May 16, 2014

Elections 2014: The Distribution of Congressional Seats in the Bolivian System


As I have been posting lately, the electoral process towards the presidential and legislative elections in Bolivia is well on its way. Not only the campaigns are in full swing but the work of the Supreme Electoral Tribunal (the entity regulating and organizing the elections) is also progressing steadily. That the work of the SET progresses efficiently and independently is important because that work is the main guarantee for the legitimacy of the results and the stability of the political process. The SET has all the burden on itself.

However, since the advent of Morales into power, the reputation and the work of the SET has been repeatedly questioned by Bolivian observers, mostly the opposition and also by myself. In recent years, there have been many allegations of partiality with the ruling party. This critique was voiced during the last electoral process when the entity was revamping the electoral register, which it turned into a digital one. This time, critics have recently voiced concern about the redistricting efforts for the oncoming elections.

While, in spite these allegations, the work of the SET has largely proven to be asserted and in good practice, there are still large doubts looming over it because of some decisions made which point to irregularities or even worst to manipulation.

One observation I have this time has to do with the application of the electoral system chosen by the country. The system is known, in political science, as the mixed-member proportional representation system or MMP. This means that the government and the entity in charge the electoral processes in the country rely on a system where people are represented in congress by politicians elected in two ways: a) the percentage of votes received by a particular party at the national level equals the same percentage of seats in the lower chamber. This assures a proportionality of representation among the parties in the lower chamber and helps to keep proportional the distribution of power within that chamber. It basically means that if a party receives 20 percent of the national vote, that party will get 20 percent of the seats in the lower chamber. In addition, and here is the second way of allocation, b) from that 20 percent allocated by the national vote, 50 per cent of the seats are occupied by what is known as direct mandates. These mandates are simply representatives who are directly elected by the population in local districts through the first-past-the-post method. The single-member representatives are called uninominals, in Bolivian jargon, and the proportional representation representatives are called plurinominals. As you can see, the system is a mixture between proportional representation and single-member district representation. Bolivia chose this system mainly because of the tendency of the single-member district method (the US method) to generate two-party systems and to benefit the largest parties. That is, not to say that the proportional representation system does not also have its biases. Bolivia has always had a multi-party system and the general consensus has been, and still is, to keep such characteristic. As the Plurinational state came along and the re-foundation of the country progressed, the electoral system was largely left the same. The MMP system is still applicable.

Now, as you can imagine, there are at least two way in which this system can be manipulated. One concerns the local districts. Not only because they have the potential to brake or make elections for candidates according to who wins them, but also because they are subject to manipulation through the drawing of boundaries. The second way is through the seat distribution method, which is the way to translate the votes into seats in congress. However, if you combine these two, manipulation can be a complicated yet lucrative endeavour for the side engaging in it. These fears have been casted over the Bolivian electoral agency before and are the current source of speculation now.

The SET has recently re-calibrated the congressional seats distribution rules. These efforts, in theory, have been following the law which mandates the adjustment of voting districts every time a new census is carried out, since these are defined according to the population growth estimates. A reapportionment process was due because the government conducted a new census in 2013.

According to the SET's publication (pdf) on the redistribution of seats, the entity, while acknowledging the principles guiding the MMP for its application, concluded that such method was inadequate to solve what it called unequal conditions in the country. The document reads:
"En esta tipología no hay lugar para el caso boliviano, porque éste incorpora el
principio de equidad, lo que lleva a establecer un nuevo modelo. Se trata de un paradigma que toma en cuenta a un país marcado por asimetrías o la inequidad, una
formación económico-social donde existen, unos espacios territoriales con poca
población y con bajo grado de desarrollo económico, otros con mejores condiciones demográfi cas y económicas." 
The inequalities the document refers to are namely that some departments have much less population as others and as such are less economically developed. Backing this argument the document presents the following seat distribution per territory in the Chamber of deputies based "only" on proportional representation.

Assuming the government used the D'Hondt method (they don't mention that detail), the graph shows indeed that when we take into account the population in each department, the departments with larger populations receive more congressional seats than the ones with less population. The emblematic example comparing La Paz and Pando highlights this problem. Indeed, the rich and most populous department of La Paz has 35 more seats than the less populous and not so rich department of Pando. Therefore, La Paz ends up being significantly over represented in the Chamber of Deputies. However, it has to be highlighted at this point that if La Paz has much more inhabitants than Pando, then by proportion, it deserves more representatives. Otherwise, the principle one man - one vote is violated. In addition, it is known that the D'Hondt method has a bias towards larger parties. That is why some countries such as the US and Germany prefer to use other methods. That is not to say that these latter methods are better. The intention is soley to mention that there are other alternatives, in any case, seen by academia as more proportional or less biased.

Furthermore, the document explains how the government of Bolivia aims at remedyin this "injustice". In effect, the new method introduces corrections to the distribution calculations taking into account the differences in department population and differences in the level of human development. The intention behind this was to set a minimum number of representatives per each department. That way departments that had less population would en up having more representation in congress. The end result of the apportionment process can be observed in the following graph.

As you can see, La Paz has less congressional seats and Pando went from having 1 to acquiring 4 more. The same thing happened with Santa Cruz and Cochabamba (the most populous departments) and the following increase in seats in the rest of the departments. Now the big questions are: is this equality? Does this remedy the biases the system might have? After all the difference in seats was significant before. While Pando had only 1 person who spoke for the region in the lower chamber, La Paz had 36. In a vote, Pando had no chance at all.

However, there is problem with the above argument and it has to do with apportionment. If we remember, the MMP system does not only seek to proportionally represent the population in general. It also has a mechanism to represent the regions. To achieve this, the Bolivian system has opted for a dual chamber approach where the regional or territorial (if you will) representation is achieved in the Senate or higher chamber and the proportional representation of the national vote is achieved in the lower chamber or Chamber of Deputies. The MMP system does not seek to correctly apportion territorial representation in the lower chamber, which is what the above table is trying to do.

The result is that Pando, as a result of the "adjustment" made by the SET might NOW be over represented. Before, let's remember, the Senate had the task of representing the department of Pando. For this task, each department had 3 senators. After the introduction of the 2009 constitution the number of senators was raised to 4 per department. That is one more per department. Proportionally speaking that should not make much difference, if at all. But now, with this new correction or adjustment (however you might want to call it), the department of Pando does not only have 4 senators representing its interests in congress but it also has 4 more deputies representing its interests. Now that does not sound proportional to me.

The basic critic is that the new adjustments forget that the proportionality in the lower chamber does not refer to the regions but to the voter support for the parties in the party system. This should represent the ideological and policy side of the equation. That means that if the MAS were to get 33 per cent of the national vote, the lower chamber should have 43 MAS deputies.

The shadow that is casted over the SET is that they rearranged the distribution of seats this way because it ultimately benefits the MAS and the government. Let's see! The three departments with the most population, La Paz, Cochabamba and Santa Cruz, which hold the largest concentration of people in their capital cities of the same names, whose citizens tend to support less the MAS, have lost each 7 (La Paz), 7 (Santa Cruz) and 4 (Cochabamba) seats in the lower chamber. Considering that the MAS tends to do much better in the rural areas, i.e. not the cities, the odds have been moved from the largest cities to departments where the rural population is more significant.

Now attentive observers of the Bolivian political process might argue that La Paz has voted for MAS in the last two elections. However, while that observation might be correct, one has to recognize that support for the MAS has been decreasing not only in La Paz but in the other two largest cities in the country. The MAS has lost its mass appeal in the urban areas as its policies and actions have all too often prompted criticism. To counterbalance this, the MAS government has been working hard on securing the rural vote with gifts, donations and help which President Morales himself has been sure to hand out in person.

This adjustment to the distribution therefore has been seen with critical eyes by the media and the opposition in Bolivia. Now, the next issue is called redistricting or gerrymandering, which will be the topic of my next post.