Only around 10% of humanity still live in so-called ‘full’, ‘liberal’ democracies. That fraction is shrinking. Yet again, leaderism and authoritarian regimes are spreading. Their rise is fuelled partly by the widely-perceived corruption, extreme inequalities, rampant exploitation and the other failings of existing so-called democracies. We urgently need to find more imaginative and successful ways to democratise democracy itself, so we can reverse our dangerous downward spiral into despotism.

UK democracy has long been distorted by our notoriously unrepresentative ‘First-Past-the-Post’ (FPTP) elections, bedevilled by safe seats, wasted votes, and a succession of pro-elite governments supported by only a minority of the population. In a long-overdue effort to fix our gross democratic deficits, the Labour Party recently voted overwhelmingly to introduce some kind of proportional representation (PR), although the leadership seems reluctant. However, existing variants of PR are still beset with multiple problems, which undermine key principles of democracy, representation and equality. 

For example, most big political parties are broad coalitions sprawling over a multi-issue political space. Individual voter preferences on particular issues may not map onto party platforms: however many political parties there are, there are always going to be far more issues. Democracies would be more effective if their voting systems could better reflect this inherent multi-dimensionality.

Existing electoral systems, including PR (even misnamed ‘open’ list PR), give party hierarchies and leaders (often elected by only partially democratic processes) far too much power over candidate filtering, selection and ordering. PR adjustments pull votes away from potential best matches between citizens and representatives, and donate them to poorer matches, perhaps at an opposite wing of the same party, perhaps in a different party. 

In a typical legislative assembly, all representatives have one equal vote each – despite being backed originally by very different numbers of actual voters. Under PR, some imperfect adjustments may occur to somewhat equalise the number of votes per representative, essentially by transferring ‘spare’ votes from elected or eliminated candidates, or by allocating votes via party lists. But most of these corrective measures are hard to grasp, undermining the principle of representativeness. 

In any case, real power is not evenly distributed between those actually elected, so some representatives have more weight than others. Moreover, many are heavily ‘whipped’ to vote along party lines. This flagrantly violates all norms of proxy voting, which is in effect what representatives are doing: voting on our behalf between elections on things they supposedly know more about than we do. Every one of our votes, delegated and entrusted to those representatives, should count the same.

We propose a far simpler, more proportional alternative to standard PR models: instead of just having one equal vote in the assembly, each representative casts the number of votes that elected them, as if they were proxying for those that voted for them until the next election. This straightforward change solves, at a stroke, many of the problems associated with most popular variants of PR such as party list PR, mixed or additional member PR and Single Transferable Vote (STV). (More discussion of existing voting systems, their problems, and alternative ideas, such as referendums, direct democracy, liquid democracy, sortition, citizens’ assemblies, and their problems, together with more details of our solution, are given here).

Our guiding principles are that no vote should be wasted as it is in our FPTP system, but also that, wherever possible, top matches between candidates and voters should be adhered to. Voters should also have more choice over policies, parties and people to implement them. This maximises the fine-grained ‘fit’ between the electorate and their representatives across the huge number of political dimensions and issues.

It is not hard to add up weighted votes, even without modern technology – indeed this has long been standard practice in many situations (for example, at company general meetings, where one-share-one-vote is the norm, different shareholders can own very different numbers of shares, and proxy voting is commonplace). One can simplify the process further by dividing the vote numbers by a convenient power of 10 (such as 1,000), then rounding (e.g. 14,789 becomes 15). Many assemblies already use some kind of within-assembly secure electronic voting. It would be trivial to upgrade such systems for weighted voting.

To keep the total number of representatives and the voting power of each within sensible bounds, we can use multi-member constituencies, each electing, say, 6 representatives. (About 110 constituencies with populations of around 600,000 works for the UK). A given party can field multiple candidates. Voters rank the candidates on their ballot papers. We eliminate the candidate with the fewest votes (first preferences, initially), and transfer their votes to next preferences. We repeat this until the right number of representatives are elected. Unlike with STV, ‘spare’ votes of those elected are not transferred to next preferences. Each representative keeps all the votes that elected them. The detailed preferences of the voters are therefore more closely adhered to than under STV, which gives away votes of popular candidates to less popular candidates, often with different views.

One could also include an ‘minimum vote threshold’, say 3% of the votes cast; candidates with less would be eliminated, their votes transferred to next preferences. This would exclude extremists with very low support, and reduce ‘hyper-fragmentation,’ which can lead small parties to gain disproportionate power. An ‘automatic election threshold’ of say 10%, above which a candidate would always be elected, could counter bigger parties ‘crowding out’ smaller parties or independents. Both measures could change the number of members returned by some constituencies.

Compared with existing systems, our proposal will provide a more exact reflection of each voter’s multi-dimensional preferences. Voters will be able to choose both between parties and within parties. Each voter will be able to pick and rank the candidates (and parties) whose preferences are reasonably close to their own in multi-dimensional political space. For the vast majority of voters, there is a good chance one of these will be elected. Our proposal thus allows a more complete, accurate, proportional and fine-grained coverage of political space than other kinds of PR – and it is far simpler!

We expect votes-weighted representation will generally lead to coalition governments, which is no bad thing, for multiple reasons. Democracy is inherently a series of compromises, and we believe it is best conducted according to the emerging scientific principles of collective intelligence, in order to recruit a wide range of talents and to generate the best ideas. Coalition building can improve governments (at least increasing the available talent pool), and may have a moderating influence, aligning them better to the overall population than under FPTP. 

Another critique levelled at PR is that it can lead to unstable governments – but so can FPTP! A changing government does not necessarily mean an unstable country, if the changes are beneficial or allow useful adaptations to new circumstances. In many northern European countries using PR, governments are fairly stable; these countries are often held up as being well-governed. Anyway, should government ‘stability’ per se be lauded above the will of the people and the good running of the country?

Votes-weighted representation will work well at multiple scales, from local councils all the way up to large countries, and even transnational federations such as the EU. Electoral districts and any election thresholds would have to be sized appropriately to keep the assembly from growing too unwieldy (even with remote online debating and secure electronic voting functionality). At these larger scales, a key advantage of votes-weighted assembly voting is that different multi-member electoral districts can have very different populations, while still achieving fair, flexible, responsive, multi-dimensional proportionality of overall representation.

We are not claiming our proposal will ‘fix’ democracy fully. To do that, we would also need to embed democracy more systematically into our everyday lives. Broader, deeper, more decentralised democratisation is needed across far more of society and the economy, including wherever practicable, in companies and other workplaces. Many existing and proposed voting systems face additional, more general challenges, including fake news and the lack of a diverse free press, an independent judiciary, and a legal system that protects everyone in society, as justly and equally as possible, not just favouring the rich and powerful or the vocal majority. However, compared with existing systems, votes-weighted representation will allow more proportional, more accountable, more responsive and finer-grained coverage of the inherently multi-dimensional political space. This should help democratise democracy, improve its functioning and reverse its global retreat.

Guy Major and Jonathan Preminger are senior lecturers at Cardiff University.