Jamus discusses boundaries.
13h ·
Residents of landlocked neighborhoods in Upper Serangoon may have wondered what wild magic conspired to have led them to becoming a part of the voting district of Marine Parade GRC. Those living in Fernvale have often also asked why they aren’t part of
#SengkangGRC, after having identified with the town since the beginning.
The carving of electoral boundaries can, indeed, appear to be a dark art. The exercise is inherently political—it will, after all, be key to the voting outcome—and all the more so if voters that support one party live in a different place from those that support another. The process of drawing these bounds in a way that advantages one party over another is known as “gerrymandering”: after a former American vice president, Elbridge Gerry, who signed a bill that created a district looking like a salamander.
But must odd-shaped boundaries necessarily be nefarious? Often, the common-sense answer rings true (“if it looks like a ducks, and quacks like a duck…”). Still, the complexity of the task means that it needn’t be. After all, the math is itself daunting. A simple 4x4 grid will offer 117 ways to subdivide the 16 squares into 4 districts (with at least 4 adjacent squares), while a 9x9 grid increases the options to 700 trillion possibilities.
Okay, so with this dizzying array of boundaries, should we just accept whatever is drawn up? Thankfully, no. Just as math presents us a conundrum for boundary drawing, it also offers us a way out, with various methods for ensuring fair boundaries. One way is to use geometry to ensure that maps are simple and coherent (these have precise mathematical definitions, but they are what they sound like in English). Odd-looking, snaking districts are unlikely to fulfill these criteria.
Another way is to minimize what’s known as an “efficiency gap”—basically, to keep wasted votes (those that aren’t needed to determine the winner of a district) to as few as possible. This is also intuitively appealing, since we’d like everyone’s vote to matter. I did some basic calculations along these lines and this approach suggested that, for
#GE2020, the PAP had an electoral advantage of 17 percent, due to districting alone. That’s larger than the final tally that would swing the outcome in most SMCs and GRCs.
A third relies on theories developed to ensure fair division (think of how one might design a system to split a cake equally between two people). One way to do this is to allow both sides to take turns with proposing and accepting splits. Interestingly, Singapore used to have such a system, in the early years. In the 1950s, boundary committees included representatives from various parties. In 1963, the PAP even invited opposition parties to submit proposals for how constituencies would be delineated!
Advances in computing power essentially now allow thousands of mathematically-fair boundaries to be simulated in finite time, and used to draw actual boundaries. These can even be adapted to constraints, like balanced ethnic composition, which we want to have in Singapore.
It’s puzzling that we don’t seem to have deployed these methods for our Electoral Boundaries Review Committee (EBRC). So when Singaporeans express distrust in the shapes of our electoral districts (see:
https://mustsharenews.com/gerrymandering-poll-singapore), we shouldn’t be surprised. If we truly want to convince Singaporeans that our boundaries are not drawn in a politically-motivated way, the tools are available for the EBRC to do so. It’s low-hanging fruit for electoral reform.
More generally, we should ask what sort of political values we reflect, when we condone intentional gerrymandering. Are we a meritocracy? If we are, does it extend to the political arena? Do all parties get an equal shot, and voters all get an equal voice, in Singapore?
#makingyourvotecount