Survey Data on Party System Change: Limited Usefulness

If candidates were quick to adapt to the new incentives, what about voters? The most direct evidence of a change in voter attitudes would be public opinion surveys. Unfortunately there is a dearth of comparable surveys containing questions about voter attitudes towards parties, and what evidence we do have is mixed. For example, a 1998 Suan Dusit poll asking why people vote for a particular candidate did not include a candidate's party affiliation as an option for respondents.Footnote ⁷ What we are left with is a handful of surveys by different groups, each asking slightly different questions, and none before 2000. This of course limits the inferences we can draw, but with this caveat in mind it is none-the-less worth reviewing what the survey results we do have tell us. A 2000 Suan Dusit poll asked respondents on what factors they based their parliamentary vote.Footnote ⁸ 27.9% of respondents reported they based their vote on the candidate, 22.2% on the party, 32.5% on a combination of the candidate, party and party leader, and 6.5% based their vote on the identity of the party leader. All total, the percentage of voters who relied on party cues for at least some portion of their decision was 54.6% (see Table 1).Footnote ⁹ In a 2007 poll Suan Dusit asked a similar question of voters, but unfortunately changed the menu of possible responses to: candidate, party, and party and candidate.Footnote ¹⁰ The percentage of respondents basing some part of their vote on party considerations in this poll was 66.3%.Footnote ¹¹ Again, the comparability of these two polls is limited by the differences in the questions being asked, but even with this limitation the increase in the number of respondents reportedly giving some weight to party identity in their voting decision is worth noting – an increase of nearly 12 percentage points.Footnote ¹² (Of course, it is not possible to tell whether this increase translates into greater loyalty to a particular party). If we break this increase down sub-nationally the biggest change occurs among Bangkok respondents, only 45.3% of whom based their votes at least partially on party considerations in 2000 compared to 85.7% in 2007. Outside of Bangkok the change was much more modest, from 54.5% in 2000 to 58.9% in 2007.

Table 1.

Select survey results pertaining to political parties.

Additional evidence about changes in voter attitudes towards parties is even more mixed. AsiaBarometer regularly asks respondents to name important social circles or groups. In the 2003 survey, 11.1% of respondents listed political parties as important.Footnote ¹³ By 2007 this number had actually declined to 3.7%.Footnote ¹⁴ In 2000, an Asia-Europe survey reported that barely a quarter of voters expressed “a great deal” or “quite a lot” of confidence in the country's political parties.Footnote ¹⁵ Seven years later two different surveys asked a similar question but with very different responses. AsiaBarometer (2007) reported that 40.5% of respondents trusted parties “a lot” or “to a degree”.Footnote ¹⁶ By contrast in a World Values Survey in the same year only 23.3% of respondents expressed “a great deal” or “quite a lot” of confidence in Thailand's political parties.Footnote ¹⁷

These survey results are summarized in Table 1. The bottom line is that the data are inconclusive. By some measures political parties appear to have become more important, while according to others the status of parties in Thai society has not changed or even declined. Fundamentally, however, the limited comparability between surveys by different organisations, using different methodologies, and asking slightly different questions makes drawing inferences about party system change from survey data extremely difficult. Fortunately, there are other strategies we can employ to try and determine the extent to which the Thai party system is undergoing significant change.

Institutional Reform Round II: Leveraging Reform to Trace Party System Change

It is possible to leverage some features of Thailand's electoral system along with its history of constitutional reform to estimate the extent to which voters began to view their vote differently than they had in the past. Recall that the 1997 constitution replaced the block vote system with single seat constituencies – thereby eliminating both intra-party competition as well as the supply of ‘surplus’ votes that many voters were willing to ‘sell’ in the pre-reform era (see below). This change, combined with restrictions on party switching, the addition of the national party list tier, and greater power for the prime minister, generated stronger incentives for candidates and voters to place party before person.

The party which most quickly and ably responded to these new incentives was the Thai Rak Thai party led by Thaksin Shinawatra. In the wake of the 2001 election Thai Rak Thai was eventually able to control a majority of the seats in the House of Representatives – the first political party ever to do so since the transition to democracy began in the mid-1970s. With a secure majority Thaksin worked aggressively to marginalise his rivals and political opponents while centralising power within Thai Rak Thai, and more specifically, within the office of the Prime Minister (Painter Reference Painter2006; Pasuk and Baker Reference Phongpaichit and Baker2008, Reference Phongpaichit and Baker2010).

The centralisation of power around Thaksin eventually generated a backlash from certain segments of the public and, ultimately, from Thailand's conservative forces. There is no need to detail here the events surrounding the September 2006 coup and its aftermath – this has been done elsewhere (e.g. Pasuk and Baker Reference Phongpaichit and Baker2010; Suphalak Reference Kanchanakhundi2006; Ungpakorn Reference Ungpakorn2007). However, it is noteworthy that constitutional reform was immediately put forward as one of the central planks of the coup leaders' (and interim government's) reform agenda. Their stated goal was to use constitutional reform to correct some of the perceived shortcomings of the 1997 constitution and the excesses of the Thaksin era. In short, the 2007 constitution was explicitly crafted to undermine the capacity of future political parties and elected leaders to challenge Thailand's conservative forces.

As in the 1997 reforms, once again the party system became a focus of the drafters. Their motives were fairly transparent – they sought to turn back the clock and return Thailand to an earlier era of multiple weak, narrowly focused parties and large, unstable governments. They included in the 2007 constitution a series of reforms designed to undermine the importance of parties and party labels and to discourage the creation of large national parties. First, the single seat electoral constituencies were replaced with the old multi-seat, block vote system. This meant that once again candidates from the same party had to compete against each other for election. Second, the single national party list election was scrapped and replaced with eight regional party lists, with parties allowed to run in just some of the lists. The consequence of this reform was that the party list campaign no longer served as a national referendum on each party's platform and prospective prime minister, since each region voted on a separate slate of party list candidates, and potentially, a different set of parties. The North and Northeast, strongholds of Thai Rak Thai, were carved up into four different zones for the purposes of the party list. Finally, the restrictions on party switching were effectively eliminated – members of parliament were now free to seek out greener pastures in the run up to elections. A further blow against budding party loyalties was a ruling by the Constitutional Court outlawing the Thai Rak Thai Party and banning its top leadership, including Thaksin, from office for five years.

Given all of these changes the 2007 election represents an interesting quasi-natural experiment. If party attachments did not in fact emerge over the previous decade, then the reforms put in place by the coup leaders should have been sufficient to return the party system to its earlier anaemic and fragmented state. By contrast, if, as I have argued, the tools and incentives produced by the 1997 constitution, combined with the policies and resources of Thaksin, helped produce new partisan identities, then we should see evidence of these identities in the 2007 election, despite efforts to turn back the clock. The 2007 election, with its re-adoption of most of the pre-1997 rules and institutions, allows us to conduct a rough ‘before and after’ comparison.

As discussed previously, prior to 1997 rural voters, particularly those in the North and Northeast, were atomized – there was no sense of a collective identity that crossed provinces in any significant way. No party existed that could credibly claim to represent the collective interests of these voters. This lack of connection to parties and the perceived lack of concern over policy and government quality among rural voters received a large share of the blame for the state of pre-1997 Thai politics and was a major target of the 1997 reformers. They hoped to encourage voters to look beyond offers of cash or personal appeals from individual candidates and instead consider the partisan, policy, and governance implications of their vote.

In large part, then, the 1997 reforms achieved their goal. Making use of the 2007 electoral data there is good evidence voters in the North and Northeast began to think about politics differently, and vote differently than they did before 1997.Footnote ¹⁸ To start with, consider the prevalence and pattern of split returns (Table 2). Under the block vote system voters are able to cast multiple votes for multiple candidates (as many votes as there are seats in the constituency). How do they cast their votes? Historically, the fact that voters had multiple votes to cast tended to undermine the value of party label and encourage voters to split their votes among candidates from different parties. One vote was reserved to support the voter's sincere preference – usually the candidate with the strongest local ties and who was best known to individual voters – in short, a person the voters could trust to be responsive to their needs. Voters often treated their one to two additional votes as ‘surplus’ votes. How those surplus votes were distributed depended, in part, on which candidate was willing to pay the most for those votes (Hicken Reference Hicken and Schaffer2007; Sombat Reference Sombat1993).Footnote ¹⁹

Table 2.

Percentage of constituency returns split between parties.

The result was that, prior to 1997, voters often split their votes among candidates from different parties, and as a consequence, many of electoral constituencies returned MPs from more than one party.Footnote ²⁰ In fact, prior to 1997 a majority of constituencies produced split returns (Table 2). As many as 90% of the constituencies split their returns in the North and Northeast, while the South tended to vote a straight party ticket – reflecting the dominance of the Democrat Party in the region. In 1997 the block vote was replaced with single seat constituencies, but then returned with the 2007 constitution. What happened then, when the block vote system was readopted in 2007 and voters faced the same opportunity to split their votes? Despite the same opportunity most voters in 2007 cast all of their votes for the candidates from a single party, translating to single party returns in nearly two-thirds of the districts. Note particularly the big declines in the North, Northeast and Central regions (Table 2). Overall, then, evidence from split returns are consistent with the idea that voters have become more partisan, and more loyal to party label, over time.

Another indicator of the increase in partisan loyalties, particularly in the North and Northeast, is the difference between the number of votes that candidates on the same party team receive. One of the advantages of a block vote system for researchers is that we can use co-partisan vote totals as a marker of party loyalty on the part of voters. Where voters care about the party label they should cast all of their votes for the candidates from the same party, and hence, all co-partisans should receive roughly the same number of votes. The bigger the difference we observe between co-partisan vote totals, the less important is the party label, and the more important are candidate or factional considerations, ceteris paribus.

Table 3 shows the differentials for the two largest parties in the 1992, 1995, 1997 and 2007 elections. Traditionally these differentials have been large. Not surprisingly, pre-1997 Democrat supporters show the most loyalty, but even there the first place Democrat candidates received as many as eight times the votes of the second place candidates on the party team. Parties gaining votes outside of the South and Bangkok pre-1997 exhibit even larger differentials, with first place candidates garnering as many as 15 times the number of votes of the second place candidates, and 25 times the votes of the third.

Table 3.

Co-partisan vote differentials.

2007 is a marked departure from this earlier pattern. In this election first place candidates received only 50% to 70% more votes than their co-partisans – a huge change. But more striking is the loyalty of Palang Prachachon Party (PPP) supporters. PPP was the clear successor to the banned Thai Rak Thai Party.Footnote ²¹ First-placed PPP candidates received only 20%–40% more votes than their co-partisans. Indeed, PPP actually outperformed the Democrats when it came to party loyalty. In short, in 2007 more voters were voting a straight party ticket than in the pre-reform era.

If we were to examine electoral maps of Thailand we would see that in the pre-1997 era these maps resemble patchwork quilts. No single party clearly dominated a region outside of the Democrats in the South. This began to change after the 1997 reforms. Specifically, voters in much of the North and Northeast began to vote as a block for a single party, Thai Rak Thai. As a result, the elections in 2001 and especially 2005 begin to resemble a two party contest between Democrats in the South and parts of Bangkok and Central Thailand, and Thai Rak Thai in the North and Northeast.

Strikingly, the North and Northeast continue to vote together even after the fall of Thaksin. In 2007, a referendum on the new Constitution took place, which was widely viewed as a referendum on the 2006 coup. The North and Northeast were the only regions to vote to reject the charter, despite aggressive tactics by the regime to try and ensure its passage. By the time the 2007 general election rolled around, voters had ample opportunity to return to the old norm of backing politicians from multiple political parties. Yet they largely refused to do so. Thai Rak Thai had been outlawed, Thaksin and all of the top Thai Rak Thai leadership were banned from the election, and the coup leaders saw to it that the old candidate-centred electoral system was reinstalled. Despite all of this, most voters in the North and Northeast continued to cast all of their votes for the TRT successor party, PPP.

Finally, in the run up to the 2011 election the electoral rules were amended once again back to a system very similar to what was in place for the 2001 and 2005 elections.Footnote ²² (The motivation for this change appears to be a judgment by the incumbent Democrat party that it stood a better chance under the resurrected mixed-member system.) This return to a version of the 1997 system presents another opportunity to assess the development of partisanship through comparing the difference in the number of constituency votes parties receive versus the number of party list votes they receive. Since each voter has an opportunity to cast one vote for a constituency MP and one vote for a party list, the difference in the number of votes a party receives across the two tiers should in part reflect the strength of partisan attachments. If all voters cast a straight ticket vote, the number of votes across the two tiers should be equivalent. As Table 4 demonstrates, the data are consistent with Thai voters exhibiting more intra-election party loyalty over time. In 2001 the difference between the two electoral tiers was 25% for Thai Rak Thai and 14% for the Democrats (in other words, the Democrats commanded more loyalty on the part of their voters than did Thai Rak Thai). The decline in this difference in 2005 and again in 2011 is consistent with the argument that voters have become more attached to political parties. Note also that by 2011 the total number of votes across the two major parties was nearly identical for both the Democrats and Pheu Thai (the successor to Thai Rak Thai and Palang Prachachon). Pheu Thai supporters now appear to exhibit as much or more loyalty as the more traditionally party-centred Democrat supporters.Footnote ²³

Table 4.

Percentage difference between constituency and party list votes.

Implications

Taken together, the evidence I have reviewed in this paper supports the hypothesis that the 1997 reforms did bring about changes to the party system – altering how voters behave, particularly those in the North and Northeast. However, in the process the provincial North and Northeast have been transformed from an atomized vote bank to a relatively cohesive voting block. Home to 52% of the eligible voters a cohesive North/Northeast is virtually unbeatable electorally. The irony here is palpable. Some 1997 reformers blamed ignorant Thai voters, who allegedly cared little for parties and programs, for dragging down the efficiency of Thai democracy. They hoped that the constitutional reforms would induce these voters to behave differently (read: like middle class voters behave). Indeed, these voters have changed their behaviour, but they are not behaving. In the eyes of their opponents, provincial voters are no longer just an annoyance or a hindrance, they have become an out-and-out threat. This shift has played a major role in the political crises and instability that has recently plagued Thailand. This has unfortunately prompted some in Thailand to re-evaluate their views about the necessity and legitimacy of elections, political parties, and democratic government for Thailand.

As many commentators have noted, current Thai politics is increasingly defined by conflict between re-invigorated cleavages – specifically, class (wealth and middle classes versus poor) and region (Bangkok and the South versus the North and Northeast). However, these cleavages have long existed (Ockey Reference Ockey2005). Only moderately politicized, cleavage conflicts have historically remained just under the surface. What has changed in the past decade is not only that these cleavages have become more politicised (Hicken and Selway Reference Hicken and Selway2012; Selway Reference Selway2007), but that they have, for the first time, become ‘particised’. We thus witness rural voters becoming partisan in ways they have not been before.

Is the Emergence of Partisan Attachment Permanent?

While it is clear that Thai voters, particularly those in the North and Northeast, behave very differently than their counterparts from 15–20 years ago, one might legitimately wonder whether this represents a lasting change, or a more temporary shift. It is certainly possible that the increase in voter loyalty that I label as partisan is merely a reflection of the extremely polarized post-coup environment. According to this argument the 2006 coup split the country into two camps, and the electoral results in 2007 and 2010 are a product of this polarization, rather than the development of new and lasting party attachments. It is also possible that this is personal, not partisan. In other words, one could argue that voters are really tied to Thaksin Shinawatra rather than his party. If so, the apparent loyalty of North and Northeast voters to TRT/PPP/PT parties is really just an expression of support for this popular, but polarising, figure.

It is too soon to definitively judge whether what we have observed in recent elections represents a sea change, or just a temporary personalised polarisation of politics. However, several pieces of evidence suggest that we may be witnesses to a more fundamental transformation. First, Thailand has experienced polarising elections in the past (e.g. 1975, 1976 and 1992) but even in those highly polarised environments voters still typically chose to split their votes between candidates from several political parties (e.g. compare 1992 to 2007 and 2010 in Tables 2 and 3). The loyalty shown by voters to PPP in 2007 and PT in 2010 represents something different. For the first time societal divisions are being written onto partisan divisions, and thus they have a political efficacy that is novel for Thailand. (One manifestation of this is the explicit link between some social movements and specific political parties [see Sinpeng and Kouhonta Reference Aim and Kuhonta2011]).Footnote ²⁴ Indeed, the recent polarisation is arguably reinforced by the reification of social divisions into partisan divisions. Second, Thailand's conservative forces did all in their power to mitigate the Thaksin effect – banning him from politics, forcing him into exile, seizing his assets, and providing ample incentives and opportunity for Thaksin loyalists (whether voters or politicians) to jump ship. And yet, most voters in the North and Northeast continued to cast all of their votes for the TRT successor parties, PPP and PT.