top of page
  • thaidatapointscom

Post 2/3. The Thai Raksa Chart Penalty: How the Decision to Split Pheu Thai Affected Final VoteShare

Updated: Jun 7, 2019

By Joel Selway and Allen Hicken

Picture: Logo for the Thai Raksa Chart Party

In this second installment of our simulation of the true decline in support for Pheu Thai in the 2019 elections, we explore how the party’s decision to not run in all 350 constituencies affected Pheu Thai’s vote share. We do this by comparing information on electoral performance in these same provinces in the 2011 elections, performance in other constituencies within the same province (where available), and average decline rates between 2011 and 2019. We find that Pheu Thai’s vote share could have been as high as almost 27% (an additional ~4.5%) had it not made the decision to engage in pre-electoral coordination with the Thai Raksa Chart party, which was dissolved prior to the March 24 elections.

The Ill-fated Alliance with Thai Raksa Chart

In the lead up to the elections, the Pheu Thai (PT) leadership was fearful that the Electoral Commission or Constitutional Court might dissolve the party—not an unreasonable concern given the junta’s antipathy towards the party and its leader, Thaksin Shinawatra.[1] Knowing that such a ban would prevent the party and its supporters from winning any seats, PT elected to hedge its bets. It supported the formation of a set of sister parties, the most important of which was the Thai Raksa Chart party (TRC). The fact that PT and TRC were part of what political scientists call a “pre-electoral coalition” was well-known. The two parties carved up the electoral map, for the most part choosing not to compete head-to-head.[2] As part of their pre-electoral coalition, PT did not field candidates in 100 constituencies.

While the pre-electoral coalition was a sensible precaution given the political environment, it proved to be a strategic mistake. On February 8, 2019, TRC shocked the Thai political scene when it announced the nomination of Princess Ubolratana as its candidate for prime minister. The princess accepted the nomination and for a short time it looked like the ruling junta’s electoral plan had been dramatically upset. However, just a few hours later, King Vajiralongkorn announced that it was inappropriate and unconstitutional for a member of the royal family to be involved in politics. Taking the King’s lead, the ECT invalidated the candidacy and then requested that the Constitutional Court dissolve the TRC party. The Court complied, dissolving TRC on March 7, less than three weeks before the election, and banning its leaders from politics for 10 years.

The dissolution of TRC left PT fielding candidates in only 250 constituencies. The obvious question is how much better would the party have done if it had fielded a candidate in every constituency? In addition to potentially capturing more constituency seats, the contests would have added to their nationwide vote share, perhaps entitling the party to seats from the party list tier. In short, we can come to no conclusions on whether support for Pheu Thai has declined without taking these 100 seats into consideration. However, as we detailed in an earlier post, we also need to remember that Pheu Thai appears to have reserved the constituencies in which it was the strongest for its own candidates, and allowed Thai Raksa Chart to contest in constituencies where it was systematically weaker. Keeping this in mind we turn now to some simulations of what PT’s vote and seat share might have been had it contested all 350 constituency seats.

The Simulation

To estimate PT’s adjusted vote share, we need a reasonable estimate of how the party would have performed if it had run in a given constituency. The main problem, however, is that the number of constituency seats was reduced from 400 to 350. In addition, constituency boundaries were changed in numerous cases.[3]We are thus unable to simply compare performance across the same constituency in 2011 to 2019.

Given this limitation, one approach is to use information on how the party faired in other constituencies in the same province. This is straightforward for the provinces in which PT ran in at least one constituency—we can look how well PT did in the constituencies within the province in which it did run in 2019, and use that information to infer the party’s expected vote share in the constituencies where it didn’t run within the same province.

The challenge, though, is that there are 17 provinces where PT fielded no candidates. As result, we have to rely on other information about PT’s fortunes. Specifically, we compute three average vote declines: at the national level, the regional level (5 regions), and the sub-regional level (12 areas). Table 1 shows how this differs by method:

Table 1. Average Vote Decline for Pheu Thai between 2011 & 2019 elections, averaged at region and across entire nation

Across the entire country, PT’s vote share declined by an average of 18.21 percentage points across the 61 provinces in which it competedin at least one constituency (column 1). In the 22/30 constituencies that PT ran in Bangkok in 2019, the average vote share was 26.07%. This is 19.91 points less than the 45.98% the party averaged in the 2011 elections. (Since Bangkok has no sub-regions, method (3) is identical to (2) in this case.) If we use the regional average method, the simulated vote share for PT would be 1.70 points lower in the 8 constituencies in Bangkok than if we used the national average method. In contrast, in the South, where PT ran in only 10 constituencies, PT averaged 3.10% in 2019 compared to the 9.95% it achieved in the 2011 elections in the same provinces. This represents a 6.16 point decline, much lower than the 18.21 point average decline nationally. Relying on the national average, then, could overestimate PT’s decline in the South by 12.05 percentage points. This possible overestimation increases to 15.02% for the Deep South once we break down by sub-region. For provinces where PT failed to run, as Table 1 demonstrates, we get better estimates of PT’s local strength the more fine-grained and disaggregated the data. Thus, we believe that the sub-regional average is likely to be the most accurate given the heavy regional patterns in party support in Thailand.

Table 2. National Vote Tally for the Six Simulations

In total, there at least six reasonable ways to estimate the PT’s vote share. We can just use the average national (6), regional (4), or sub-regional (2) decline to estimate PT’s vote share in each of the 100 constituencies, or we can supplement each of these estimates with provincial information for provinces where PT fielded candidates in 2019 (1, 3, and 5). All six methods show that PT’s national vote share would have increased from the 7,920,630 votes, or 22.29% of the national vote it received. The most modest estimate of PT’s adjusted vote share is 25.04%, from an addition of 1,633,102 votes. The more fine-grained information we use, the higher the estimated vote share for PT. The combination of using subregional average decline data for provinces where PT failed to run, and provincial average decline data for provinces where PT ran in at least one of the constituencies provides us with an estimated PT national vote tally of 9,553,732. This is an increase of 1,633,102 votes, or 4.6 percentage points. The final tally of 26.89% of the vote is still significantly lower than the 44.3% of the constituency tier vote PT won in the 2011 elections.

At first blush, 4.6% of the vote may not seem like a large amount given a 40% increase in seats contested. But remember, TRC ran in constituencies where the average level of support for PT was lower than in the constituencies where PT fielded candidates. Perhaps the biggest indicator of this is that 40 of the 100 uncontested seats were in the South, PT’s historically weakest region, where it averaged just over 11% of the vote in 2011. In contrast, only 4/100 seats are in PT’s historically strongest region, the Northeast, where it averaged 59.12% of the vote in 2011.

In the next post we look about what these simulations imply for the distribution of seats.



[1]To be precise, Thaksin was the founder of the Thai Rak Thai Party (TRT), the predecessor of PT that was dissolved following the 2006 coup. A second predecessor, the People’s Power Party (PPP), was likewise dissolved in 2008.

[2]PT and TRC fielded competing candidates in only 22 districts.

[3]Some participants and observers accusedthe military-influenced electoral commission of intentional gerrymandering districts to put anti-military parties at a disadvantage.


bottom of page