May 24, 2012

Prof Wadan Narsey: A Proposal for Reform of Fiji's Electoral System

May 23, 2012

[This article is addressed:

* to the YashGhai Commission
* political parties in Fiji (old and new)
* those intending to make submissions on possible reforms to the electoral system, and
* those interested voter educationin the run-up to the elections.

It would be useful for the Ghai Commission to publicise  an email and a Skype address so citizens (abroad or locally) can communicate their views to them, without being in Fiji in person.]

There is consensus that the electoral system in the 1997 Constitution must be changed and a system without any ethnic constituencies, as the Regime insists, will be an improvement.

The weaknesses of the current system are many and I won't repeat them here. Interested readers can look at my analysis of the 2001 elections.

The current system is not unfair on ethnic grounds. See this article examining the results of the 2006 elections from an ethnic perspective.

The real weakness in the results is that small parties with reasonable national, but scattered, support are not likely to win in any one constituency, and so are under-represented or totally unrepresented in parliament.

A practical weakness was that it was extremely difficult to explain the system to ordinary voters, leading to high percentages of invalid votes.

There is an annex at the bottom of this article, which the Ghai Commission might look at, on my politically neutral educating attempt in 2005, and how this 2005 practical workshop exercise could be easily replicated, for a new electoral system, to reduce invalid voting.

Both these sets of weaknesses will be reduced by a simple "proportional" system, which gives each party the same share of seats in parliament, as its share of votes nationally (allowing for Independents).  The Party does not even have to win a single seat in a local constituency.

There is also consensus that Fiji should try a "List" element which will have two very important advantages for Fiji.

First it will help to elect larger numbers of women by forcing Parties to put women candidates high on the Party List (which every voter will see before the election).

Second, parties will be able to bring capable people into parliament, without having to be elected in any particular constituency.

A third advantage of the proportional system is that electoral boundaries will become totally unimportant: there will be no incentive for parties to try to manipulate electoral boundaries(as they used to in 199, 2001 and 2006).

A fourth advantage is that the results will be extremely predictable - even if any new party comes on the scene.

There is no one perfect system: all systems have some advantages, some faults.

This article suggests recommends a system that is relatively simple to understand and operationalize. Readers are advised to go through the arithmetic examples in a group, for discussion and easier understanding.

The system in operation

This is a simple working example based on 50 seats in Parliament: 25 for local constituencies; and 25 from the List part. You can easily change the total number.

If you cannot understand the simple arithmetic, get your secondary school children to explain to you.

Each voter will get 2 ballot papers: one for the local constituency, and one for the national party of choice.

The National Ballot paper containing the list of all the political parties, is the most important vote, which will determine how many seats each party will have in total in Parliament (in addition to any Independents who may be elected at the local constituency level), adding up to 50.

Local constituencies will elect 25.

The Party List will then provide 25 to ensure that the totals are as determined by the national party vote..

Election for the local constituency

Every electoral system in the world has local parliamentarians whose primary responsibility is to serve the local constituency needs in roads, bridges, electricity, health, jetties, investment projects etc.

If they don't perform, voters don't vote for them the next time.

We can imagine the same 25 "open" constituencies that were used in the 2006 elections: no ethnicity criterion for either candidates or voters (except to be over the age of 18).

You will not get hundreds of names on any one ballot paper (as the current NCBBF/David Arms proposal for only four large constituencies would give you, leading to a total confusing mess at election and counting time).

The ballot paper will only have the names for the candidates for that constituency, in random order, with their party symbols (Independents allowed).

The voters will place numbers in order of preference of candidates: 1, 2, 3, etc  so that if the first preference candidate does not win, that vote is not wasted but goes to the second preference, etc.

Voters can use their own personal preference order,or use the ones their party gives them.  But the voter decides in the secrecy of the voting booth, not the political party (i.e. no confusing "above the line" or "below the line" nonsense).

Voters can stop at any number, without the vote being disqualified- that vote would simply not be counted further if the preference counting goes beyond that number. So advise voters to fill in all the numbers.

The winner is the candidate who manages to win 50% of the votes, either on the first count, or following the counting of preferences.

For example, the following could be the result for the 25 local constituencies:

Table 1

Party
(A)
Local seats
won
(B)
List seats entitled
(C)
Total seats
entitled
(D)
Perc. of  National party votes
Party A
10



Party B
7



Party C
6



Party D
0



Independents
2



Total
25
25
50
100%

The national result for parties

Each voter uses the National Ballot paper (which has all the political parties on it) to tick against his or her party of choice.  No numbers are required.

These National Ballot papers are counted and aggregated throughout the whole country to get each party's share of the total votes- eg. as in Column D:this is the most important column, more important than A.

Table 2


Party
(A)
Local seats
won
(B)
List seats entitled
(C)
Total seats
entitled
(D)
Perc. of  National party votes
Party A
10

19
40%
Party B
7

17
35%
Party C
6

10
20%
Party D
0

2
5%
Independents
2

(2)

Total
25
25
50
100%

You have 48 seats to distribute between the Parties, because of the 2 seats won by Independents.

Those percentage shares in Column D, multiplied by 48 seats then gives you each Party's total number of seats entitled in parliament (here in Column C).

In the example here,      Party A is entitled to (40% of 48) = 19 seats altogether.

                                    Party B is entitled to (35% of 48) = 17 seats altogether.

With a total of 50 seats in parliament, any party that can get a minimum of 2% of all the national votes (roughly 12,000 votes), will get one seat in Parliament.

In other word, Column D is the target for all political parties:  what percentage of national votes will they get?

If they appeal only to one ethnic group- they will be limited by the numbers in that ethnic group.

Any party which can appeal to all the ethnic groups, will have larger percentages in column D, hence larger numbers in parliament.  FULL STOP.

So it does not matter at all (except to the Independents) how many seats the party won in the Local Constituencies!

Because Column B (= Column C minus Column A) will then give you the number of parliamentarians coming from the "Party List" by simple subtraction: the total in parliament stays the same- as in Column C.

The Party List

Every political party will publish a "List" of their candidates in order of importance, at the closing date of nomination of candidates i.e. before the elections (there will be no surprises sprung on voters).

The List must start with their Leader, who will be Prime Minister if that Party forms Government. (Let there be no doubt about that,as there was in the 1999 elections, remember?)

The party will be expected to put names of their candidates in their order of importance.The public will be able to see clearly from the order in the List,

(a) how multi-racial the party is
(b) how much importance they give to women at the top?
(c) how regionally representative this party is? Viti Levu? Vanua Levu? urban? rural? etc.

Since Party A won 10 local seats, and is entitled to a total of 19 seats in Parliament, they will get the remaining 9 candidates from the top of their List, in order, elected into Parliament.

Since Party B won 7 local seats, and are entitled to 17 in Parliament, then will now pick 10 from their Party List.

Note that Party D did not win a single seat in the Local constituencies, but gets 2 seats from the List, because they got 5% of all the votes in the country.

They do not have the total freedom to select who they want.  They must work down the List, in order.

Table 3

Party
(A)
Local seats
won
(B)
List seats entitled
(C)
Total seats
entitled
(D)
Perc. of  National party votes
Party A
10
9
19
40%
Party B
7
10
17
35%
Party C
6
4
10
20%
Party D
0
2
2
5%
Independents
2
0
2

Total
25
25
50
100%

If any local constituency winners are already on that List (and they can be), they are simply skipped over.

Party List will encourage Gender Balance in Parliament: MDG Target

All Pacific countries, including Fiji, are totally failing the MDG target of having gender balance in Parliament.

Women generally are reluctant to go on the campaign trail for all the usual reasons, and voters often are reluctant to elect women parliamentarians, for all the usual reasons.

The List system will force political parties, to women candidates, fairly distributed at the top of the list, to make sure that there is gender balance in parliament, without women having to win in local constituencies.

The List system will also give political parties the ability to introduce good candidates into parliament, even if they do not stand, even if they are not selected, or if they are not elected from a local constituency. 

A certain Taufa Vakatale will remember this bitter experience.

If a Party puts undeserving candidates at the top of their List, they will lose voters.

Absolutely no need for Electoral Boundary manipulation

This system will have the huge advantage that it will not matter at all where the boundaries are, as all votes for any party will be counted, wherever the voters are, and no party's votes are wasted, even if their local candidate is not elected.

This will totally eliminate all the political nonsense that used to go in the Electoral Boundaries Commission where political party representatives used to try to move a boundary this way and that in order to win some local constituency.

There will be big saving in costs for the Electoral Office and the Fiji Bureau of Statistics, who traditionally had to go through these laborious boundary exercises, trying to satisfy ridiculous criteria such as ensuring reasonable proportions of ethnic groups in each constituency.  All of that can be chucked out the window.

I recollect from my own 1999 experience how illogical many boundaries were, especially in urban Suva.

There is also no need to even make constituencies around the same size, since the national votes are all aggregated, from big and small constituencies.

The boundaries can be designed primarily to make it convenient for the local elected member to serve his/her constituencies, without any confusion, although you would not make them too small or too large in number.

Once boundaries are defined, the Electoral Boundaries Commission can go into hibernation, unless there is a need to increase the number of constituencies because of population movement.

Minor arithmetic problems


If you experiment with different examples, you find small arithmetic difficulties on how to determine proportional seats where parties are entitled to fractions of a seat.   e.g.  10.8  or 7.6    or    4.9


A simple rule is to choose the highest fractions, just enough to ensure that

(a) the total seats in parliament, including the Independents, add up to 50; and

(b) the total in the List column adds up to 25.

Predicting the future?

It will be very simple to predict how many seats each Party, old or new, will winunder the system.

The easiest (and cheapest) way is to start with the 2006 Open Constituency percentage results for SDL (47%), FLP (42%), NFP (7%), NAP (4%), PANU (1%).

You can even add a new party - let us call it NPBBF (National Party for Building a Better Fiji), led by You-Know-Who.

You can make an intelligent guess how much support (in percentage points) NPBBF will draw away from each of the other parties, and reduce the other parties' percentage support accordingly.

OR, if you want to be really scientific (like the Gallup Polls etc) you can spend some money to conduct a small random sample survey of all the rural and urban constituencies, and adults over the age of 18, to find which party they would vote for. A small sample of 3000 households would be very accurate indeed. Tebbutt Poll or some university entrepreneur could do it easily, using some help from the Fiji Bureau of Statistics household survey unit.

Put the revised percentages in column D in my Table 3 above, then just work out Column Cas "percent of 50"(assume 0 Independents).

Column C will then give you the final result in Parliament, with each party's numbers  exactly in proportion to their percentage support in the country.

You don't need to worry at all about Column A and who gets elected in individual constituencies: regardless of how many are won by each party in the local constituencies, the total number in Parliament remains the same- as in Column C.

Only the number from the List will change.

Simple, isn't it?

Multi-party government?

The Ghai Commission should be advised by all political parties, not to interfere with the multi-party government provision in the 1997 Constitution (as the NCBBF very unwisely did).

The only change required would be the minimum number of seats required to eligible to be invited into Cabinet: 10% of 50 seats will give you a minimum of 5 seats.

While Fiji's political leaders failed to make use of this mechanism in 1999 (Chaudhry) and 2001 (Qarase), it was being made to work in 2006 (by Qarase), until cut short by the 2006 coup.

Planners of any new parties should remember that it is relatively easy to get the minimum of 5 seats (or 10% of national votes) which would give them the right to be invited into Cabinet.

It is much harder to get 50% of all votes through secret ballots which a party would need to form government on its own.

It would be useful for all the political parties in Fiji to first discuss amongst themselves, perhaps facilitated and moderated by independent advisers, what they would like to see in the revised electoral system, given their experiences in the past.

A broad political agreement and a consensus set of recommendations to the Ghai Commission  should also facilitate and encourage the Ghai Commission to report quickly and with some degree of confidence to the Constituent Assembly.

Annex: Suggestions to the Ghai Commission for Voter Education Campaign

Having seen the poor performance of the electoral system in 1999 and 2001, I developed in 2005, a more detailed and hands-on voter education training kit for the Fiji Elections Office.  I used these in workshops held in Suva, Lautoka and Labasa, for voters, district officers, and returning officers.

These training kits may be seen here in Englishin Fijian and in Hindi

These "hands-on" voter education kits used simple language and cartoons to explain (a) the workings of the Alternative Vote Electoral system (b) the establishment of the multi-party government and Senate after the elections, and (c) as well as all the good governance issues associated with a democratic electoral system.

Essentially: the work-shop participants individually or in groups, acted like Returning Officers, and actually counted the votes in 5 model elections, transferring preferences where needed, in order to identify the "winners" for each constituency.

This was extremely useful in understanding how that complex system worked.

Thetraining manuals and the accompanying sets of Ballot Papers that go with the exercise, may be seen at the USP Library.

It will be miles simpler with the electoral system I am proposing here.

What the Ghai Commission could do


A comparable workshop training kit for the electoral system that I am proposing here (or some variation of it which is a compromise with the NCBBF/David Arms system) would be relatively easy to devise and agree on.


All the political parties could be taken through a national workshop to ensure that they fully understand how the system will work. Their election campaigns can then focus totally on their policies for development and good governance.

The voter education materials could be printed, within three months of the Electoral System being approved by the Constituent Assembly (and this could easily be the first task achieved by the Ghai Commission, as the NCBBF committees have already discussed alternatives in 2008).

The whole of Fiji could be made familiar with such a system, in just three months of workshops and media campaigns, with the co-ordination of NGOs and "good governance" international institutions, all completed by March of 2013.

The public media campaign would just show two colour coded ballot papers that all voters would fill out (if they wish):

1. Green National Ballot paper: Tick against the party you want (only a tick).

2. Yellow Local constituency paper: Write down the numbers 1, 2, 3, 4, etc against the candidates.

With such an easy system, the elections could be easily held by June 2013.


No comments: