In the first part of this two-part article, I examined the underlying attitudes of the DAP and PKR leaders before coming to the negotiating table.

I also analyzed split ticket voting patterns using polling station data collected from the 2004 general election to show voters of different races chose to split their votes in favor of either DAP or PKR, depending on the nature of the contest and the race of the candidates.

In this part of the article, I want to further probe the short and long term dynamics underlying the seat allocation exercise between DAP and PKR as well as discuss some longer term issues of political identity and political space in regards to these two parties.

Let's start this section with the following question given the evidence, based on the 2004 general election results, that DAP Chinese candidates performs much better than PKR Chinese candidates, why does PKR still insist on negotiating for some Chinese majority seats to be allocated to it, as the current negotiations seem to indicate?

I can propose a few reasons for this.

Firstly, it is not inconceivable that both DAP and PKR leaders might not have analysed the 2004 general election results in the manner shown in Part 1 of this article and hence, are unaware of these voting preferences.

Secondly, leaders of political parties sometime give themselves too much credit when it comes to their reputation and ability to swing votes in their favor.

Very few opposition political leaders can go to a seat which they have not contested in previously or have not worked the ground and hope to have a good chance of unseating the BN incumbent.

Anwar Ibrahim is such a candidate. Lim Kit Siang is also such a candidate, as shown by his decision to contest in Ipoh Timur in the 2004 general elections. Lim Guan Eng may be such a candidate and this will be put to the test if he decides to contest in a parliamentary seat in Penang in the next general elections.

Beyond these three leaders, within DAP and PKR, it is hard for me to imagine any other opposition leader with the ability to swing votes in their favor in a seat which they do not have established grassroot and local ties.

The PJ Selatan example

An apt illustration would be, again, the parliamentary seat of PJ Selatan (Apologies to R Sivarasa for beating up on him). This seat was contested by Sivarasa Rasiah, one of the more senior Indian leaders in PKR, in the 2004 general election.

A well known lawyer and activist, especially among legal and NGO circles, he put up a good fight against Ong Tee Keat in the parliamentary seat of Ampang Jaya in the 1999 general elections (Ong won the seat by a much reduced majority of approximately 10 percent).

In the 2004 general election, he shifted to PJ Selatan, a seat that was vacated by Dr Syed Husin Ali who contested as a PRM candidate in 1999.

PJ Selatan, which was won by the MCA candidate by a majority of eight percent or roughly 4,000 votes in 1999, saw the majority for the MCA candidate increase to 21,000 votes or a majority of 44 percent in the 2004 general elections.

Granted, the cause of the margin of victory swelling to such a large amount in 2004 was probably due more to the overall swing back to the BN among the electorate and less so to do with the PKR candidate himself.

But still, this illustrates the limits of individual candidates being able to swing votes in his or her favor just by the force of personality or by his or her national reputation.

I have already shown in Part 1 how voters in the Bukit Gasing state seat split their votes in favor of the DAP candidate in the tune of roughly 3,300 votes.

The importance of party-over-candidate becomes even more stark when we contrast the experience and profile of the DAP Bukit Gasing state seat candidate with Sivarasa.

John Chung, the DAP candidate for Bukit Gasing, was running his first campaign, was not even 30 at the time of the elections, had fewer financial resources at his disposal and did not have a large number of volunteers working for him, certainly far fewer than someone of Sivarasa's experience and stature.

And yet, John out-polled Sivarasa by more than 3,000 votes in the Bukit Gasing polling stations.

The fact that John was from the DAP and that Sivarasa was from PKR were obviously more important to many of the Chinese voters in Bukit Gasing than their individual qualifications and experience.

Leverage of powers

While individual opposition candidates can often decrease the level of popular support for the opposition, especially if their individual profiles do not fit the profile of the majority of the voters in a seat, it is much rarer to find opposition non-incumbent candidates who can swing voters to vote for them by reputation or even grassroot work alone.

Take the case of Ronnie Liu, a DAP leader who is a tireless activist and grassroot worker, who contested in the 1999 as well as 2004 general election in the parliamentary seat of PJ Utara, a Chinese majority seat (77 percent Chinese).

The majority for the BN incumbent, Chew Mei Fun, increased from 2,500 votes in 1999 to over 13,000 votes in 2004.

Ronnie, while popular in some of the lower and lower middle class areas in PJ Utara because of his grassroot activism, was seen by the largely middle to upper middle class voters of PJ Utara as too much of a rabble rouser rather than a potential opposition parliamentarian.

Indeed, split ticket voting analysis revealed that the BN candidate at the parliamentary level received 2.5 percent more votes than the BN candidates in the two state seats in PJ Utara.

Thirdly, if PKR is to maintain its public image as a genuinely multi-racial party, it needs to find winnable seats for its non-Malay leaders to contest in.

If its non-Malay leaders are to be continually fielded in un-winnable mixed seats, then it is only natural that these leaders would not have standing to appeal to non-Malay interests both within and outside the party.

PKR can only pass the 'smell test' of being a genuinely multi racial party if it has elected representatives from all races, but specifically from the Malay and Chinese communities, the two largest communities in Malaysia.

That is one of the main reasons why PKR fought for and obtained the Chinese majority state seat of Padungan for one of its Sarawak Chinese leaders, Dominic Ng, to contest in during the 2006 Sarawak state election.

The ability of opposition political party leaders to fight for and to place candidates in certain seats is one of the few leverage of powers which the opposition parties possess.

It is one of the main instruments by which opposition party leaders can motivate existing grassroot members to work hard in their respective constituencies, in the hope of being 'rewarded' with approval to contest in either a parliamentary or state seat.

Keeping supporters happy

This is one of the important reasons as to why DAP is reluctant to give up contesting in some of the mixed seats such as Taiping, Sungai Siput, Teluk Intan or Puchong, even though it realises that it stands little chance of winning these seats without significant Malay support who comprise significant minorities in all these mixed seats.

This is not to say that the DAP has not made any concessions whatsoever. It ceded the mixed parliamentary seats of Bayan Baru, Nibong Tebal, Gopeng, Tanjong Malim, Selayang and Teluk Kemang, which it contested in in the 1999 general elections, to PKR in the 2004 general elections.

In return, PKR allowed the DAP to contest in the parliamentary seat of Bentong in 2004, a seat which it had contested in in the 1999 general election. But the lack of agreement between DAP and PKR resulted in three corner fights in the following mixed parliamentary seats: Batu Kawan, Taiping, Teluk Intan, Sungai Siput and Puchong.

Contesting in these mixed seats has symbolic as well as practical use. It enables parties like DAP and PKR to demonstrate, symbolically, its multi-ethnic character by placing Indian and Malay candidates (for DAP) and Chinese and Indian candidate (for PKR) in some of these mixed seats even though (or perhaps because) these seats are usually un-winnable seats for the opposition candidates.

For example, DAP placed Malay candidates in the mixed parliamentary seats of Bentong and Bakri and Indian candidates in the mixed seats of Cameron Highlands, Batu Kawan and Sungai Siput and PKR placed Chinese candidates in the mixed parliamentary seats of Batu Kawan, Nibong Tebal, Bayan Baru, Gopeng, Selayang, Kapar, Batu and Gelang Patah and Indian candidates in Ampang, PJ Selatan, Taiping and Teluk Intan.

It is also a way by which supporters of both parties in these constituencies are 'appeased' in that the national leaders of the respective parties are not seen as 'giving up' the right for these parties to contest in these mixed areas.

These supporters, some of whom have aspirations to be candidates themselves, will then not rebel against the national leadership of either party and also, importantly, continue to support party candidates who might actually stand a better chance of winning state seats in these areas.

In addition to the factors outlined above, the issue of determining which party and what kind of candidates are best suited to contest in different configurations of mixed seats, further complicated the seat allocation and negotiation process.

The mixed seats

What kind of mixed seats and who should contest in them?

Different analysts and politicians have used different definitions as to what constitutes a 'mixed' seat.

One definition might be seats where no one race constitutes more than a majority or 50 percent of the electorate. Twenty seven out of 165 parliamentary seats in Peninsular Malaysia fulfill this condition of being a 'mixed' seat.

Another definition might be seats where no one race constitutes more than 60 percent of the electorate. Using this definition, 66 out of 165 parliament seats in Peninsular Malaysia would be considered 'mixed' seats.

If both DAP and PKR want to lay claim to these mixed seats, there would be undoubtedly an overlap of seats which both parties want and would need to 'barter' over.

Further complicating the matter are arguments over what type of mixed seat would the DAP perform better than PKR and vice versa.

Arguably, PKR would do better than the DAP in mixed seats where the Malay electorate is between 50 percent and 60 percent of the electorate and the DAP has already conceded most of these seats to PKR as well as to PAS.

Indeed, of the 30 parliamentary seats with a Malay electorate between 50 percent and 60 percent, the DAP only contested in two of these seats Alor Gajah and Tanjong Piai.

One could even make the case that PKR would do better than the DAP in mixed seats where the Malay electorate is between 40 percent and 50 percent of the electorate.

Indeed, of the 19 such parliamentary seats, the DAP only contested in four seats in the 2004 general elections Bentong, Raub, Puchong, Segamat and Bakri all very difficult seats for any opposition party to win.

Most controversial are the parliamentary seats where the Malay electorate is between 20 percent and 40 percent, the Chinese electorate is between 40 percent and 60 percent and the Indian electorate is between 10 percent and 25 percent.

There are 17 such seats and these are seats which are wanted by both parties. It is difficult to argue that either DAP or PKR have a comparative advantage in these types of seats.

Factors such as the grassroot network of both parties in these areas, the history of voting for the opposition and DAP specifically (since PKR is a relatively new party) and the race of the candidate which either party wants to field in specific seats come into play.

Indian voters and Indian candidates

While the voting patterns of Malay and Chinese voters have been discussed in Part 1, the voting patterns of Indian voters have thus far been ignored.

This is important in discussing the type of mixed seats which both parties want to contest in.

While PKR can claim to be able to win more Malay votes and the DAP can claim to be able to win more Chinese voters, can either party claim to be able to break the BN stranglehold over the Indian voters?

Is either party willing to field Indian candidates in any of these contested seats, even if it meant that it would decrease the electoral prospects for that party in that seat? These are important questions which need further discussion.

The level of Indian support for the BN has traditionally been the highest among the three main ethnic groups in Peninsular Malaysia.

Some estimates have put this at between 70 percent to 80 percent compared to between 55 percent to 65 percent for Malay and Chinese voters, depending on the election year.

The opposition parties have found it difficult to make any electoral headway among the Indian community but these voters could make the difference in whether a DAP or PKR candidate would perform better in the third type of mixed seats.

One strategy which has proved relatively successful for opposition parties in terms of winning over some Indian support is by fielding Indian candidates in seats with a significant minority of Indian voters.

But fielding Indian opposition candidates come at a cost in that the opposition also loses a significant number of Malay and Chinese voters.

This is probably the reason why the PKR Chinese candidate in the Chinese majority parliamentary seat of Batu Kawan (56 percent Chinese) won more votes compared to the Indian DAP candidate in the same seat and why the Chinese DAP candidate in the parliamentary seat of Teluk Intan (45 percent Chinese) won more votes compared to the Indian PKR candidate in the 2004 general election.

Hence, both parties are in a quandary. Fielding Indian candidates might win them more of the Indian vote but it also loses Chinese and Malay votes.

At the same time, both parties understand the necessity of fielding Indian candidates to appease their Indian supporters and members as well as project, symbolically at least, that both parties are multi ethnic parties.

Both parties fielded five Indian candidates at the parliamentary level but DAP was more successful in that two out of its five Indian candidates were elected into office Karpal Singh in Bukit Gelugor and M. Kulasegaran in Ipoh Barat.

In fact, Kulasegaran was remarkably successful (by standards of appealing to Indian voters) by winning between 40 percent and 60 percent of the popular vote in many of the heavily Indian polling stations in the Ipoh Barat seat.

From this standpoint, the DAP seems to be able to marginally argue that it would appeal more to Indian voters compared to PKR.

Furthermore, one can argue that the recent decision by PKR not to field an Indian candidate in the Ijok by-election much to the disappointment of then PKR leader KS Nallakaruppan (who recently formed the Malaysian Indian United Party) might have damaged PKR's reputation among Indian voters.

The question of finding and fighting for seats to field Indian candidates in would feature at the back of the mind of DAP and PKR in their seat negotiations.

Suitable coalition partner

At the same time, leaders of both parties would be weighing in the electoral cost benefit analysis of fielding Indian candidates in seats which they might arguably perform better with Malay or Chinese candidates and the consideration of presenting a multi-ethnic platform and appeasing its own Indian members and supporters.

In the longer term, both DAP and PKR face challenges which require both parties to cooperate with one another but which will also inevitably create electoral tensions.

The DAP realises that the percentage of Chinese majority seats where it stands a reasonable chance to contest and win in are decreasing as the proportion of the Chinese population in Malaysia decreases.

It requires the support of a significant proportion of the Malay electorate, especially in seats where the Chinese voters only comprise between 50 percent to 60 percent of the electorate and the Malay voters make up between 20 percent to 40 percent of the electorate.

To do this, it needs to cooperate or be in a coalition with another party which can command the support of more Malay voters without alienating a significant proportion of its Chinese and Indian voter base.

PKR is a much more suitable coalition partner for the DAP compared to PAS because of PKR's much more moderate stance on issues like the NEP and the Islamic state.

At the same time, PKR needs the support of Chinese voters in seats where the Malay voters only comprise between 50 percent and 60 percent of the electorate and Chinese voters make up between 20 percent to 40 percent of the remaining electorate.

But one of the conditions for PKR to be able to appeal to a significant enough number of Chinese voters is for it to boost its 'Chinese' credentials by having some Chinese leaders who are popularly elected.

These Chinese leaders will not only present a more genuinely multi-ethnic face to PKR but they will also act to 'moderate' some of the more 'extreme' elements which may be present in PKR.

At the same time, if PKR had more Malay leaders who were popularly elected from seats with a significant percentage of Chinese voters, these Malay leaders would also act more 'moderately' within the party and in the public sphere.

The same rationale also applies to DAP leaders who are elected in more 'mixed' seats and with a significant proportion of Malay votes. It is more likely that these leaders will act more 'moderately' when it comes to issues to do with national policy and religion and so on.

But this potentially symbiotic relationship can only be predicated if the DAP concedes to allow PKR to contest in some winnable Chinese majority seats, to get the ball rolling, so to speak. And the PKR needs to correspond by conceding certain seats which the DAP has a better chance of winning compared to PKR such as Batu Kawan, Taiping and Teluk Intan.

The BN way

But these mutual concessions are not without risks and potential conflicts. If PKR does manage to win some Chinese majority seats, what is to prevent the PKR from asking for more Chinese majority seats for its candidates to contest in? Similarly, if the DAP starts to win some mixed seats, what is to stop it from asking for more mixed seats?

Unlike the BN, which wins a large percentage of the seats it contests in, most opposition parties end up on the losing side more often than not and this leaves open the door for the other opposition party to make a 'claim' on a seat by saying that their candidate would have a better chance of winning that seat.

Furthermore, there is no 'final arbitrator' between DAP and PKR, unlike the BN which has Umno and the Prime Minister. Within the BN, the Prime Minister, who is also the Umno president, plays the role of final arbitrator when it comes to seat allocation.

This role is particular important after delimitation exercises when a certain number of new parliament and state seats are created and need to be allocated among the different BN component parties.

While some of the BN parties might not be happy with their seat allocation, the strong internal discipline within the BN prevents leaders within these parties from actively and publicly campaigning against other BN component parties and candidates.

For example, Umno in Perak was extremely unhappy that it had to 'give up' the parliamentary seat of Bukit Gantang to Gerakan who had to 'give up' the parliamentary seat of Taiping to PPP.

While voter turnout in Bukit Gantang decreased significantly, this did not prevent the Gerakan candidate from winning this seat.

While there was probably some internal reluctance on the part of Umno and Gerakan grassroot workers to support the PPP candidate in Taiping, the candidate still managed to win that seat (with a small majority of roughly five percent).

Similarly, Umno members in Kuantan were unhappy that it had to 'give up' this seat to MCA in exchange for another parliamentary seat in Pahang but this did not prevent the MCA candidate from winning this seat.

There are no such disciplinary procedures between DAP and PKR. If talks do break down, the end result would be three corner fights featuring candidates from both parties in addition to the BN candidate.

While last minute interventions by PKR leaders prevented three corner fights in Bukit Gelugor and Ipoh Timur (which were eventually won by DAP) , it was not sufficient from preventing three corner fights featuring these two parties in five other parliamentary seats.

Throw in the state seats

Thus far, the discussion here has involved only parliamentary seats. Throw in state seats into the mix and the situation becomes even more complicated.

PKR and DAP are primarily concerned over parliamentary seats but state seats can be important bargaining chips in which parties which concede parliamentary seats can be compensated by the other party conceding state seats in return.

The necessity for DAP and PKR to cooperate should be apparent to leaders of both parties even though some of the rank and file members would not necessarily agree with a level of cooperation beyond preventing multi corner fights.

The Sarawak state elections showed that it was possible for PKR and DAP to cooperate and allocate seats in a manner which prevented multi corner fights.

It is likely that the number of multi corner fights will decrease in the next election compared to the 2004 general elections.

But whether the level of cooperation among the two parties will increase to a more formal level, for the mutual long term benefit of both parties is much less likely, given the many inherent tensions and challenges.

ONG KIAN MING is currently pursuing his PhD in political science at Duke University .