What we can learn from the factors that broke UK politics

Early headlines from the 4 July elections in the UK focused on Labour’s thumping win, one of the biggest majorities in a long time.

Keir Starmer’s moment, it seems.

But not so fast.

In terms of vote share, Labour with 33.7% had a swing of only 1.5% compared with 32.2% at the last election in 2019 when they were so soundly trounced that Jeremy Corbyn was forced to step down as leader. In fact, in the election before that, in 2017, Labour under Corbyn scored 40%.

Cyril Ramaphosa will be green with envy. With a smaller vote share than the ANC’s, Starmer has had a resounding victory with no GNU in sight. A gnu is what they call a wildebeest in the UK.

If you look at the vote count, it’s even more stark. In 2019, supposedly a horrible result for Labour, under Corbyn they scored 10 269 051 votes. In 2024, the Labour vote was 9 712 011, a drop of more than 5%. Part of this can be explained by a lower turnout, 60% in 2024 versus 67% in 2019. 

But whichever way you look at it, Corbyn’s Labour had a lot more enthusiasm behind it than Starmer’s. Which is not terribly surprising, because Starmer moved the party hard to the right, including cleansing itself of “antisemites”, essential Labour members who oppose Israeli policies — a good fraction of whom are Jewish.

So what is the explanation?

The big story is not a surge of popularity for Labour but fragmentation of the Tories (the Conservative Party). The Tory vote was split three ways: those who stuck with the party and two groups who were disgusted with it. The latter group split into those who went to Labour as Tory Lite and those who went for the Reform Party, a warmed-over version of the Brexit Party.

Another big gain for Labour was the collapse of support for the Scottish National Party, which was punished after a wave of scandals.

But it did not all go Labour’s way. The Liberal Democrats staged a massive comeback after being pulverised for taking part in a coalition with the Tories in 2010, winning 72 seats after holding only 11 in the previous parliament. The LibDems were founded in 1988 as a merger of the old Liberal Party, which waned in influence when Labour grew, and the Social Democrats who split from Labour in 1981.

However, Labour also lost on its stance on Gaza, which follows the US line — Israel, wrong or wrong. Five independents were elected in Labour seats on a pro-Gaza protest vote, including Corbyn, who ran as an independent after being barred from running as a Labour candidate by Starmer. 

Starmer himself suffered a significant loss of votes in his own constituency of Holborn and St Pancras to former South African ANC MP Andrew Feinstein, who campaigned against genocide in Gaza. In 2019, Starmer received 36 641 votes, 64.5% of the total, a thumping majority by any standards. This year, Starmer received 18 884 votes, just over 50% of his previous total, for about 49% of the total. Feinstein scored over 7 000 votes and the Green vote also increased. 

A big part of the drop in total votes was the reduction in turnout, over 18 000 fewer than in 2019. Whatever the reason for these changes, the vote in his own constituency is hardly a ringing endorsement of Starmer’s leadership.

There is other evidence that Starmer’s position on Gaza and general rightward shift cost Labour votes, including an increased Green vote, with an increase from one to four seats.

Starmer enters government with a massive majority, even if the thumping was mainly between forces of the right – sometimes barely harder right than Starmer himself. But what is the lesson for others?

Clearly the UK electoral system is broken. For someone to hold nearly two-thirds of the parliamentary seats with a little over a third of the vote is absurd — even if his major opponents formed a circular firing squad. Given that Starmer’s purged Labour Party is not much different from the Tories on a good day, you might wonder why this is a big problem. After all, if the policy differences aren’t that great, why shouldn’t the least chaotic bunch win?

It’s the principle of the thing. Government of the people, by the people, for the people should start from a representative legislature, which the UK House of Commons clearly is not.

On the other hand, the South African proportional representation system, while ensuring fair representation, does not result in constituency MPs — elected representatives that are directly responsible to a portion of the electorate. The nearest we have to that is ward councillors at municipal level.

I have before proposed a constituency-based proportional representation (PR) system; this is doable and is worth considering. I present it here in a simplified form that I hope is easy to understand. I only describe the process for the national vote but the same principles apply to provincial legislatures.

The basic idea is that voting occurs in constituencies, as in the UK system, but instead of a first-past-the-post result, the nationwide count is used to determine results in two steps.

In the first step, all votes across the country are counted and pooled for each party. Independents at this stage are all grouped together as a “party”. Once this stage is complete, the seat count for each party (and all independents as a group) can be calculated, exactly as it is in the present PR system.

In the second stage, the votes for each candidate are sorted within their party from highest to lowest. This forms a PR list for each party. Independent candidates are placed in one list, on the same basis as a party. From the calculation in the first stage, each party is allocated seats from the PR list that is calculated in the second stage, from the top down, until their allocation is filled.

The effect of this is to switch from a closed party-list PR system, in which the voters have no say in the list order, to an open party-list PR system in which the voters determine the order.

There are several benefits of this approach. Everyone in parliament will be elected from a particular constituency, making them directly accountable to those voters. If an MP leaves parliament, it becomes possible to have a by-election. 

Independents can also naturally be accommodated, whereas they do not fit the concept of a closed-list system. A closed PR list also has a big problem: it is susceptible to internal party corruption to manipulate a candidate’s position on the list.

A drawback of this approach is that the process of allocating seats does not guarantee exactly one MP per constituency. Some could end up with more than one, others with none. However, any party aspiring to win votes in a constituency with no MP will have the incentive to work that constituency with a view to winning over those voters next time.

We have been living with a pure closed-list PR system for 30 years; can we honestly say it works as well as it should? On the other hand, envious though Ramaphosa might be of Starmer taking nearly two-thirds of his parliament with a third of the vote, do we really want that either?

So, how about it? My idea for combining a PR system with a constituency system will not be hard to implement — the hardest issue will be splitting the country into 400 areas of roughly equal numbers of voters but we already do something similar to create municipal wards. 

The nomination, ballot printing and vote counting systems will not be hard — they will be of similar difficulty to the ward component of a municipal election, though with fewer individual counts as there will be only 400 constituencies; there are over 4 000 wards. Creating a national PR list out of constituency votes will not be difficult, nor will calling the final result.

But, overall, there will be big benefits — a closer connection between MPs and their constituents, a more natural model for independents and the possibility of by-elections when someone leaves parliament. But the biggest benefit overall will be taking internal party corruption out of deciding PR lists.

Philip Machanick is an emeritus associate professor of computer science at Rhodes University and a Makana Citizens Front councillor.