Three-Member Constituencies (3MCs)
A compromise between First-Past-the-Post (FPTP) and Proportional Representation (PR)
FPTP
From 1947 to 1977, Sri Lanka’s Parliament as selected through First-Past-the-Post elections.
The most important benefit of FPTP was that it was simple. Voters knew exactly who their representative was, because he or she was whoever got most votes in the constituency.
The most important drawback was that seats in parliament was not always proportion to the votes each party got.
Famously, in 1970 the UNP got 38% of the popular vote, but only 17 seats in parliament, while the SLFP won 37% of the votes (52,546 fewer votes, to be exact), but won 91 seats — a clear majority in the 151-member parliament.
PR
Starting with the 1989 general election, subsequent elections have been held under the Proportional Representation system, where seats won are proportional to votes received.
While PR might seem fairer, it has a serious problem. In 2020, the Colombo Electoral District was assigned 19 seats, which meant that every voter in Colombo had 19 representatives. There is no link between candidates on the party list, and geographical constituencies. For example, a voter in Borella (like me) cannot point to any one person (or even several people) and say, “that’s my candidate.”
MMPR
As a compromise between PR and FPTP, a Mixed-member proportional representation (MMPR) system has been proposed. This system was, indeed, used in the 2018 local government elections.
In a mixed system, part of the seats are assigned on a FPTP basis, and the rest using PR.
For example, in the 2018 local election, 60% of the seats were assigned on a FPTP basis, and the remaining 40% using PR.
MMPR doesn’t completely solve the problems of FPTP and PR. We still have PR’s problem of the Lack of Constituency Link. And like FPTP, seats and votes can still be disproportional.
More importantly, MMPR is complex, and few people understand it. The lack of this understanding is a major problem for its adoption in a democratic election.
Three-Member Constituencies (3MCs)
Both PR and FPTP have serious problems. The solution I’d like to propose is, like MMPR, a compromise between FPTP and PR. But it is much simpler than MMPR.
It is inspired by two insights:
- PR’s problem of a Lack of Constituency Link happens in only large electoral districts with a large number of seats (like Colombo). Small electoral districts don’t have this problem (like Trincomalee with only 4 seats).
- FPTP is actually an extreme form of PR, where each electoral district has only one seat.
So, what if we retain PR, but such that every electoral district is assigned exactly 3 members? In each 3MC, any party or independent group can field up to 3 candidates, and the 3 candidates who get the highest number of votes, win the seats.
If we apportioned the 225 seats in parliament into 75 Three-Member Constituencies, and assigned them to electoral districts based on the number of registered voters during the last (2020) general election, then the number of 3MCs in each electoral district would be as follows:
Larger districts like Colombo, would be split into multiple 3MCs. Smaller districts like Trincomalee would remain unchanged.
More specifically, if we split Colombo into eight, 3MCs it might look like this:
(I use Blind Partitioning, to redistrict)
Questions
Why 3?
Technically, the requirement is that the constituency should be small. So, 4 or 5, or even 6 might be reasonable. But as the number of seats increases, so does the problem of the lack of Constituency Link also increases.
Won’t this be unfair for small parties?
Yes. But there is a question of whether a party that cannot win at least a third of a small constituency deserves a seat in parliament. Personally, I think not.
Any other questions?
Appendix: Other Districts
Gampaha (8)
Kurunegala (6)
Kandy & Kalutara (5)
Galle & Ratnapura (4)
Nuwara-Eliya, Matara, Jaffna, Puttalam, Anuradhapura, Badulla & Kegalle (3)
Matale, Hambantota, Batticaloa, Digamadulla, Polonnaruwa & Moneragala (2)
Vanni & Trincomalee (1)
