Every vote counts when the system is controversial

Every vote counts.

This oft-repeated mantra is perhaps nowhere truer than in the case of Macau’s unusual voting method for the 14 directly elected seats of the Legislative Assembly.

The controversial method, which invariably favors stability of the legislature and an even distribution of seats over the unpredictability of representative democracy, is about to be put to the test again on Sunday.

Three hundred thousand voters are registered to participate in the election, with many of them blissfully unaware of how few votes it can take to elect one lawmaker in lieu of another.

The Times explains why your vote could make the difference, why winning four seats on a single list is practically impossible, and why political associations are taking a gamble by running multiple lists:

HOW THE VOTING WORKS

On Sunday, some 24 lists will compete for just 14 directly elected seats in the legislature, while an additional 12 will be nominated by the functional constituency system. The 14 directly elected seats will be decided through a voting system based on the popular D’Hondt method, but with a local twist.

The D’Hondt method, named after 19th century Belgian mathematician, Victor D’Hondt, is a form of proportional representation that seeks to allocate seats in as close accordance as possible with the actual distribution of votes. The method is used to elect the legislature in many countries, including Portugal, Spain, and Japan.

A man casts his vote during the 2013 legislative election

The method dictates that each list is assigned a quotient, calculated by dividing its total number of votes by the number of seats it has already been awarded plus one. The list with the highest quotient figure is awarded the first seat, after which its quotient is reduced and the next seat is contested. Lists that do not win a seat in each round do not have their quotient reduced.

After winning its first seat, a list’s quotient becomes exactly half the number of votes it garnered in the election. In order to win a second seat, the new quotient must exceed the remaining, unchanged quotients of all other lists.

Direct elections to the MSAR’s Legislative Assembly use a “Macau Modification” method, which decrees that the quotients for lists are not decided locally – that is, by how many seats that particular list has already been awarded – but by a global modifier. The quotient in this system is calculated by dividing the votes for each list by a fixed integer: first one, then two, and then four. A table is compiled (see Figure 1), whereby one seat is awarded to each of the 14 highest quotient numbers, irrespective of ‘round’ or list.

THE IMPOSSIBLE FOUR SEATS

The Macau Modification method favors a more even distribution of seats between election contenders than a pure D’Hondt method, strengthening less popular lists and weakening more popular ones.

Critics say that the method was deliberately designed to keep the political landscape fragmented and prevent the emergence of a cohesive bloc of directly elected legislators that could challenge the status quo.

Political commentator Eric Sautedé says the system is “unfair” because the number of seats awarded to a list does not correlate with the number of votes it attracted.

“Because of the ‘modification’, it is quasi-impossible to get more than three candidates elected on a single list,” he told the Times. “I am all in favor of party-list proportional representation, but it has to be more representative.”

The representation problem stems from the third quotient of the Macau Modification, which unusually divides the number of votes by four instead of three. This means that, after attaining a second seat in the legislature, a list must perform extraordinarily well vis-à-vis its rivals to get a third.

A fourth seat is practically unimaginable because it would require the number of votes for one list to be divided by eight and still surpass that of its competitors.

“The problem with this method is that it doesn’t respect the proportionality of the different forces in Macau,” said Sergio de Almeida Correia, a lawyer and political scientist. “Because of this, the number of democratic votes is given less importance than it should be in any democratic system […] and this undermines democracy.”

MORE SEATS MIGHT MAKE IT WORSE

The Times modeled the 2013 Legislative Assembly direct elections using both the Macau Modification method and the D’Hondt method. The two systems contained significant variance in the order in which seats were assigned to the lists, which for the final few seats can be the difference between election and non-election.

For example, the 10th seat in the 2013 election went to Melinda Chan’s Alliance for Change (MUDAR) with a quotient of 8,755. Under a pure D’Hondt system, this seat would have been awarded to Chan Meng Kam’s United Citizens Association of Macau (UCUM) with a slightly higher quotient of 8,795, while the 11th seat would go to MUDAR.

More importantly however, the two systems produce different election results using the same data.

Under the pure D’Hondt method, Chan Meng Kam’s ACUM would have secured four seats, instead of the three that his association managed under the Macau Modification method. The extra seat, belonging to Leong Veng Chai, would have reduced Pereira Coutinho’s New Hope to a single legislator.

The fact that the Macau Modification method tends to distribute seats more evenly between election contenders is not that obvious in the 2013 election modeling due to the small number of seats up for grabs.

But a hypothetical experiment, using only those lists that were awarded at least one seat in the 2013 Legislative Assembly election, shows that in a 20-seat scenario, Chan Meng Kam would have been awarded five seats under a pure D’Hondt system and four using the Macau Modification method. All nine lists shown in Figure 1 would have been awarded at least two seats under the incumbent system, while MUDAR and the New Macau Democratic Association (ANMD) would have been the only lists of the nine to win just one each under the pure D’Hondt method.

Repeated calls for more directly elected seats – which have been consistently ruled out by the government in recent years – might make the problem worse under the current system. They are likely to result in excruciatingly close election results, with the potential for a seat in the legislature to be decided by just a few votes.

This is because the voting quotients become mathematically tighter after each seat is awarded, which often means that the final ones can go to either unpopular lists seeking to elect their first legislator or popular ones seeking a second or third. It’s anyone’s game.

2,000 VOTES COULD BE DECIDING FACTOR

In 2013, the 14th and final seat in the legislature was decided by just 32 votes, or a difference of 16 in the quotient figures. The seat went to New Hope (NE), narrowly beating Angela Leong’s New Union for Macau’s Development (NUDM).

Thirty-two votes determining the election of one lawmaker over another is striking in the context of a proportional representation system that claims to have represented some 150,000 voters in 2013.

Sautedé expounds on this as a focal point of his criticism against the system currently used in Macau.

“The fight is really for the last three or four seats, given the type of voting system that we have,” he said. “This time, this could extend to the last six seats, so Au Kam San, Melinda Chan, Zheng Anting, Wong Kit Cheng, Song Pek Kei and Leong Veng Chai are the ones in the hot seats.”

Moreover, “the spread between Au Kam San and Leong Veng Chai is only 2,262 votes… meaning that six seats are attributed within a spread of about 2000 votes.”

To put that in context, more than 10 times that number (about 14.5 percent of ballots cast) went to lists that failed to elect a single lawmaker in 2013. The proliferation of list-splitting tactics means that these “wasted” votes together constituted the second most popular choice.

THE GAMBLE OF SPLITTING LISTS

Political associations in the territory have adopted list-
splitting tactics to benefit from the structural way in which legislature seats are allocated and to increase their chances of winning more than one seat. But it is a fine line to tread.

This was evident in 2013, when the New Macau Liberals, under the leadership of Jason Chao, submitted their own bid to the Legislative Assembly, separate from New Macau veterans Au Kam San and Ng Kuok Cheong.

Hoping to repeat the success of the previous election in 2009, the New Macau gamble did not pay off and cost the political association one seat in the most recent legislative term.

Chan Meng Kam’s ACUM was able to take three seats in 2013 running on a single list. His share of the vote was 18 percent which, in the context of the 14 directly elected seats, corresponds to 2.52 seats.

Nevertheless, on Sunday Chan will seek to replicate New Macau’s 2009 victory by dividing their candidates across several lists with the aim of nabbing an extra seat or two.

Each list will only aim to elect two lawmakers, thereby circumventing the exponential number of votes needed for a third and fourth.

“It is much easier to elect one lawmaker running on his or her own list, than to elect three [lawmakers] from one list,” said Correia.

However, the political maneuvering – seen by some as an exploit of the system – has been criticized for undermining democracy through the use of underhanded tactics and a lack of transparency.

In theory, the more lists there are, the fewer votes needed to take a candidate across the hypothetical threshold. But critics point out that list-splitting is essentially a gamble, and the 2013 case of New Macau shows that it can backfire when the number of lists and the distribution of votes are highly uncertain.