People don't seem to understand that the alleged weakness of the first-past-the-post system is its greatest strength.
Ontario is facing a referendum on adopting a mixed member proportional representation system. Essentially you retain a geographically-based riding system, but you add extra seats assigned to each party according to popular vote. It is intended to compensate for the fact that the first-past-the-post system awards the seat to whomever wins a plurality of the vote. Smaller parties can't win a plurality in any one riding, and so are frozen out, even as they might provincially earn a sizable fraction of the total popular vote.
Sounds good on paper. All sorts of arguments pro and con can be made on the political implications of the MMP scheme.
I don't look at a body like a legislature as the political expression of the collective will of the people. I leave that to philosophers and other chronically unemployed types.
I work for a living. I get things done. And for that I need power. Power drives a system. But it has to be clean power. Filtered power.
I look at everything as systems, with inputs, outputs, feedback loops, memory, and a source of power. It's an engineering thing, but it works for me.
So what does this have to do with legislatures and voting schemes? A legislature is a system which outputs governance for a society. A machine for making laws. That machine needs power, and in a democracy, that power is derived from the popular will of the people.
The challenge, though, is to provide clean filtered power. Imagine a lamp (a machine for making light) being powered by a power source in which the voltage varies randomly and the current switches direction with wild abandon. The light would be bright, dim, off, flickering, steady -- all over the place and utterly useless.
If we tried to work a legislature off of the popular will as measured moment by moment, it would be the same thing.
So we apply filters (just as power supplies are filtered in any electrical circuit).
One filter is to stage elections in intervals measured in years.
Another filter is to convert millions of votes by individuals (as sampled only at those intervals) into a small and manageable group of super-voters. We call them our representatives (members of the provincial parliament, or MPPs, in Ontario).
More filtering is done using a first-past-the-post system. The area over which the government has sovereign control is broken into sections (ridings). Then we sample the voters preferences within those ridings, and whatever candidate earns the most votes earns the privilege of voting on behalf of those people. That is another level of filtering.
The filtering continues. Want a research budget and guaranteed time during Question Period? Then you need to belong to a party that has elected a certain number of minimum of members to Queen's Park.
The problem with MMP and other proportional systems is that they tend to make it easier for smaller parties to get into the legislature.
Why the heck would I want the Green party to get a handful of seats and the Socialists Workers Party to get a handful of seats and the Nationalist Party of Canada to get a handful of seats?
Lowering the bar to their entry is like allowing higher-order signals to pollute the power line in a circuit. Suddenly my circuit has to contend with extraneous signals affecting the output, signals that frankly are not part of the working design.
Engineers call that noise. Engineers hate noise.
The Green Party and the Socialist Workers Party and the National Party of Canada are noise. That's not my personal value judgement on their policies. That's the point. My individual judgement is irrelevant. I know that in the collective voting system we use, these parties cannot earn enough votes to win a seat.
That tells me that as far as the filtering system we have constructed, these parties are treated as noise and are filtered out.
The question of switching to MMP or something like that, therefore, is whether we believe that too much is being filtered out by our system. Are we filtering out valuable signal instead of pointless noise?
I don't think so. The legislature passes laws after intense debate and public consultation. Many of the policies of fringe parties are incorporated into the policy platforms of larger parties in yet another filtering process, the party policy convention. In a way, they are not truly filtered out.
And not all fringe parties remain fringe parties forever. The Reform Party and the Bloc Quebecois and the NDP all found a place amidst established parties.
The Green Party has never won a seat. Well, many of their policies are already reflected in the policies of other parties. Other Green policies can't find enough traction with voters. They simply can't earn enough votes as a standalone party to win a seat. That puts the onus on them to construct policies that are more appealing to a broad cross-section of voters you would find in any given riding.
An MMP scheme will give one or more seats to the Green Party based a few votes here and there, isolated and disconnected, spread everywhere in the province, but in no single place a significant enough to make even a dent in the votes cast for other parties. Isolated. Disconnected. Tiny. Noise.
But wait, Steve, the Green Party still got a bunch of votes. Doesn't that earn them something?
Sure, if they got all those votes in one or two seats, sufficient enough to win a plurality. Just like any other party. The Green Party is not special. They lose because their votes are distributed in tiny amounts across the province. As an engineer, a bump in a signal line that is infrequent and tiny in amplitude is not a signal at all, but noise. The electoral system we use challenges a party to show that with just over 100 ridings, that is, just over 100 opportunities, it is able to convince a broad cross-section of people to vote in sufficient numbers to give that party a plurality win.
One hundred kicks at the can. Just once did the Green Party manage to convince a broad group of people, a mix of office workers and doctors and housewives and househusbands and construction workers and so on and so forth to vote for them in preference to everyone else?
No. Not once. Over and over again they fail.
Now if I collected all the, say, academics together, and counted their votes, I might get a Green win. But what sort of riding is just academics? I mean, who would mow the grass and clean the windows and do the accounting and so on.
The brilliance of the riding system is that each geographical area is in effect a microcosm of Ontario as a whole. Some will be more urban than others, some more affluent, and so on, but no riding is likely to be so entirely dominated by one tiny segment of society that a single-issue party can succeed.
To succeed, a party must appeal, at least to some extent, to many different segments of society simultaneously.
That is one of the most important filters of all, and I can't imagine why I would want to undermine it with MMP or anything else.