Politics and their Outcomes

Cartoon by J.J. McCullough at filibustercartoons.com. You should be reading it!

 This post is largely about the upcoming Presidential Election, but I'm going to tie it back to local politics and urban planning. I swear!

Depending on where you've been reading about U.S. Politics (particularly on TV or newspapers), you might believe that Barack Obama stands a really good chance of losing to Mitt Romney on November 6th. After all, we've got articles such as this one give the impression that Obama's job was in deep trouble.

And of course, if you read Nate Silver's FiveThirtyEight blog regularly, you probably are scratching your head over the idea that Romney could have ever been described as a "favorite". Even at his best (the period after the first debate), Romney has never held the advantage over Obama, according to Nate Silver's numbers. To date, Silver predicts that Obama has approximately an 83% chance of winning the election; it may be premature to call his victory a "sure thing" (Silver does believe he has almost a 20% chance of losing after all!), but clearly the narrative his data is pushing differs from much of the media coverage. One therefore, must have it wrong. Figuring out how this can "happen" requires a deeper analysis.

In any election, there's really only a handful of "swing states" that really matter. One of the quirks of the Presidential Election in the United States is the idea of the Electoral College. "Winning" involves getting a majority of the 538 electoral college votes up for grabs (i.e. getting at least 270), that are roughly distributed to states based on their population. Winning a state's electoral vote comes down to getting a plurality of votes in the state. As some states are so liberal leaning that Romney has no chance (e.g. California) and some states are so conservative leaning that Obama has no chance (e.g. Texas), this means that each candidate will have a "foundation" of electoral college votes that are pretty much a sure thing.

Image from RealClearPolitics.com

As an example, Obama is pretty much guaranteed to get 154 electoral votes, and Romney is pretty much guaranteed to get 127. Other states may not be as close to 100% as you'll get, but they lean significantly enough that you can pretty much count them for a candidate. For Obama, this means he likely can count on another 47 electoral college (putting him to 201/270 votes needed) and another 64 for Romney (giving him 191/270 votes needed). The remaining states (noted in grey above) are the ones that will therefore decide the election.

From this perspective, the election certainly does look like a horse-race. If the national polls are close (and they are), then potentially either candidate could win the election. These eleven states therefore are the real "battleground" for the U.S. election.

But that's the interesting quirk about the Presidential election; it's not really about getting a plurality of the votes overall. Where you get those votes is much more important under an electoral college system. If a candidate gets a plurality of votes in New York, they get all 29 votes. You don't get a number based on how much of the vote you recieve.

(I should note that there are two states - Maine and Nebraska - that do not do an "all-or-nothing" approach to the electoral college, but the 9 notes*** between the two do not generally have an impact on an individual election.)

Imagine if the electoral college was tied (i.e. about 222 votes for each candidate) and only two states were in play: California (55 votes), and Texas (38). If all your concerned about is winning, then the only thing that matters is getting a plurality of votes in California. Texas (in this scenario) doesn't matter. You can win 100% of the vote in Texas, but if you don't get 50%+1 of the vote in California you lose.

And it's once you start drilling down into the state-to-state polling numbers that you begin to understand why Nate Silver believes Obama is the favorite to win: his polling in the swing states (particularly Ohio) is simply better than Romney's. Consider the following polling numbers that come from RealClearPolitics. Each number represents how far ahead they are (percentage wise) based on recent polls against their opponent. Surveys often give a confidence interval of about 3.5 to 5%, which means that the actual vote could be 5% in either direction:

Virginia (13): Romney +0.3
Florida (29): Romney +1.4
North Carolina (15): Romney +3.8
Total: 57 votes

Colorado (9): Obama +1
New Hampshire (4): Obama +1.8
Iowa (6): Obama +2
Ohio (18): Obama +2.5
Nevada (6): Obama +2.7
Michigan (16): Obama +3.5
Pennsylvania (20): Obama +4.6
Wisconsin (10): Obama +5.4                   
Total: 89 votes

Now consider that Obama has a lock on 201 electoral college votes, and Romney has a lock on 191. If you need 270 to win, this means that Romney needs to win more than Virginia, Florida, and North Carolina (total: 248 votes). He would also need to win in states where Obama is the favorite (if perhaps, only slightly so). Obama could lose in his three weakest states (Colorado, New Hampshire, and Iowa) and still win the election.

Incidentally, this is why Ohio is considered to be such a crucial state in this election. Romney's path to "victory" practically requires it, and without Ohio, Obama's chances of reelection drop dramatically, forcing him to "win" most of the other states he is favoured in.

But let me try and tie this back to Toronto (and to a lesser degree, Canada).

As you may have guessed (or already knew), the effect of the electoral college distorts how politicans "approach" winning an election. If you're Romney and you've got cash to spend, you'll get the most value for your money spending it in states like Ohio rather than states such as California or Texas. Certainly, there might be some long-term strategy for "flipping a state" to your side, but in the short-term of an election, you're going to want to "impress" the undecided voters in the states that you can potentially win in. This means putting resources (ads, appearances, promises) in that you normally wouldn't in the more "solid" states.

Similar things happen here in Canada, where the electoral "strength" of provinces such as Ontario and Quebec (which elect more representatives to the House of Commons than provinces such as Alberta and Nova Scotia) are often thought to be "pandered to". Like the Electoral Map above, a prospective party leader could win every seat outside of Ontario and Quebec (that's 112 seats) but you still won't get a majority without finding 30 more seats between Ontario and Quebec. It is little wonder why (from an electoral perspective) that "the west" has a difficult time competing for political interests.

But let's talk about Toronto. Municipal government in Canada doesn't typically have a "party affiliation" (at least not in the strictest sense; there are obviously left and right-wing city councillors). Even here we see the effect on the electoral system on politics (and by extensino, land-use planning). Any particular councillor is elected from a bounded geographic area (a "ward"). Hence, if you want to "win" an election, you need to appeal to the voters within your ward. One of the big complaints about the ward system is that it encourages politicians to think hyper-locally.

For example, it's often said that politicians will "pander" to the residents in their ward, and reject sound planning applications, passing them off to the Ontario Municipal Board (OMB) to take the heat rather that face a backlash from their residents. The result is that planning in Toronto can end up costing developers, residents, and the city itself both time and money fighting cases. One can imagine that if politicans in Toronto were instead elected "at large" that there would be a smaller incentive to bend to pressure from "NIMBY" residents, as losing their vote wouldn't significantly hinder their chances for reelection. It's also contributed to hindering transit in Toronto, as councillors outside of the downtown core fight for modes of transit (ie. subways) that are financially infeasible.

(Amusingly, one of the barriers to OMB reform is that it's hard to enact pressure on the Ontario government to change the legislative framework, given that planning reform must compete with schools, services, and taxes that provincial governments are more associated with.)

But of course, it's not like the "ward" system has nothing but downsides. It;s certainly great to have someone to direct inquiries to when you have a problem (in fact, it's probably one of the biggest reasons why Rob Ford - who was known for trying to deal with problems outside his ward - was elected). The solution is often therefore to find ways of addressing the shortcomings of any system. As an example, a larger emphasis on per-consultation and education about the planning process (and how the public can help) might reduce some of the reactionary revulsion that neighbourhoods often have to "change".

Politics might hinder city-building, but it need not prevent it!