Hacker News new | ask | show | jobs
by tjl 3309 days ago
In Canada, we have a first past-the-post voting system (if you get the majority of the votes in the riding, you get the seat) and we have at least 3 parties federally (Conservatives, Liberals, and NDP). Technically, we have 4 (including the Bloc Québecois but they're only in one province). So, it's entirely possible to have a winner take all voting system and have more than two parties. Things have fluctuated over the years and we've had 5 parties when the Progressive Conservatives had a split and the Reform Party was created, but they eventually merged once they realized that they couldn't win if they split the votes.

There's been at least one attempt to try and change the voting system, but the side involved in presenting the pros of the new system really did a terrible job.

1 comments

Well, it sort of depends on the size of the "voting blocs" in which there are winner-takes-all. If the voting bloc was a single person then you have proportional representation, more or less. In the US, the voting blocs are the states, and they are usually very big, so the US is very far from PR.

How big are these voting blocs in Canada?

Anyway, it's interesting to me that even PR systems usually end up having two major parties (with a host of smaller parties around them). I guess that's because many voters cannot bother to educate themselves about the smaller parties or because there's usually one big question of the day that can be answered in a yes/no fashion...

I work in lobbying so I'm on the periphery of something very interesting: The idea of voting blocs, and being able to identify who is going to vote which way, based on discernible facts such as race, gender, culture, religion, occupation, income, etc. This is the new holy grail in politics.

I believe that with the right algorithms and data stream (i.e. Facebook) you can calculate a parties relative strength in the electorate, based on voter characteristics. Then apply layers of other data. For example, apply crime statistics to it. Ex. Murder victim has characteristics that would make it more likely than not they are liberal. Democrats -1. Murderer is caught and has characteristics that they are more than likely Conservative. Republicans -1. Result null.

Terrorist walks into a gay nightclub and kills 49 people who have characteristics that they are most likely o vote liberal. Democrats -49. Terrorist enters a military base and murderers 13 people who have characteristics that would appear most likely to vote conservative. Republicans -13. All +/- a statistical norm.

Very cold calculus... I think government officials and political parties are doing this right now, harvesting immense data to say with certainty: In a particular voting district, we can reasonably say with x turnout we will have a result of y; x1:y1, etc.

In Canada, there's a few hundred districts across the country, so each one is no more than a few hundred thousand people at most. But, I'd say most are under 50k people. Those districts each get to elect one person to parliament. If one party gets a majority of the seats in parliament, they form the government (a majority government). If they don't (but have the largest number of seats for a single party), they try and form a coalition and form a minority government. If they can't form a coalition, the next largest party gets a chance to do it and if so they form the government. Realistically, this second option doesn't really happen.

So, doesn't a district elect a congressperson? If so, the voting bloc isn't a whole state, except I guess for electing the President and senators. Canada doesn't have an elected senate and we don't have anything like an elected President. We have a Prime Minister, but they're the head of the governing party in parliament. We also have a Governor-General, who represents the Queen as our head of state, but it's mostly a ceremonial position, except when there's a change in government. Even then, they're highly constrained and it's basically ceremonial there too.