrelievedly earnest material you may seriously damage your brain. One has to pick up the Critchley volume to remind oneself that politics is often funny. Critchley bids an appalled farewell to his colleagues, ‘Mother’ Teresa Gorman, ‘Little Norm’ Lamont, ‘Big Norm’ Tebbit and all the rest:

John Carlisle, the right-wing MP for Luton North, is vin de table ... I hope he will take up residence in a cardboard box in Luton Airport. Steve Norris, the party’s chief garagist, will be found on the forecourts of the East End. John Patten will not be missed. At best a ‘third growth’, he will probably teach geography at some secondary school in a deprived area of a Northern city. On the other hand, he might be appointed manager of Hartlepool United. Dame Elaine Kellett-Bowman will be much missed. A feisty lady with a good left hook ... shrill of voice and short of temper. I have been keeping her for years.

And so on.

We still await a British version of Fear and Loathing on the Campaign Trail and it is probably too much to hope that those who write our election studies will lighten up a bit. But what gnaws at one even more is the feeling that one is missing the big picture. In one sense these studies show the British empirical tradition at its best: there is endless detail about how much impact the Referendum Party made, what the results might have been like under other electoral systems. It is in the nature of this sort of analysis that it sheds very limited light on trends: indeed, it is liable to give the impression that elections are simply too complicated to yield any underlying meaning.

The key to the British election of 1997 and the uniquely large swing it produced in defiance of economic considerations was surely the notion of pent-up change. We have seen this before – in 1945 and 1979. The huge 10.3 per cent Labour swing of 1997 is simply the most recent example. The key point was that by 1990 voters were heartily sick of Thatcher and Thatcherism and Labour ran steadily ahead in the polls by 10-20 per cent. Even Thatcher’s replacement by Major did no more than reduce the gap to near parity and Labour went into the 1992 election with the polls forecasting a comfortable win. In the event, faced by Kinnock’s sheer lack of credibility, voters baulked and the Tories won comfortably with a lead of over 7 per cent. But the minute Smith succeeded Kinnock, Labour shot into a lead which was consolidated beyond recovery by the effect of the forced exit from the ERM in September 1992. The Government had endlessly told the electorate that leaving the ERM would be a disaster and had even raised interest rates 5 per cent in a day in order to stay in. Then, when we fell out, low inflation, faster growth and lower unemployment ensued. This was why the Government was given no credit for the economy: it had got it right only by mistake and against its will. The lesson learned by many Tories was that Europe was Entirely A Bad Thing. This only made matters worse, for it guaranteed that the Party would be split.

The strange symmetry of 1979 and 1997 was that in both cases voters were making up with a vengeance for the fact that, at a previous election, they had failed to take the opportunity to vote against something they vehemently disliked, a deeply aggravating fact as the years rolled by. When parties get elected against the odds like this, their factions learn all the wrong lessons – essentially that they can get away with almost any amount of self-indulgent behaviour and still win. Victories create a new set of myths – in the current case that Labour owes everything to Blair and the new yuppie style. Actually, Labour would have won crushingly under Cook or Brown in 1997. And, of course, whatever the New Labour style, Labour still rests on the votes of the poorer half of the population, with certain immanent consequences that Blair is bound to discover. But very few politicians are at all good at understanding their own victories: the one real constant seems to be the arrogance of power, endlessly reproducing the need for us to ‘take our democratic revenge’ – sortir les sortants.

Send Letters To:

The Editor
London Review of Books,
28 Little Russell Street
London, WC1A 2HN

Please include name, address, and a telephone number.


Vol. 20 No. 3 · 5 February 1998

All elections generate their own myths and R.W. Johnson merely repeats the most recent (LRB, 22 January): that the Tories lost even though the economy was doing well. He ought to ask: whose economy? His confusion is only possible because, despite listing no fewer than seven economic indicators, he omits wages, the most important of them all. For millions in the public sector, for instance, several consecutive pay freezes ensured that there was to be no share in any extra benefits the economy was producing, to compensate for blows such as large increases in indirect taxation – another factor he neglects. Even the advantages of lower mortgage rates were often cancelled out by fears of negative equity, while the fate of those having to pay private sector rent increases seems beyond the ability of Johnson, or any other commentator I have read, to recall. Many people work far harder than they used to, yet have less money to show for it. That reality is what matters, not any ‘feel-good factor’ or lack of it, or any ‘perception’ about whether the Tories got it wrong over the ERM. Johnson does not mention the word that matters most today: ‘inequality’.

Ed Horton

send letters to

The Editor
London Review of Books
28 Little Russell Street
London, WC1A 2HN

Please include name, address and a telephone number

Read anywhere with the London Review of Books app, available now from the App Store for Apple devices, Google Play for Android devices and Amazon for your Kindle Fire.

Sign up to our newsletter

For highlights from the latest issue, our archive and the blog, as well as news, events and exclusive promotions.

Newsletter Preferences