I recently looked at Labour/Lib Dem tactical voting since 1983 and noted that the analysis didn’t work quite so well for 1997.

In 1997, the Lib Dems won 46 seats. This was up 26 on their total of 20 in 1992 and was achieved despite their share of the GB vote falling by 1.2 percentage points (pp). To what extent was this made possible by tactical voting? Boundary changes for the 1997 general election make this a little tricky to answer, but broadly we can draw some conclusions.

The table below shows the change in vote share for seven types of seat in 1997. It also shows the increase in the absolute number of Labour and Lib Dem votes as a percentage of the fall in Conservative votes. In total, the increase in the Labour and Lib Dem vote was only just over a quarter of the fall in the Conservative vote. But where it mattered, the percentage was higher.

a) Seventeen seats won by the Lib Dems in 1992 and held in 1997.

b) Sixteen seats gained by the Lib Dems in 1997 that would have been lost by the Conservatives on UNS. Brecon and Radnorshire, which Labour would have won on UNS, is included here.

c) Eleven seats gained by the Lib Dems in 1997 that the Conservatives would have held on UNS. Eastleigh and Newbury are not included here (see below).

d) Eastleigh and Newbury, which were by-election gains held by the Lib Dems in 1997. Boundary changes made these seats quite a bit smaller, which is why no figure is presented for the increase in the Labour and Lib Dem vote as a percentage of the fall of the Conservative vote.

e) Christchurch, which was a by-election gain not held by the Lib Dems in 1997.

f) Ten seats held by the Conservatives with a margin of less than 5 percentage points over the Lib Dems in second. It does not include Christchurch (see above) or Mid Dorset and North Poole, which was a new seat in 1997.

g) Eight seats gained by Labour in 1997 having been third behind the Conservatives and Lib Dems in 1992.

Changes are measured against the nearest equivalent seat in 1992.

Seats successfully defended in 1997

The Lib Dems won 17 seats in 1997 that were the same or quite similar to seats they won in 1992. Their share of the vote remained the same in total across these seats. However, this hides some quite big differences. In Tweeddale, Ettrick and Lauderdale, Labour came within 1,500 votes of gaining a seat in which they were fourth in 1992. This may have been partly due to the retirement of David Steel. At the other end of the scale, Labour actually share fell by around 1.5pp in Orkney and Shetland.

It’s worth noting that the Lib Dems “lost” three seats that they had won in 1992. All three were affected by boundary changes. Having won an exceptionally close four-way battle for Inverness, Nairn and Lochaber in 1992, Russell Johnston stood down in 1997 and the Lib Dems were soundly beaten by Labour and the SNP in the successor seat.

Also standing down in 1997 was David Alton, who won the 1979 Liverpool Edge Hill by-election before defending the seat at the general election two months later. He won the successor seat of Liverpool Mossley Hill at the following three general elections. Unsurprisingly, the Lib Dems didn’t get close to winning any of the three seats that consumed Mossley Hill in 1997.

In Rochdale, the incumbent Lib Dem MP Liz Lynne was defeated by Labour. This was one of only two seats won by the Lib Dems in 1992 in which Labour came second. In the other, Simon Hughes held off Labour in North Southwark and Bermondsey.

Lib Dem gains on UNS

In 1992, the Lib Dems were second to the Conservatives in 145 seats. Applying uniform national swing (Conservatives down 11.4pp, Labour up 9.1pp and Lib Dems down 1.2pp) would have resulted in the Lib Dems winning 17 of these seats and Labour gaining six.

In reality, the Lib Dems gained 16 of these 23 seats including Brecon and Radnorshire which Labour would have gained on UNS. Having gained Littleborough and Saddleworth in a by-election, the Lib Dems failed to hang on to the successor seat in 1997. In that case they were competing with Labour, so anti-Tory tactical voting was not such a factor. In Eastbourne, the Lib Dems missed out to the Conservatives by 2,000 votes with Labour and the Referendum Party the main recipients of the fall in the Tory vote. However, Wikipedia indicates that boundary changes meant that this seat was just beyond the Lib Dems on UNS.

Whereas the Labour vote increased by 6.2pp in the seventeen seats retained by the Lib Dems in 1997, the Labour vote increased by 3.8pp in the 16 Lib Dem gains. Once again, there was a lot of variation. The Labour vote increased by 10.7pp in Portsmouth South, while it fell by 0.8pp in St Ives.

Lib Dem gains beyond UNS

The Lib Dems gained 13 more seats in 1997. These included two seats gained in by-elections (Eastleigh and Newbury), four seats in South West London (on new boundaries), and seats as varied as Lewes and Sheffield Hallam.

Unsurprisingly, the Lib Dems did especially well in the three seats they had gained at by-elections. In Christchurch, the Labour vote actually fell 5.2pp, indicating clear evidence of tactical voting that saw the Lib Dems fall 2,000 votes short of the Conservatives.

Excluding Eastleigh and Newbury, the increase in the Labour vote was 3.0pp in these seats, so lower still compared with Lib Dem holds and easier Lib Dem gains. The change in the Labour vote ranged from a fall of 6.6pp in Sheffield Hallam to an increase of 10.3pp in Colchester where 1,500 votes separated the Lib Dems in first to Labour in third.

In the 10 seats the Conservatives held by less than 5pp over the Lib Dems in 1997, the Labour vote increased by 4.7pp. I think it’s fair to say that the extent to which Labour supporters were persuaded to lend their votes to the Lib Dems did make some difference. And, of course, the increase in the Labour share of the vote was considerably lower in most of these seats than the national change.

However, there are other factors that affected the prospects of the Lib Dems in the longer shots. In the 11 seats gained beyond UNS, the Conservative vote fell by 14.3pp. By contrast, it fell by 11.9pp in the 10 seats held by the Tories by less than 5pp. Furthermore, in those seats held by the Conservatives, the Referendum Party performed better. Local factors may have also made a difference. In Harrogate and Knaresborough, for example, the desire to defeat Norman Lamont may have contributed to the 18pp increase in the Lib Dem vote.

Labour gains from third

As well as defeating the Lib Dems in Oldham East and Saddleworth, Labour picked up Conwy, Falmouth & Cambourne, Bristol West, Leeds North West, and Shrewsbury & Atcham as they would have with UNS. The largest deficit to the Lib Dems that they overcame in those seats was 5.9pp in Bristol West (compare with the 9.5pp gap between them in Brecon and Radnor in 1992). However, Labour didn’t stop there. They also picked up Harwich, St Albans, and, most stunningly of all, Hastings and Rye where the Lib Dems had been 19.5pp ahead of Labour in 1992.

Overall, in these eight seats, Labour’s vote increased by 13.2pp, while the Lib Dem vote fell by 4.8pp. The polls may have given voters in these places confidence that Labour would win the election comfortably and were prepared to risk ending up with a Tory MP.

Conclusions

The Lib Dems certainly outperformed UNS in the 1997 General Election. Even if we give the Lib Dems the eight seats won by Labour from third place in 1992, that’s still fewer than the 11 seats the Lib Dems won beyond UNS. However, it wasn’t entirely due to potential Labour voters backing the Lib Dems instead. Just as important was the ability of the Lib Dems to win over Conservative voters in areas like South West London.

Tom Leveson Gower

Tom Leveson Gower posts on PB as TLG86