The troubling rise of Prime Minister Boris Johnson in the United Kingdom and President Donald Trump in the United States can both be traced to the inherent problems of geographically based electoral districts.
Before Alberta’s election in April, we had a chance to fix those flaws. Instead, the work of the 2016-2017 Alberta Electoral Boundaries Commission resulted in a new electoral map that was deeply flawed, created by a flawed process and based on flawed assumptions.
Our system of geography based electoral districts has deep roots that go back centuries in the United Kingdom. At their inception, geographically delineated constituencies were not created for the purpose of democracy, but rather as an expression of the power of aristocrats.
In essence, MLAs are philosophically the inheritors of archaic systems designed for Henry III’s barons in the 1200s.
Later legislation (such as the Reform Act of 1832) may have added democratic apparatus to constituencies, but the idea of a geographic constituency was fundamentally not designed to advance the representation of voters. As a metaphor, you might add an outboard motor to a bathtub, and it’ll get you across a lake, but the tub is ill suited to that task.
The first advent of democratically elected representatives of geographic constituencies occurred in a low-data world that is entirely unlike the information-rich environment in which our current policy makers operate. This meant that while 18th Century MP Robert Walpole, regarded as the first prime minister of Britain, may have known roughly whether the average person in his district of Castle Rising was wealthy or poor, it was very rudimentary knowledge compared to the household-by-household and street-by-street data available to modern election campaigns and to map designers.
We’ve seen in recent years how a corrupt redistricting process can make use of this kind of data. North Carolina’s congressional map comes to mind. A Republican-controlled process devised and implemented a congressional map that concentrated Democratic voters in three districts and allowed a statewide minority of Republican voters to win the other 10.
It is one of the finest examples of Gerrymandering – designing a map to favour one party – and would have been impossible to accomplish in earlier decades due to the paucity of granular data about voters.
There is, however, an upside to this data-rich environment: when implemented correctly, this information could be used to mitigate the anti-democratic underpinnings of geography based representation.
Historically, Alberta has largely avoided the worst pitfalls of redistricting. We have yet to see overtly partisan gerrymandering – though some of the low-population northern ridings have sometimes borne a significant similarity to the ‘rotten boroughs’ that once infested the British Parliament.
Premier Rachel Notley’s appointment of a panel composed of representatives of both major parties to oversee this electoral boundary review fit well into the provincial tradition. But the terms of reference for the committee had the same problems that every previous redistricting had since the dawn of our electoral district system: the wrong priorities, inappropriate skill-sets, and blindness to the structural issues inherent in geography based electoral districts.
Consider the commission’s priorities. As per constitutional precedent, and as per the Supreme Court case Reference re Prov. Electoral Boundaries (Sask.), , they set out with the noble goal of ensuring “effective representation.” But while effective representation sounds high-minded, all it really means is that there should be a similar number of voters in any electoral district, and that there are no obvious impediments to them addressing concerns of any particularly remote group of voters.
There is, essentially, no real way to measure how “effective” a representative is, and we are left with the most reductive interpretation of this ruling.
The second priority for electoral boundary commissions has been more problematic: To adjust the population size of districts while keeping most boundaries as static as possible. In fact, during the public hearings of the Alberta Electoral Boundaries Commission, one question was asked of almost every member of the public who attended: “Which boundary of your home riding would you change to address population growth, while leaving the rest unchanged?”
There is little reason to think keeping district boundaries as static as possible does anything to advance democratic aims. In fact, boundaries that change little over time may only serve to entrench incumbents.
Elected officials seeking re-election already enjoy a strong advantage – approximately 10 per cent, according to some estimates. Incumbents who have that kind of head start, often guaranteed re-election regardless of the work they do, are less likely to feel the need to work hard representing their constituents. In other words, perpetuating the incumbency advantage leads to less effective representation.
In the United States, constituencies that have a high partisan lean – that is, they are “safe” for one party or another – tend to be represented by the most radical wing of that party. Because their elected representative is most vulnerable to an intra-party challenger, they are incentivized to avoid cooperation or collaboration with people from across the aisle. This leads inevitably to gridlock and bad policy.
It could be argued this dynamic also contributed to Brexit, Boris Johnson, and the ongoing breakdown of government in the United Kingdom.
So how should a boundary commission go about redrawing an electoral map?
The answer is “mathematically.”
The gap in vote efficiency is one effective measurement that needs to be considered in this redistricting process. If the voters of one party are packed into a given district, then all votes beyond the 50 per cent needed to win that district are “wasted,” as are all votes that are cast for a losing candidate.
Comparing the percentage of all votes for each party that are “wasted” provides one measurement of how advantageous a riding map is. In the last election, the efficiency gap between the NDP and the UCP was 6.25 per cent in favour of the UCP.
Now, every map will inherently have a partisan lean. It’s inescapable in part because of the way groups of voters tend to cluster together. Nowadays, left-leaning people tend to congregate in cities; rural areas tend to be more conservative. This clustering means that votes will be wasted.
Although it’s difficult to be certain, the partisan lean of Alberta’s map does not appear to be egregious by the standards of other jurisdictions. This is, however, faint praise.
To determine if a map’s partisan lean is excessive, mathematicians Alan Frieze, Wesley Pegden and Maria Chikina proposed an ingenious solution: Create thousands of randomly varied maps and measure the partisan lean of each variation. If the partisan lean of the actual map is greater than that of the majority of these random districtings, then the map should be redrawn.
Such a method can be used to find optimized solutions for a number of metrics. You can develop maps that have ridings whose makeup best represents the overall population on demographic measurements like ethnicity, age, and income. You can create a map that prioritizes having a representative sampling of left-handed people versus right-handed people. But the metric that will probably lead to the best outcomes for voters is one that elected representatives will not like: largest number of competitive districts.
Competitive districts lead to representatives that have to work to earn the votes of their constituents. In Alberta, more competitive maps would make the phenomenon of the MIA MLA less prevalent. It would also mean that strong partisan bias for an individual elected official would be a losing strategy, since they would be more worried about the general election than getting re-nominated.
It’s worth noting that in its final report, Alberta’s recent electoral boundary commission made no mention of such metrics. And those omissions make sense – the commission was made up mostly of lawyers and led by a judge, Justice Myra Bielby of the Alberta Court of Appeal. Lawyers can be counted on to ask lawyerly questions, and interpret terms of reference carefully, but with a narrow and limited perspective.
To be effective in the 21st Century, an electoral commission needs at least one person with an advanced degree in mathematics, preferably someone who has studied topology or global optimization. It would also be appropriate to include an economist, preferably one with a background in game theory. Individuals who have studied geography and sociology could also provide important perspectives. The key questions surrounding modern electoral maps have less to do with the law than they have to do with these other disciplines.
The terms of reference and appointment process of the 2016-2017 panel were in line with past practice. But these terms of reference were insufficient. Geography based electoral districts with a first-past-the-post system are a bad way to divide people for the purposes of political representation. The terms of reference could have tasked the commission with creating boundaries that mitigate the flaws inherent in the fact that we divide voters up by where they live.
Our adherence to this antiquated system – which we inherited from the United Kingdom without ever asking if it was the best for governance – is based on anti-democratic assumptions that may have been blunted over the centuries, but whose influence is still felt.
As long as we are stuck with this system, we should work to subvert its biases, and the best tools we have to do so are complex mathematical models.
We need election commissions that include the skills to do so, and have a mandate that empowers them.
Olav Rokne is a writer who lives in Edmonton. There’s nothing I can do about how WordPress makes it look as if I wrote Olav’s piece. Nevertheless, as before, I merely facilitated its appearance! DJC