NOTE: BP is still on hiatus. These posts are an exception. These posts will simply explain MMP as explanations that BP has seen offered so far are inadequate. This post does not endorse any electoral system or represent the views of anyone else.
Thailand’s new constitutional framework is beginning to take shape. While there are still multiple opportunities for reform proposals to be amended or rejected—at the Constitutional Drafting Committee (CDC) drafting stage, by the National Reform Council (NRC), or by the National Council for Peace and Order (NCPO)— we now have a pretty good idea of the slate of reforms Thailand’s CDC is considering. This is the first of three posts about the potential effects of those proposed reforms on Thailand’s political parties and party system. This first post focuses on the effects of the proposed reforms on party system fragmentation (the number of parties), the balance between small and larger parties, and the balance of power between the two largest parties.
B. MMP replaces MMM
The CDC is proposing a Mixed-Member Proportional (MMP) electoral system, modeled after the system in Germany. Like Thailand’s previous Mixed-Member Majoritarian (MMM) electoral system MMP gives voters two votes: one for a constituency MP in a single seat electoral constituency, and one for a party list. However, rather than simply adding the party list seats to a party’s constituency seat total, as is done under MMM, the party list vote is used to determine the total number of seats a party receives. The goal of MMP is to make the number of seats each party obtains as proportional as possible to the percentage of party list votes the party receives.
The CDC has yet to flesh out the details of how Thailand’s MMP system will operate, but we have the broad outlines. The total number of seats in the House of Representatives will be a minimum of 450 and a maximum of 480 seats, at least 20 fewer seats than the previous parliament. The number of constituency seats has been dramatically reduced, from 375 in the 2011 elections to a proposed 250, with about 250,000 people per MP. The number of seats set aside for the party list increases from 125 to 200. At 44 percent of total seats this represents the largest percentage of seats set-aside for the party list since Thailand adopted a two-tier system 2001. Finally, as in 2007 the party list seats are to be divided across 8 electoral regions. Table 1 summarizes these proposed changes.
C. Unanswered Questions
So what does MMP mean for Thailand? Obviously we cannot predict future MMP electoral outcomes with certainty. For one thing, much depends on the future strength and strategies of Thailand’s two largest parties—Pheu Thai and the Democrat party. For another, the CDC has yet to report its position on several key details of how the MMP would operate in Thailand. Three deserve particular mention.
- Thresholds: We don’t yet know whether or not there will be an electoral threshold. The German system, which is held up as the model for the current proposal, requires parties to garner 5% of the party list vote or 3 seats in the constituency elections in order to be eligible for party list seats. The MMP system in New Zealand requires 5% of the vote or one seat. Higher thresholds (more restrictive electoral systems) favor larger over smaller parties. Higher thresholds prevent smaller parties from winning seats on their own. Opponents of thresholds object to the fact that supporters of those small parties are denied direct representation. Proponents, on the other hand, argue that thresholds are useful at producing more stability in the party system and preventing hyper-fractionalization, while also making it more difficult for extremist or fringe groups to win seats.
- Overhang Seats: The CDC has not said yet how the new electoral system will handle overhang seats. Overhang seats occurs when the number of constituency seats a party wins exceeds its proportional share of the party list votes. For example, if there are 100 seats in parliament and Party A won 5% of the party list vote. This would entitle them to only 5 seats, but as Party A won 6 constituency seats this means there is one overhang seat. There are three common ways of handling overhang seats. A) The parties with overhang seats are allowed to keep the overhang seats and the number of list seats allocated to other parties is decreased by a corresponding amount. In other words, the number of parliamentary seats is fixed, so one party’s overhang gain comes at the expense of other parties. This method is used in Bolivia and Lesotho. B) Parties keep all of their overhang seats and the size of parliament is simply increased to accommodate those extra seats. So if there were 2 overhang seats the size of the parliament would be increased by two seats to accommodate. New Zealand uses this approach, as did Germany prior to 2013. Both these first two options have the effect of reducing the seat share of the parties that do not have overhang seats. For this reason, Germany’s Constitutional Court recently ruled its use of option B was unconstitutional and insisted on the use of a third option. C) Parties keep their overhang seats and all other parties receive compensatory seats to keep the distribution of seats proportional to each party’s vote share. Our presumption is the reference to a parliament with between 450 and 480 seats implies the use of method B, though the drafters have yet to say as much.
- Regional Lists: It is still unclear how the regional party list system will be itablmplemented. Will parties present separate lists of candidates in each of the 8 regions, or will parties present the same list nation-wide? Will party list votes be allocated at the regional level, as is done in Germany, or will the party list votes be allocated nationally, as was the case in the 2007 Thai election? All else equal allocating the seats nationally is more advantageous to small parties than allocating seats regionally. Given that the CDC is proposing to use Open List Proportional Representation (OLPR), we think it is likely that they will use 8 separate regional lists and aggregate and allocate seats on a regional basis. We will the effects of regionalizing the party list in the next post.
Since we don’t yet know answers to these questions in the analysis that follows we explore what the effects would be of different thresholds. We ignore, for now, the possibility of overhang seats and use the national party list vote as the basis for allocating seats. In our next post we will explore the effect of overhang seats and regional party lists.
D. The Effects of MMP
To get a sense of what we might expect from the switch to MMP we examined how the outcomes from the 2007 and 2011 elections might have been different had Thailand used MMP for those contests. We keep all other aspects of the electoral system the same. For 2007 there are 400 seats elected from constituencies using the block vote, and 80 seats elected from the regional party list, and for 2011 there are 375 first-past-the-post constituency seats and 125 national party list seats.
So, how might the outcome have been different if Thailand had used MMP in 2007 and 2011? Figure 1 shows the outcomes for different MMP scenarios in 2007, while Figure 2 shows the results for 2011. We analyze the effects of MMP with no threshold, with a 1% or one seat threshold, and a 5% or 3 seat threshold. Again, all the scenarios assume party lists seats will be distributed nationally, with no overhang seats allowed.
The first thing to note is that the new system does nothing to disturb the rank ordering of the election results. Palang Prachachon (the predecessor to Pheu Thai) and the Democrats still run number one and two in 2007, regardless of the specific scenario, and Pheu Thai and the Democrats are still the top two finishers in 2011.
E. No big gains for smaller parties
Some reformers and commenters are reporting that the switch to MMP will decrease the power of the largest parties and lead to more seats for small and medium-sized parties. For example, a recent article in The Nation asserted that, “there are likely to be more medium-sized political parties and the gap between the two main parties – Pheu Thai and Democrats – and the other parties, will reduce.” Our analysis suggests that this is not the case. Even in the most permissive scenario, with no threshold, the seat share of smaller and medium sized parties barely budges. In 2007 the seat share of smaller parties moves from 17.1 percent of the seats to 18.3 percent, while in 2011 the increase is from 15.2 to 16.4.
However the number of parties winning seats does increase dramatically under MMP with no threshold. Seven parties won seats in 2007. When we introduce MMP with no threshold eighteen parties win seats, with ten of those parties winning 3 seats or less. In 2011 the number of parties winning seats increases from eleven to twenty-one, with thirteen parties winning 3 seats or less. Such a large numbers of small parties raise concerns about political stability and the methods the larger parties might use to win the support of the micro-parties. However, as we introduce thresholds the number of smaller parties winning seats and their share of total seats predictably declines. Table 2 shows what happens to the number of smaller parties and their share of seats as we increase the threshold.
F. Big Gains for the Democrats
The fact that the two largest parties retain their dominance under our MMP simulations does not mean that nothing changes. The most striking and significant change we observe is the narrowing of the gap between the two largest parties. To be blunt—MMP is really good for the Democrat Party. In 2007 MMP would have been enough to almost completely erase the 98 seat gap between the Democrats and PPP. In 2011 MMP cuts the seat gap between the Democrats and Pheu Thai by as much as 38 percent. So, while the two largest parties remain dominant under these scenarios, the distribution of legislative power between them shifts significantly. Pheu Thai is still the largest party in all of our scenarios, but barely so in 2007. In 2011 Pheu Thai fails to win a majority when there is no threshold, but hangs on to a majority if thresholds are used.
What explains why MMP is so advantageous for the Democrats? We can dismiss the simple notion that the Democrats struggle in constituency contests, but do better in the party list contests. This may have been true in 2007–when the vote share gap between the two largest parties is smaller for the party list than it is for the constituency election. But in 2011 Pheu Thai out-distanced the Democrats by an even larger margin in the party list election than it did in the constituency election (see Table 3).
Instead, the Democrats benefit from MMPs emphasis on proportionality. As the second place party it was often in the situation of failing to win a seat in the first-past-the-post constituency elections, despite winning a large number of votes. In short first-past-the-post elections resulted in disproportionality: over-representation for winning parties (PPP and PT), and under-representation for the losing Democrats. We can see this visually in the figure below drawn from the 2011 election. The each bar represents the difference between the percentage of seats a party won in constituency elections, versus its vote share. We show this difference at the national level, and then break it down for each of the eight electoral regions used in 2007. The pattern is obvious. As the largest party Pheu Thai received, on average, more seats that its seat share would strictly warrant, while the Democrats usually received fewer seats than we might expect give is vote share. Disproportionality is greatest in Regions 1-4, areas in the North and N.E. the strongholds of Pheu, and in Region 8, the South, where the Democrats are dominant. Nationally and on average, Pheu Thai’s seat share advantage due to disproportionality was around 11 percentage points. It is not surprising then, that adopting a more proportional electoral system improves the Democrat Party’s prospects vis-à-vis Pheu Thai.
NOTE: Changed an error in Table 2.