Lewis Baston is author of Reggie: The Life of Reginald Maudling and several books about British general elections. He is a consultant on politics, elections and constituencies.

Last month, I wrote about the events of 30 years ago: the Westland scandal that divided the Thatcher government and caused two Cabinet resignations in early 1986. I allowed another anniversary to pass by, namely, the 110 years since the last Liberal overall majority was elected in the landslide of 1906. It is just about feasible that the same individual may have been able to vote both in the 1906 election and in the 1987 election that followed Westland – in the latter case as a man aged 103 or so, but there are few other threads linking the two.

The events of the 1906 campaign, and the broad shape of the results, are a familiar enough story. The Conservatives, tired, divided and beset by resignations after a long spell in government, stumbled into one of the worst defeats in their history. A chain of consequences was set in motion that led to the establishment of the welfare state, the rise of the Labour Party, Irish independence and the supremacy of the Commons over the Lords. It was quite an important election in historical perspective, but the coming of the 1906 election was the result of a rather strange manoeuvre that suggests that it was a bit of a game that got severely out of hand.

We are accustomed to the expectation that a sitting government calls an election at the end of its parliamentary term, and that if the election result is negative the Prime Minister then tenders the resignation of the government, or in some cases remains in office to explore the possibility of forming a coalition (1974, 2010) or faces parliament to be formally defeated (1886, 1892, 1923-24).

Occasionally, though, a novel variation can happen in that a government resigns and puts the other side in before an election takes place, even without suffering what would generally be regarded as a defeat on a vote of confidence. Gladstone and Rosebery did this in 1885 and 1895, essentially turning the keys of Downing Street over to Salisbury. It was therefore odd, but not unprecedented, for the Conservative Prime Minister Arthur Balfour to return them to the Liberals without a fight in December 1905.

Party leaders and strategists are often not very good at assessing events and tendencies in the other party. They sometimes believe their own propaganda and either underestimate their opponents, or else endow them with imagined near-supernatural powers of political manipulation in order to find psychologically satisfying explanations for losing. This happened even in the clubland environment of Edwardian politics where, despite the vicious intensity of party conflict ,the Liberal and Conservative elites were socially intertwined. The relative ease with which Campbell-Bannerman formed a broadly united Liberal government should not have come as a surprise, but it seems to have done. The issues that divided the Liberals in 1900 had slipped down the agenda and, by promoting Tariff Reform, Chamberlain had at once divided the Unionists and pushed the Liberals’ most unifying cause of Free Trade to the top of the agenda.

Free Trade also gained the Liberals the tacit or overt support of small-c conservative interests that normally favoured the Unionist cause, in the press and among the leaders of local society in rural and small town England. The Liberal Party organisation had revived strongly since 1900, and the party was ready to field an almost full slate of candidates in Great Britain, in contrast to 1900 when 163 Unionists were returned unopposed. The Liberals benefited from a surge in turnout and enthusiasm.

The Liberal landslide was enormous. The Conservatives and Liberal Unionists fell to 157 MPs, the worst-ever showing on the right hand side of the aisle (a little worse than 1997, when there were 165 Tories elected), against 400 Liberals and another 113 MPs from the Liberals’ allies among Labour and the Irish Nationalists on the other side. The electoral system, as it does sometimes, had produced a rather strange transformation in the votes cast, because the swing to the Liberals was greatest where they needed it most. Although on the face of it the Liberal lead in vote share was comparatively modest – 49 per cent to 43.6 per cent for the Unionists – the electoral pact with Labour means that the Lib-Lab combination should probably be counted together to give the left 54.9 per cent support. But the reward in seats was still particularly high.

The principal holdouts of Conservative support in 1906 were Chamberlain’s fortress of Birmingham, although the Liberals did make inroads in the wider West Midlands; Liverpool and west Lancashire; central and west London; and Ulster. The 1906 election was the only time, ever, that the Tories have lost their majority of seats in south east England, with Liberal gains appearing in strange seats such as Chertsey, East Grinstead and Henley. These gains were transitory, and reversed at the next election in January 1910, but some aspects of the political map of 1906 stuck. Liberal strength in Lancashire and the west of Scotland was maintained in 1910 when southern England and rural areas flipped back to the Unionists; the map of 1910 was much more about class politics than the map of 1885.

The Balfour gambit was not quite as suicidal as it may seem to us now. While Rosebery’s resignation in 1895 led in short order to electoral disaster, Gladstone’s resignation in June 1885 had more ambiguous consequences because the Conservatives governed ineffectively for a few months before the election and the causes of Gladstone’s unpopularity in the summer had receded into memory. The three acts reforming elections passed during 1880-85 – franchise extension, new constituency boundaries and tough laws against corruption – all meant that calling an election was more uncertain than usual, and that the Liberals would be likely to benefit. The Liberals won a big majority in Great Britain, something that had seemed very unlikely a few months previously, and it was only because of the fluidity of the Irish situation and British high politics that it all fell apart again so quickly in 1886.

Another factor making Balfour’s gambit seem reasonable was that Britain in 1906 was not a full democracy. The value of the prize at stake in an election was different for each party. A Conservative government essentially got to do whatever it wanted, while a Liberal government had to tiptoe around the power of the Lords with its entrenched wealthy, landowning interests and its outsized Unionist majority of 479 peers to 88 Liberals. The Tories had a back-stop against any Liberal measures that came too close to threatening their interests, and they used it militantly during the parliament of the Liberal landslide. They responded to their landslide defeat in 1906 with a notable lack of humility, Balfour stating that ‘the great Unionist Party should still control, whether in power or opposition, the destinies of this great Empire,’ and F.E. Smith making his career with a maiden speech brimming with wit, defiance and chutzpah. If the Tories had won in 1910, after hobbling a majority Liberal government, Balfour’s resignation in 1905 might have seemed a brilliant longer-term version of what Gladstone had worked in 1885-86. It was not until the Parliament Act of 1911, which might not have happened had they been more restrained in 1906-11, that losing an election became a particularly serious matter for the Conservatives. It may be noted that the Tories got a lot better at winning elections from 1918 onwards.

The change of government in 1905-06 was probably the last time that a Prime Minister underestimated the importance of the power to dissolve Parliament, and indeed the inherent power of being the tenant of Downing Street. Part of the choreography of elections is to make it difficult for people to imagine the opposition party taking power, and to exploit the aversion to change among many electors. Admitting that you’re having trouble governing, getting the Leader of the Opposition to try on being Prime Minister for size and him then announcing a few popular policies and call an election…it seems to break all the rules. But in rejecting the Balfour Gambit, our modern understanding of politics actually concedes that the other side might do a decent job of governing for a bit rather than being a chaotic disaster. If a Prime Minister really believed that the opposition were unelectable and incapable of governing, elections would be more frequent and would sometimes involve the Balfour Gambit.

But wait a moment. The rules about dissolutions were changed in 2011 by the Fixed Term Parliament Act. Early elections are still entirely possible, but they now require some parliamentary procedures that would have looked strange to Edwardians – major-party collusion to pass an election motion, or an engineered no-confidence vote. In the event of a hung parliament, an updated version of the Balfour Gambit might come into play either by accident or design. Imagine Government Party A, in the depths of unpopularity and recession, after successive Commons defeats, deciding to let Government Party B have a go. By exhausting the possibilities of the existing parliament, it would be in the (easily subverted, but still) spirit of the Fixed Term Parliament Act, and might have political benefits. Party A could prevent Party B from calling an election during the honeymoon period of the new government, and allow it to share a bit of the blame for bad times and get a buffeting from the smaller parties in parliament. Perhaps in 20 years’ time it will be easier to understand what Balfour, the mildly decadent author of A Defence of Philosophic Doubt, was up to in December 1905. Like the famous flap of the Amazonian butterfly’s wings, Balfour’s languid gesture may have had far-reaching consequences.