The Implications of MMP for Thailand: Regional Party Lists
by Allen Hicken and Bangkok Pundit
This post is co-authored and cross-posted with Bangkok Pundit.
For part one of this series, see here.
Picture: Yaowarat at night
Source: Wikipedia Commons
In the last post we considered the effects of MMP on Thailand’s party system. For ease of comparison we looked at electoral results for 2007 and 2011 under MMP assuming everything else stayed the same. Specifically, we kept the total number of available seats the same for each election, aggregated and distributed the party list votes and seats at the national level, and ignored overhang seats. Based on that analysis we found that MMP does little to boost the seat share of small and medium sized parties, and does not help close the gap between the two largest parties and the rest of the field. However, MMP is a boon to the Democrat party, allowing it to significantly narrow the gap between the Democrats and the Thaksin-backed Palang Prachachon (PPP) or Pheu Thai (PT) parties.
In this post we rerun the 2007 and 2011 election, this time adhering as closely as possible to the parameters of the draft constitution. Specifically that means adjusting the following:
Number of seats: We reduce the number of seats available from 480 and 500 seats in 2007 and 2011 respectively, to 450, as proposed under the draft charter. We also adjust the ratio of constituency to party list seats to the proposed 250:200.
Overhang seats: In the last post [XXXinsert link] we ignored the issue of overhang seats. However, as we stated then, our presumption is that the reference by the drafters to a parliament with between 450 and 480 seats implies that overhang seats will be allowed, and specifically that the size of the parliament will be adjusted upwards to accommodate parties that earn overhang seats. In this analysis we incorporate overhang seats by increasing the size of the parliament by one seat for every overhang seat. We don’t, however, reward compensatory seats to non-overhang parties to maintain overall proportionality. In other words, we assume that Thailand will use the New Zealand approach to overhang seats (Method B) rather than the current German approach (Method C).
Regional party lists: Unlike 2007, where parties presented the same list in each of the 8 electoral regions and list seats were aggregated and allocated nationally, we assume that this time around the party list tier will be truly regional. That is, parties will run different lists in each of the 8 regions, and party list votes will be aggregated and seats allocated at the regional level. The fact that the CDC has proposed using open list PR for the party list we think implies this approach to the regional party list. We call this system “regional MMP”.
Thresholds: We compare outcomes under three different threshold scenarios: (a) No threshold, (b) 1% of the party list vote or one constituency seat, and (c) 5% of the party list vote or three constituency seats (German system). We assume that thresholds will work as they do in Germany—that is, the thresholds will be determined based on national votes and seats, and this will determine whether parties are eligible for party list seats, regardless of the region. In other words, the same set of parties will be eligible for party list seats in all 8 regions.
Let’s begin by looking at the possible regional breakdown of provinces and seats in the new system. Assuming that the drafters elect to use the same 8 regions that were used in 2007 this is what the breakdown would look like. (See a map of these 8 regions here). We assume that the share of constituency seats allocated to each region will remain the same as it was in 2011 (which is roughly commensurate with each region’s population), and that each region will receive 25 party list seats. The calculations use the Sainte-Laguë method for allocating seats as it is the system used in Germany (and New Zealand). The resulting system is free of any serious malapportionment, with a Malapportionment score of .006.
B. The Effects of Regional MMP
We turn now to the effects of regional MMP on the Thai party system.
No Help to Smaller Parties
As we saw in our national-level analysis in the last post, regional MMP does little to help boost to the overall fortunes of small and medium sized parties—despite claims by some commentators to the contrary. In 2007 small to medium-sized parties claimed 17.1 percent of the seats. Under regional MMP with no thresholds the share of seats barely budges, increasing to 17.9 percent of the seats. Once we introduce the thresholds, the seat share for small and medium-sized parties falls to 16.4 percent (see Table 2). In 2011 small and medium-sized parties earned 15.2 percent of the seats. With regional MMP the percentage of seats falls to 14.8% for no threshold, 14.2% with 1%/1 seat threshold, and 11.1%, with a 5%/3 seat threshold.
The shift to regional MMP has the negative impact on the number of parties we would expect. Under national MMP there are 125 seats up for grabs, which means that even very small parties can still win a seat if no threshold is used. For example, if we use national MMP for the 2011 election, 1 party can capture a seat in parliament with significantly less than 1 percent of the party list vote. However, with the switch the regional MMP number of seats available in the party list falls to 25 in each region. This makes is much harder for small parties to win seats, even when there is no threshold. Of those 15 small parties only four win seats under regional MMP—Palang Chon, Rak Santi, Thaen Khun Phaendin, and Matubhum. (We see a similar story for 2007 as well.) Smaller parties also do more poorly as thresholds are introduced, as we would expect.
Big Gains for Democrat Party
Regional MMP, like its national counterpart, disproportionately benefits the Democrat party. While Palang Prachachon and Pheu Thai still emerge as the largest parties in the 2007 and 2011 elections respectively, the gap between the Thaksin-backed parties and the Democrats narrows dramatically (Figures 1 and 2). In 2007 the introduction of regional MMP is enough to reduce Palang Prachachon’s seat advantage over the Democrat Party to a slim 3 seats if no thresholds are used, and 6 if we use thresholds. In 2011 Pheu Thai loses its majority under the no threshold and 1 percent threshold scenarios, and just clears a majority with a 5 percent threshold. Pheu Thai retains a commanding lead over its rival in terms of the number of seats, but the gap between two parties shrinks by as much as 40 percent. Taken together it is not hard to conclude that use of regional MMP in the last two elections could have significantly altered the balance of power within parliament.
Regional Breakdown of Seats
Tables 3 and 4 show how each party would fare by region in terms of the number of seats. (We use the 1%/1 seat threshold scenario for these examples). In 2011 the Pheu Thai dominates regions 1-4 in the North and Northeast, the Democrats dominate the South (region 8) and the two parties run neck –and-neck in the remaining three regions (Table 3). Three parties that won a single seat in 2011 fail to capture any seats under regional MMP—Rak Santi, Prachathipatai May/New Democrat, and Mahachon. Note that the total number of seats in 2011 exceeds 450. This is because three parties benefit from overhang seats, increasing the size of parliament. Bhumjaithai picks up overhang seats in region 4 (1 seat), region 5 (1 seat) and region 7 (3 seats). Palang Chon receives 2 additional seats over its party list allocation in region 4 and Chart Thai Pattana gets a single overhang seat in region 7.
In 2007 Palang Prachon handily wins regions 1-4, the Democrats win regions 7 and 8 and the two parties run closely in regions 5 and 6. The beneficiaries of the overhang seats in this scenario are Pheua Phaendin (two seats in region 5), Chart Thai (three in region 2 and one in region 4), and Matchima (one seat in region 4).
The Democrat Party and the South
According to Thai Rath, Nipit Intarasambat, Deputy Leader in the Democrat Party, expressed his opposition to regional MMP, arguing that it would harm the party’s fortunes. Specifically, he claimed the party does so well in Southern constituency elections that it would not be entitled to any additional the party list seats. Thus, MMP would end up hurting the party relative to its competitors. The evidence nationally clearly contradicts Nipat’s fears—regional MMP is about the most favorable electoral system one could design for the Democrat Party. But perhaps what is true nationally is not true in the South. What does the evidence say?
At one level Nipat’s concerns are clearly misplaced. It is true that the Democrats win nearly all of the constituency seats in region 8—85 percent in 2007 and 91 percent in 2011. But the party also wins the lion’s share of the party list votes in region 8—80 percent in 2007 and 75 percent in 2011. Thus, most of the party list seats also go to the party. Figure 3 shows the constituency seat baseline for the party in 2007 and 2008 (in blue) and the total number of seats the party captures in each of our threshold scenarios above this constituency baseline. The claim that the party will be shut out of the party list seats does not hold up.
To make it clearer, using MMP, the Democrats would have won 30 constituency in 2011 and would be entitled to a top-up of 16 party list seats (no threshold), 17 party list seats (1%/1 seat threshold), or 20 party list seats (5%/3 seat threshold). Therefore, even though the Democrats dominate the South, they will still be entitled to a top-up from the party list because of how well they do on the party list vote.
However, it is the case that regional MMP allows gives parties other than the Democrats to win seats in region 8. For example, despite claiming nearly 8.5 percent of the party list vote in region 8, Pheu Thai was unable to capture a single constituency seat 2011. MMP compensates for this by awarding Pheu Thai seats from the party list to bring its total percentage of seats to around 8.5 percent. As a result, while the Democrats benefit from regional MMP in absolute terms in region 8, in relative terms they do give up some ground to other parties (Figure 4), particularly in 2011 once proportional seats are factored in.
 The issue of using a one seat threshold could lead to an increase in vote-buying for some small parties who are close to the 1% threshold, but are unlikely to reach 1%. For example, Mathabhum. won .77% of the party vote in the 2011 election, but with only a 1% threshold they would not win any party list seats under MMP. They would only have the single constituency seat they won. However, if the threshold was either 1% or one constituency seat then because they won one constituency seat, they would be entitled to two party list vote seats. Hence, winning that constituency seat becomes so important because it does not just bring one MP, but brings them three MPs in total.