Why electoral formulas matter

02.17.2006

Technorati tags:

It's always easier to call for a constitutional assembly than it is to actually call a constitutional assembly. What do I mean? Well, one of the key priorities for the new Bolivian government is to call for a new constitutional assembly to write a new magna carta for the Bolivian polity. This was one of the demands of the 2003 October protests, it was promised by Mesa, and was part of Evo's campaign platform. To be fair, it was part of everyone's campaign platform. But once you decide to have a constitutional convention, you then have to decide who gets to go. Or, rather, how we decide who gets to go.

Evo's insisting on pushing through parliament a proposal to send three delegates from each of the country's 70 legislative districts. The electoral formula (see previous post) he proposes is virtually guaranteed to give his supporters a majority (if not supermajority) of representatives. Whatever you think of Evo or MAS, the formula itself could lead to an extreme case of winner-take-all (one could win 100% of the seats with only 51% of the vote). Clearly, there are better formulas available. Or, at least, alternatives.

The key issue, of course, is whether departments w/ smaller populations will have their interests protected. If convention delegates are elected on the basis of population (which has a ring of fairness to it), then three departments (La Paz, Cochabamba, Santa Cruz) would elect nearly two-thirds of the delegates, leaving small departments marginalized. If, however, convention delegates are elected w/ an equal share by department, then large populations in La Paz, Cochabamba, and Santa Cruz would be under-represented. What to do?

Part of the dilemma, of course, is that Bolivia is a centralist state (as opposed to a federal state). Nevertheless, it currently has a Senate in which each of the nine departments has three seats (small departments are over-represented in the body). To further complicate matters, Bolivians just elected (for the first time ever) departmental prefects. This was in large measure to appease demands by several departments for greater political autonomy.

Actually, the government's also supposed to hold a referendum on departmental (or regional) autonomy. There's now a debate as to whether the referendum will be binding (Evo says no), or whether it will even take place at all (Evo suggests that should be decided at the convention). Some departments (especially Santa Cruz & Tarija) are demanding autonomy "Sí o sí" ("one way or another") — at the very least they want a binding referendum.

So far, only La Paz & Potosí back Evo's proposal. The other departments are suggesting a variety of formulas. Actually, there are currently thirteen suggested formulas (w/ more to come, I'm sure). Most are coming from departmental parliamentary caucuses, among these:

Santa Cruz: A formula to give each department five delegates (times 9 departments makes 45 delegates), plus another 125 to be divided according by population (for a total of 170).

Tarija: A formula to simply give each department 15 delegates (for a total of 135)

Oruro: A formula to give each department 10 delegates, plus another 26 for indigenous communities (for a total of 116).

Beni: A formula to give each department 10 delegates, plus another 24 for indigenous communities (for a total of 114).

Pando: A formula to simply give each department 10 delegates (for a total of 90).

Evo's already decided he doesn't want special quotas for indigenous communities. And such formulas are unlikely to work, especially since they'd get bogged down into defining which communities, how many delegates per community, etc. In terms of practicality, the issue will likely boil down to population and/or regional representation. Of all the proposals, the Santa Cruz formula seems to fairly balance population & regional representation.

Still, is there anything wrong w/ simple population-based representation? Not entirely. I rather like the first part of Evo's proposal, if only for its simplicity. Three delegates elected from each of the country's 70 legislative districts seems like a good number. A convention of 210 delegates should provide a large enough number of delegates. But only if they're elected fairly. Right now that's a big "if".

A simple solution would be to take Evo's suggested number of delegates per department, and elect the delegates using a department-wide PR formula. While this lacks the close connection between voters & delegates (districts are smaller than departments), it would come closer to ensuring fair representation. If we're using party lists. This won't work if delegates are expected to run as "independents" (no one who runs for political office is really an "independent"). And Evo's proposal clearly implies party lists.

But giving all three delegates to the party w/ a simple majority (50%+1) is grossly unfair. Why not set up a quota system to distribute seats? Say the front runner is automatically awarded a seat & the party list only wins a second seat if it wins 60% of the vote (not quite two-thirds). The second place list automatically wins a seat. If no list wins 60%, the third place list wins a seat. This seems more fair. I'd also suggest we tack on a 5% (or even slighly higher) electoral threshold to be applied on a by-district level.

Example 1: Say there are five candidate lists. Say in District N the parties win the following vote shares:

Party A — 45%
Party B — 30%
Party C — 20%
Party D — 3%
Party E — 2%

It's clear that the three largest parties are fairly evenly matched, it makes sense to give each one seat. After all, the difference between B & C is smaller than between A & B (so giving two to A and none to C seems unfair). Under Evo's proposed formula, 25% of the district's voters (those who voted for C, D, E) wouldn't be represented; under my suggestion, only 5% wouldn't be represented.

Example 2: Say there are five candidate lists. Say in District N the parties win the following vote shares:

Party A — 60%
Party B — 25%
Party C — 10%
Party D — 3%
Party E — 2%

Here A wins two seats, B wins one, but no other party wins a seat (but A & B together make up 85% of the votes, so this seems fair). Under Evo's proposed formula, 40% of the district's voters (those who voted for B, C, D, E) wouldn't be represented; under my suggestion, only 10% wouldn't be represented.

Example 3: Finally, say there are five candidate lists. Say in District N the parties win the following vote shares:

Party A — 80%
Party B — 10%
Party C — 5%
Party D — 3%
Party E — 2%

This is an unlikely scenario (results close to this happen on rare occasion in a district or two). Here A wins two seats, B wins one seat, but no other party wins a seat. I still thinking giving only two seats to A is fair. Under Evo's proposed formula, 20% of the district's voters wouldn't be representated; under my suggestion only 10% wouldn't be represented.

Evo's proposal, some may argue, is similar to how seats are currently awarded for the Senate. W/ the caveat, of course, that no list ever wins three seats. Still, two parties split all three seats (even if their combined vote share is less than 50%), which seems unfair. But the highly disproportionate Senate is counter-balanced by the more proportionate House of Deputies. That wouldn't be the case in a unicameral constitutional convention.

In the end, the rules for the constitutional convention will decide whether Evo's mainly interested in securing his power, or whether he's willing to compromise.

----
NOTE: One could asky why not simply amend the constitution (as was done in 2004). But Evo's declared that he's not willing to do this; he wants a complete new constitution that would start w/ a blank slate. But the more "blank slate" the new constitution starts from, the higher the stakes for all participants.

Posted by Miguel at 08:33 PM