 Hong Kong leads the world in much. It was the world's freest economy before 1997 and still is by several measures. Business professionals are amongst the world's most talented and China's most creative.
Not to be outdone by the private sector, however, the post-97 government has also set new standards in an unusual area: fixing elections. They managed to transform a popular vote for pro-Democrats (62% of votes cast) into a minority of seats in LegCo, the Hong Kong parliament (where pro-democrats ended up with 42% of seats). Here's how the fix happened.
1) Different seat types and voting methods
First, it was decided long before the 2004 election that adults with a vote could only vote for 50% of the seats. Specifically this meant that only 30 from 60 seats would be geographical constituencies, or GCs. Inherited from the British the significance of the GCs is not what they stand for, but what they do not stand for. The functional constituencies, or FCs, are the other half of the seats. And these are voted for by a rich and loyal minority of around 5% of special voters. These functional constituencies (FC) include banking, financial services, agriculture & fisheries etc. and are carefully pre-selected. Predictably three quarters of the candidates 'elected' in 2004 were pro-Beijing: 23 from 30 seats.
There is a special irony about the FCs. The voting method used for FCs is, mostly, FPTP (First Past the Post). Here, most votes win. It is a useful method. It's not the best method, but it is the same method used by the UK, from where it was borrowed. However, whilst this FPTP method is acceptable in the FCs, the Hong Kong government decided it was unacceptable for the GCs. Instead, the GCs have a different method. They were elected by universal suffrage using LRM, the largest remainder method.
2) Keep people confused
As the HKSAR government well knows, frankly you needed a spreadsheet and a PhD to predict or understand LRM. That makes it difficult (though not impossible) for voter coordination, where people try to vote tactically to produce a preferred result. LRM voting includes Hare quota methods (and jolly interesting discussions about whether this is preferable to Droop quotas, more of which later), remainder votes, steps, lists of candidates not individual candidates, and other counting techniques about which you don't want to know. Possibly the only easy-to-understand feature of the system designed by the Hong Kong Electoral Affairs Office is that candidates must choose numbered balls from a bag in the event of a tie.
The net result was that some voters don't have the time to enquire too closely. Some commentators don't have the energy to explain the detail. And in the run up to the 2004 election neither the government nor pollsters at Hong Kong University and other civic activists who sometimes say they want people to vote tactically, released predictions about how many over-votes would occur. No wonder.
3) Flush away majority over-votes
The Hong Kong government used LRM to eliminate some votes cast for the majority party (ie, pro-democrats). Consider Hong Kong island constituency. There were 618,000 eligible voters and 6 seats in 2004. If everybody had voted then under LRM each seat would have required a quota of 103,000 votes, arrived at by dividing 618,000 (votes) by 6 (seats).
In reality, however, close to 1-in-10 of all votes cast on September 12 were one of two types of 'over votes'. Either they went to candidates with more than their required quota. Rita Fan (pro-Beijing), for example, got 6,645 votes more than required by the quota of 59,016 for Hong Kong island. Some votes went to candidates with no hope of winning anything - Stephen Char in NT West, for example, received only 9,116 votes, nowhere near enough to do anything. He lost his deposit.
These over-votes went disproportionately against the Democrats. Over two-thirds of over-votes went to pan-Democrat causes. The pro-Beijing cause, on the other hand, only accounted for one-fifth of wasted votes.
4) Give 'cheap seats' to your friends
A related creativity in the Hong Kong system was to, in effect, give away seats on the cheap to the minority party (pro-Beijing). Cheap in this instance means fewer votes were used to win versus other candidates. This is significant. Nearly half of the GC seats (14 from 30) were won in this way - ie, they were seats where the quota was never met. Examples include C S LAU, elected in Kowloon West with 43,460 votes although the quota was 56,924. Both MA Lik and CHOY So-yuk were elected in HK Island with 74,659 votes to their list. Given the quota was 59,016 then the second seat, Ms Choy's, was in effect purchased with 15,643 votes (74,659-59,016). By contrast, the Democratic Party's Martin Lee paid the full quota price (59,016 votes) plus a further 13,756 wasted votes which were unnecessary for victory and insufficient to win a second candidate. This over-vote almost certainly cannibalised enough votes from another pan-democrat candidate, Cyd HO, to cost her seat.
For these cheap seats, the Democrats were only able to deploy 514,440 votes, or 47% of all their votes. Owing to the imbalance, however, the pro-Beijing DAB managed to use 443,494 votes, or 68% of their votes. In simple terms, therefore, over two-thirds of the pro-Beijing votes were counted in the area of cheap seats, whereas less than half of the Democrats votes were so well spent.
Very much related to diminishing votes cast for pro-democracy causes, the Hong Kong government used a very careful quota. Hong Kong's use of LRM, as has been noted, was based on a quota that divides the total votes cast by the number of seats. This is known as the Hare method. In the Hong Kong Island constituency example the quota becomes 618,000/6 or 103,000 votes per seat.
However, this is only one method of imposing a quota. Alternative methods increase the size of the denominator (the figure at the bottom of the equation) and therefore decrease the size of the quota. The Droop quota, for example, adds one to the number of seats and the Imperiali quota adds two to the number of seats. Three different methods produce three different quotas:
Hare: 618,000 votes / 6 seats = 103,000 quota
Droop: 618,000/7 = 88,285
Imperiali: 618,000/8 = 77,250
Reducing the quota has a clear effect: it makes it harder for smaller parties to win seats. The popular parties will win more seats, if not all seats, in the first step. This very obvious point was even noted by the Hong Kong government when they were evaluating proportional representation after the 1997 handover: "Among the quota systems, proportionality decreases as the quota [denominator] decreases.... The larger the denominator, the more seats will a larger party capture."
Put another way, Hong Kong's quota system is deliberately designed to trade the wishes of the majority over to the minority. It's a solution that even Stalin would have appreciated.
5) Voting for lists and not individuals foments back-room deals and discourages the adversarial approach
Finally, an inherent feature of LRM is that people vote for lists and not individuals. Aside from confusing voters who intuitively relate to voting for individuals this has another effect on behind-the-scenes political life. For the 30 geographical seats on offer in Hong Kong, a total of 35 lists emerged which contained 89 candidates. In simple terms this meant that 1-in-3 of candidates could get elected and two-thirds were pissing in the wind and they knew it.
Most candidates stood within alliances where three or more people appeared on one list. Only 15% were singletons - people who only have their name on the ticket like. Examples included Rita FAN (list# 2 for Hong Kong Island). A similar number were pairs, having two names on one list like Audrey EU and Cyd HO (list #6 for Hong Kong island).
This means that two-thirds of candidates amalgamated their efforts in a gesture of self-sacrifice. The list structure means they have a firm idea that they have zero chance of winning. In first-past-the-post there is always the hope of victory and thus the adversarial system is much stronger. Undeterred, bless, these two-thirds of candidates are donating their support to grander causes which, we must presume, they are doing from the goodness of their heart and with not a single thought for the favours they can extract from the winners at a later date and behind the scenes. And if you believe that...
The central effect of the list approach is to create alliances that glue similarly minded causes. For every candidate that chips in with the few hundred or few thousand votes that might just project the name on the top of the list comes with a price tag.
All in all, a thoroughly fixed election.
|
