In the earlier parts I've been discussing how the party-system has evolved in different sorts of democracy, and how this reflects the mechanics of the underlying electoral system. I've so far covered political systems employing First Past The Post (FPTP) and Proportional Representation (PR) electoral systems. These are the two basic and most common ways of electing governments (particularly parliamentary ones) so loosely describing their strengths and weaknesses seemed appropriate. I see the electoral system as an engine that drives the party system - release a screw here and you get more parties, tighten a bolt there and you get different issues defining them. But, I've taken an ad hoc approach, and I risk bias towards systems which best represent my own views. Now that we're moving into subtler territory we're going to need some model of what makes a better party-based democracy.
By focusing on how the electoral system defines the party-landscape, I've underplayed how the parties responds to the views of the citizens. Each citizen will have their own unique balance of opinions, with each person sitting somewhere on a broad spectrum of positions. Parties offer the citizens a series of categories that reduce the spectrums into boxes they can vote for. The more boxes you offer, the more subtle a choice they have come election time. The spectrums of opinion are often categorised using the terms Left and Right, or Conservative and Liberal. It's a tricky categorisation because every political position seems to fit somewhere into this single spectrum. And yet these poles aren't set in stone, and the ideas that define them are era and culture specific. An example of an era specific issue would be unilateral nuclear disarmament, which in the British 1980s was at the core of the Left coalition. In the Britain of 2011 neither Left nor Right would consider it worth discussing. Alternatively, viewed from Britain the US-centric issues of abortion and affirmative action that define Liberalism and Conservatism can appear bizarre. In response to this malleability of Left and Right the political compass has emerged in the last few decades as a more accurate way of defining political orientation. This retains the same basic emphasis on Left and Right, but adds Libertarianism vs Authoritarianism as separate axes. It's a nice idea, but my instinct is that this is a result of the rise of internet libertarianism; a cheeky attempt to subordinate the rest of politics to that movement's essential concerns.
In a two party system describing Left and Right appears relatively straightforward because each party ends up embodying the issues that are important to each half of politics - the Hard Left is whatever is believed by the most radical end of the Democrats, the Hard Right is whatever is believed by the radical end of the Republicans. And yet radicalism is often seen in the supposed centre as well, particularly in foreign and financial policy. In part one of this essay I described the American Left and Right this way, as convenient tools for cutting society into two equal-sized blocks. It's beneficial, because these adversarial systems allow society to permanently question the decisions being made in their name. At first you'd think it would be harder to define these poles in a multi-party system. Yet in European multi-party democracies we still often see a division of all political parties into two roughly equal-sized blocks of opinion, and new issues often fall into line with Left and Right groups (the emergence of Environmentalism as a major issue was rapidly absorbed by the European Left). This model of four flavours of politics with Soft, Hard, Left and Right is pervasive. So what does it mean?
A thought experiment - a snapshot of society's views and diversity at one moment in time
In visualising how these two loose and changing worldviews relate to society's diverse opinions let's imagine a thought experiment. Visualise subjecting everyone in a democracy to a survey; a census that collected all the information about their lives and opinions at that moment. Since everyone and every political position sits somewhere on the Left/Right scale, then somewhere in all this data we've just mined there should be something that can identify their position on it.
Let's start with what Libertarians attempted with the political compass: they tried to divide politics into four polar positions on two axis. This polar view could be a useful way to conceive of diversity, so we'll try creating opposed positions to define all diversity; let's start with the obvious ones such as Libertarian/Authoritarian, urban/rural, free market/state regulated, male/female, believers/agnostics, nomadic/sedentary, pro-immigrantion/protectionist. In the political compass the positions were represented as scales, so let's weight these poles onto scalar axes (1-10). We have a few problems here since some of these issues don't fit easily a onto a scale (particularly male/female which is more a binary trait than a spectrum) but we'll overlook this for convenience at this point, this is after-all only a thought experiment. Now, with our infinite pile of survey results we should be able to score every person on all of these scales. Here, we're just taking the idea of the political compass to its next logical step, compensating for its bias towards Libertarianism by adding every other issue that matters. With all of our data and weighted scales we should be able to reduce this to the essential issues in politics, so let's say we've cut this down to 50 axes. Our enormous pile of data can then be plotted on a multidimensional graph, with every person represented as a single point in 50 dimensional space.
Human beings are notoriously bad at visualising more than 3 spatial dimensions, but this situation is common in statistics and would present no problem to a computer. Let's spin around in 50 dimensional space looking at all this data. A lot of these issues seem to relate to each other with high scores on one linked to high scores on another: people who feel strongly in favour of the redistribution of wealth are often found to favour a soft policy on immigration - both issues fit into a philanthrophic view of helping the poor, so these issues co-vary. Alternatively rural people tend to be used to being self reliant, and self reliant people care less about public services than urban people who wouldn't be able to function if the sewers brokedown, so favouring government action tends to relate to urbanism. Within this 50 dimensional space there are some planes along which groups begin to form, linked together by the positive relationship of more than one of our axes.
This approach of hunting for variables that relate to each other is common in statistics, and spinning around in 50 dimensional space we would eventually find a two dimensional plane, cutting through this space, along which the most co-variance occurs. Visualising this surface in broader three dimensional space, we would probably see something that looked like a lumpy sausage, along which almost all of society formed into two groups. The tip to tip dimension of the sausage would have quite a strong element of the redistribution/wealth ownership axis defining it, because this is an issue that represents everyone's self-interests (the poor usually would like more money, the rich would typically like to keep the money they have), whereas the male/female axis would probably sit in the middle since women and men are found pretty evenly spread on the political Left and Right. This major axis is the adversarial plane on which the most people can find common ground in trying to define themselves, a composite axis with elements of every major issue in defining what it means. From tip to tail our sausage runs from Hard Left to Hard Right.
How a descriptive model of Left/Right relates to what matters in running a democracy
Initially this purely descriptive model of the Left and Right appears a little unwieldy. Why do we need a descriptive approach when we can just get ourselves a nice simple ideological axis with Marx at one end and Ayn Rand at the other? But it begins to appear useful when it comes to working out what interests parties represent. While a lot of people on one side might appear to have relatively little in common, they are united by the fact that they have more in common with each other than any of them have with those at the tip of the other end of the spectrum. Each coalition is held together by how much their opponents appear to be other to all of them. So once you start identifying people as Left and Right this view becomes self-reinforcing.
Another characteristic of this view of Left and Right is that it can rapidly change when a new issue emerges. If you conducted our imaginary census in America in 2001 on September the 10th and then again on September the 12th the balance of factors that defined the Left-Right axis would have shifted considerably. The rise of environmentalism and the decline of nuclear disarmament make perfect sense in this context, nukes simply stopped polarising people and became a non-issue. Alternatively, in defining Left/Right this way we are weighting it equally for everyone in the democracy, so were we to conduct the same imaginary census on a subsidiary group (like a state or a large region) we might find that the Left/Right lines were defined by entirely different issues. This is one of the fundamental limitations of adversarial democracy: Left/Right only works for a single scale of population.
Here I'd like to point out some of the failings of this view: this only works as a thought experiment. In the 1960s and 1970s numerous attempts were made in early sociology to prove the underlying existence of Left and Right as real, descriptive, things using a similar methodology to the one I suggested here. These experiments only showed that every experiment produced a different result. This essay is not intended to be original research, in fact this view would be considered hideously out of date in modern political science. However, by presenting each issue as an axis unrelated to all the others and Left/Right as an aggregate of opinion I hope I've provided a way to visualise a functioning democracy. It's something akin to a lever, balancing every issue by breadth and depth of opinion, and where the weight of every individual's views balances against the rest's come election time. Left/Right allows society to carve itself into two equally sized blocks, and then political parties carve these ranges of opinion into smaller units they can vote for.
Our politicians are eternally trying to either force new issues onto the population to define Left/Right in new ways, or sense how the shape of opinions has changed over time and how best to now represent it. Often this is about how the personality of a new leader embodies the current dividing issues in themselves, so for instance in America in 2004 when the Urban/Rural, North/South divide was as strong as its been in decades the Democratic candidateship of John Kerry, a patrician candidate from the North East, allowed the Republicans to define Left/Right firmly on Middle America/Coastal America lines. Left/Right provides an excellent strawman for debating purposes, since everyone has a broad idea of the coalition of interests of those they are opposed to. Whenever a commentator of one side uses the term to describe their opponents (typically Liberal or Conservative in the US) it should immediately suggest to us a strawman formed of the least attractive elements of the opposing coalition.
So, Left/Right is a changing, flightly relationship that exists between and defines the relationships among everyone in society. This then comes up against the mechanics of how a democracy works and how the media, institutions and politicians want to steer the national debate. It's a central trait of western democracy that the media acts to channel these debates into very specific directions. The real Left/Right may be a reflection of the opinions of the people, but the fourth estate moulds this and is the primary way to see it both for the citizens and the politicians, leading to the enormous influence illustrated by the recent UK phone hacking scandal. This system of having a mouthpiece/forum that is in itself a vested interest in the process is far from the ideal democracy, but a better system hasn't yet evolved.
Drawing parties from our model of social diversity
Before we work out how this would impact on our theoretical perfectly responsive democracy, it's worth pointing out how our current politics fails to live up to this. In the constituency-based democracies of Britain and America votes in different regions have different weights, with those in battle-ground states/constituencies having a far greater effect on the outcome of elections. Thus our supposed Left/Right choice is already bent out of shape by being heavily weighted for the preferences of people in certain key regions. In a PR based system we see a similar problem for different reasons - if two large parties fight each other to a standstill then other minority parties can piggyback on their joint weakness. They acquire coalition power on marginal percentages of the vote, making their votes worth porportionately more than those who voted for the two battling parties. PR relies on an imaginary, composite, will of the people that doesn't weight equally for all the citizens and doesn't treat society as being made up of individuals.
Cutting our axes into unitary parties has turned out to be pretty inevitable in any democracy. We can either have the tightly defined small parties of PR or the loosely defined larger parties of FPTP, and this means blurring out other diversity. If we were only to offer people a two-party system then we would simply go for a party of the Left and a party of the Right and hope that was enough representation to keep everyone engaged. Were we to change the electoral system and release the valve on the party system a little the next step would probably be 4 parties - Hard and Soft, Left and Right. This would allow our party system to represent radicalism a little better, but would encourage ideologues the way we saw PR systems do in essay 2.
Were we to look again at our multi-dimensional plot we may get a better idea of how to define subsequent parties - remember those lumps and bulges on the surface of our sausage? These are other two dimensional planes along which smaller groups within the population co-vary. These planes cut against the planes that define the basic Left/Right, but may partially relate to it. The people identified within these trends all plot somewhere on the Left-Right axis, but they have far more in common on one of the subsequent planes. These groups would often be communities such as religious, single-issue, or ethnic minorities that have far more in common within themselves than any of them do with the rest of the spectrum. These groups would be far better defined by discrete parties than merged into the spectrum of majority interests. In a responsive party-system these groups would be able to wield discrete parties with influence proportionate to their size, wider sympathy, and strength of opinion. This way, when elections takes place they would still form part of the blocks that define the overall Left and Right coalitions, but they would do so independantly and would have to publicly negotiate for their positions.
Using this view of democracy to build a better democractic system
Viewing the traditional axis of politics as descriptive rather than ideological has some interesting implications. It explains why whatever the nature of the population is, western democracy ends up as a battle between two roughly equal-sized blocks calling themselves Left and Right. In the course of this battle claiming the Centre is often the most electorally rewarding ground between the two, allowing politicians to portray themselves as a moderate against polarisation. This technique, the triangulation approach beloved of centre-left leaders like Bill Clinton and Tony Blair, is based on stealing the most electorally-acceptable positions from your opponent and calling them centrism. And yet as a position this hypothetical Centre is merely a lack of opinion on the issues the community cares most about. It does not exist.
The most worrying element is how this deals with subsidiarity (having each issue decided at the lowest feasible level, as in states rights or regionalism), because the bigger the population a democracy serves the less responsive these blocks are to the key issues in any specific individual's life. Going back to the issue of the first essay (why is Scotland's politics drifting away from that of the UK?) it's clear that the issues that define Left/Right in Scotland are no longer the same as those defining Left/Right in the UK. However well we build our democratic mechanism, as long as the parties behave as tribes and are the same parties at all scales (local/national) then the politics of the regions are always at the mercy of the politics of the nation. The failings of a Left/Right system are at their worst in a single-pool democracy of the PR sort, but the maleable at all scales representative democracy of the FPTP sort is equally flawed. To build a better democracy we need a system that responds well to subsidiarity, and has this built into its mechanism. We need multiple parties as options as well as functioning subsidiarity. For this, the two best choices are the Single Transferrable Vote system used in Ireland, and the Additional Member System used in Italy and Germany. These systems will be addressed in the next part of this essay.