31 August 2015

Modelling the 2015 Singapore general election I

How much of a swing in the electorate would the opposition need
to dethrone the PAP in 2015?

In a very accessible article, Jeraldine Phneah writes about the advantages the People's Action Party has over the opposition in Singapore. Note that several of the advantages are institutional and arise from the PAP's position as the dominant party in Singapore politics.

In PolSci speak, Singapore belongs to a subset of democracies that are called dominant party states: these are countries whose political landscape are overwhelmingly dominated by a single ruling party, often for decades. As it turns out, a phenomenon like the PAP is not that unique to history or politics. And as it turns out, phenomena like the PAP do come and go.

Before we examine how much of a swing in the electorate the opposition will need to deny the PAP its parliamentary supermajority or even dethrone the PAP, we need to model the electorate. But how to model the election system of an "authoritarian democracy", where elections are mostly free and fair but institutionally rigged via constant gerrymandering, the colonization of the state apparatus and bureaucracy by the ruling party, the control of media, and the takeover of the grassroots by the ruling party?

Polling in a limited democracy like Singapore is notoriously difficult. According to some accounts, the PAP's vaunted grassroots system failed to detect the impending loss in Aljunied GRC in 2011. Pollsters like Blackbox even resort to asking people who they think other people might vote for, instead of asking them who they might vote for. In most places, a polling company doing this would be laughed out of the room. Here in Singapore, we understand their difficulty: very few people will tell you honestly who they'd vote for.

We can perhaps work backwards to GE2011. What was the vote swing in 2011? How many seats did that cost the ruling party? But we don't know if further swings in the same direction have the same or larger effect - and how large.

As it turns out, political scientists have in the past few years built electoral models of dominant party states. Patrick Dunleavy and Won Taek Kang are at the forefront of this field of research, and I believe their models are easily applicable to Singapore.

Yes, there is a rational voter in a dominant party system

Climate of fear? Social and political engineering to create a compliant, disengaged, and low information populace? At the end of the day, voters still go to the polls, where they cast their votes in an election that is free of violence. Even in an authoritarian democracy or a dominant party system, voters exercise choices which are very rational.

You can download Dunleavy's ‘Rethinking dominant party systems’ but I'm going to digest it for you and comment on his rather interesting paper here.

The first principle is that every political party has a ideological position. Mapping its utility (or appeal) against the left-right spectrum, we'd get a bell-shaped curve. The further one moves left or right from the party's position, the less appealing it is to the voter.

The second principle is that a dominant party, regardless of how it got to be one and precisely because of its colonization of the state apparatus and so on, is seen as especially effective by voters - in and above its ideological positioning. In other words, the bell curve of a dominant party is higher (because it gains the "effectiveness bonus") and wider (because as the only party that matters, it can afford to be a little of everything to everyone).

The immediate implication: it is very common for a voter V to rationally choose a dominant party P1 that is further away (P1-Ui1) from his ideological position over an opposition party P2 that is closer to his ideological position (Ui2-P2) because of its effectiveness advantage.

In real life, we point towards the case of Alex Au, who was one of the leaders of People Like Us. During the 1990s (that is, before Au became a left-wing political activist), Au had reasoned online that the PAP was the best bet for Singapore's LGBT movement, if only it were open to fielding a gay-friendly candidate, if not a gay candidate. It would've been such a small gesture that surely a monolithic party like the PAP could afford, without denting its overall, overwhelming popularity.

In GE2015, we note the PAP indeed making a gesture to Singapore's liberals by fielding Louis Ng, an animal rights activist. Presumably was done with the liberal voter in mind. The PAP rationally expects some percentage of Singapore's liberals to acknowledge that despite being a socially conservative party, the PAP is still the "most effective" to push any liberal agenda.

Why you should not trust anyone who suggests a united opposition party to combat the PAP

Let's move on to a simple 2-party model of a dominant party system.

Just applying basic principles, we arrive at another immediate implication: a dominant party will always wipe the opposition party in a 2-party system, thanks to its efficacy bonus. In the words of Dunleavy, "If the ideological space is ‘uncrowded’ by viable parties, then one-party dominance will be sustained by the strong logic of opposition parties adopting ‘clear water’ positional strategies." Mathematically, the dominant party P1 will always get the lion's share of the votes and seats in the legislature.

Beware, beware of the wide-eyed, angry old man in the opposition rally or function who claims he'll never vote for an opposition that is splintered and fractured, and demands that all the opposition parties join together to defeat the PAP! That wide-eyed, angry old man is advocating the defeat of the opposition. And is probably a mole.

Undercutting a dominant party, or, everyone wants to be PAP's Team B

Give a multi-party situation in a dominant party state. Theoretically now, there are several strategies available.

1. The divergent approach: Differentiate yourself ideologically so you cannot be mistaken for the ruling party.
2. The convergent approach: Adopt an ideological position that is close enough to the dominant party, eating into its support. But not exactly the same position, so the party still attracts voters that only it can command.
3. The deeply convergent approach: Whatever the ruling party is offering, you offer exactly the same thing. Wherever the ruling party stands on an issue, you stand there as well.

In the Singapore example, the Singapore Democratic Party demonstrates clearly the divergent strategy. It is as liberal as the PAP is centrist or conservative. On the policy front, it challenges most of the PAP's policies. But it is also "very niche" appeal. And in the Singapore example, that niche is too small to put it into parliament.

The Worker's Party, as seen by its 15% discount approach on the PAP's targets in its deeply unpopular Population White Paper, exemplifies the convergence approach. Of course, it offers maybe a few positions or platforms that distinguish it from the ruling party. In the Singapore example, we have seen the PAP shamelessly pick a few unique proposals from centrist parties like the WP and NSP to adopt piecemeal every few years, often without attributing or thanking the opposition. In Dunleavy's words: "P1 may shift its policy position towards the convergent rival, and easily erode part of its support."

Tan Jee Say has gathered a team of retired senior civil servants in his Singaporeans First Party, running on the platform of "restoring our nation". It wants voters to see the party as a competent team that will deliver what the PAP should be delivering but have failed to deliver because they have swayed from the truth path... I predict its eventual manifesto will reveal itself to be an old school social democrat party like the Old PAP. And like the former NSP's positioning between 2011-2015, the SFP would likely appeal to voters who remember the good old days of the 1980s and early 90s, just before the economic bubble, uncontrolled immigration, and PM Lee Hsien Loong.

The immediate implication, in the words of Dunleavy again: "Only when some opposition parties adopt ‘convergent’ or ‘deeply convergent’ positioning strategies will support for dominant parties tend to be seriously eroded. Greater crowding of the ideological space is a key stimulus to some opposition parties adopting convergent or deeply convergent strategies."

The million-dollar question for GE2015 then is: are there enough viable opposition parties this year to seriously erode the PAP's dominance, and its dominant share of the vote?

Dunleavy doesn't state, imply or suggest this next bit, but I do. Going by first principles, clearly the billion-dollar question for GE2015 is: Given the 5 years of very visible policy failures in transport, infrastructure, the labour market and other areas, will the opposition parties directly expose the myth of the PAP's "effectiveness" in an effort to wrest its efficacy bonus or halo away?

No comments: