Third Candidate, Second Preference, First Job
The Dilemma of the 2019 Sri Lankan Presidential Election
For the first time in 20 years, the JVP is fielding a presidential candidate. While few expect Anura Kumara Dissanayaka (AKD) to win outright, many expect him to influence the final result. Possibly, by “splitting the vote”.
How might a “third candidate” change the result?
I try to answer this question by analysing a few hypothetical scenarios.
Hypothetical Scenario I (No Third Candidate)
Let’s suppose that no third candidate ran. If the outcome in 2019 is the same as in 2015, the final result would be:

[I’ve excluded the “other” candidate votes. Hence, the total valid votes is less than that of 2019.]
“Non-Hypotethical” Note: For 2019 to be like 2015, SP would have to win a significant proportion of the “Floating Vote”. This will be quite difficult.
Hypothetical Scenario II (Third Candidate, and Diligent Second Preference Voting)
Now, let’s suppose AKD did run. And let us also suppose he won all 17% of the Floating Vote (of which we assumed that SP would win 13% and GR 4% in Scenario I). The final result would be:

The instant-runoff would come into play. AKD is eliminated, and his second preference votes are counted. If all AKD’s supporters cast their second preference votes for the candidate they would have voted for had AKD not run, then the second preferences for SP and GR would have been as follows:

The final result would be the same as Scenario I.

Hence, if all AKD’s voters use their second preference, his candidature would not change the final result of the election.
The “instant-runoff” system is designed for precisely this scenario. It tries to make sure that the result of a three-horse-race is precisely that of a two-horse race. However, it also assumes that voter “participate” in the system. I.e. they cast their second preferences. This might not be the case.
“Instant-runoff” also makes a second assumption. That voters will accurately state their second preference. What I mean by “accurate” is that the voter’s second preference is actually who they would have voted for, if their first preference didn’t exist. This assumption might also not hold.
Hypothetical Scenario III (Third Candidate, and 20% Second Preference Voting)
Second preferences have never been counted in a past election. It is unclear if all voters understand their significance, and understand how to exercise them.
Hence, if (say) only 20% of AKD’s voters use their second preference, the final result would be:

Hypothetical Scenario IV (Third Candidate, and 70% Second Preference Voting)
On the other hand, if 70% of AKD’s voters use their second preference, we get:

Hence, the more diligently voters use their second preference, the smaller the effect of the third candidate.
So what should you do?
Suppose your top preference for the “first job” is a third candidate (e.g. AKD). But you’re sure that that candidate can’t win. And you don’t want a particular “non-third candidate” (GR or SP) to win. What do you do?
- Option 1: Do you vote for the other “non-third candidate” (SP or GR)?
- Option 2: Or do you vote for your first preference third candidate (AKD), and put your second preference for the other “non-third candidate” (SP or GR)?
On the one hand, you have a strong desire to break away from binary politics and prove that a “third-way” is possible. In which case, you take Option 2.
On the other hand, you want to “be safe”. What if other voters don’t exercise their second preference? You “default” to Option 1.
Clearly a dilemma.
Happy Voting!

