How should we be represented in Philippine Congress?
One of the core features of a modern representative democracy in a unitary setup is a legislature that fairly represents the citizenry in terms of numbers. This means that the variance of the size of constituencies of representatives should be as close to zero as possible. Following the “one person, one vote” precept, political territories are supposed to be symmetric, and there is no reason to assign more legislators per person on one area than the rest of the country.
Sadly, the 1987 Constitution gave us a weird legislative districting system that does not uphold this core feature. The inclusion of the word “progressive” on top of “uniform” in Article VI, Section 5 (1) as criterion means that it is possible to have legislative districts that have different sizes, especially since the same provision empowers Congress to “fix by law” the number of representatives (over the supposed “maximum” of 250).
The wording on Section 5 (3), which requires that “each city… or each province” with a population of at least 250,000 shall have at least one representative, is rather vague, as it can be taken to mean that cities or provinces with less than a quarter of a million citizens can have a representative[1]. As for Section 5 (4), it is left to Congress to reapportion the districts within three years of every new census, following the vague standards above.
Unfair Districting
The result is a legislative district system that is not only subject to Congress’ arbitrary gerrymandering, but also not flexible enough to reflect population changes. We have, on the one hand, the first legislative district of Caloocan City containing 1.19 million people, and on the other, the lone legislative district of Batanes with roughly 17,000 people. Both are represented by one legislator each. This is not a recipe for fairness.
We cannot say legislative districts follow geographic area either. Consider that Agusan del Norte (including Butuan City for geographic purposes) has land area of 3,546.86 square kilometers and two representatives in total. Abra, with land area of 4,165.25 square kilometers, has just one representative.
This is true even within territories previously organized as unified province, like Mindoro, which was only split in 1950. Oriental Mindoro (including Calapan City) has two representatives in total compared to Occidental Mindoro, even if it is more than 38% smaller than the former.[2]
One can always invoke “historical affinity”. But history itself will tell us that it holds little portent in terms of representation, or even having a set of elected officials. Our provincial, municipal, and city territories as well as legislative districts are dynamic constructions by our Congress, imperfectly tracking population growth and with great allowance for gerrymandering.
We’ve noted the case of Mindoro post-1950 earlier. Until 1975, Las Piñas, Parañaque, Muntinlupa, Taguig, Pateros, Makati, Mandaluyong, San Juan, Malabon, Navotas, Pasig and Marikina were municipalities of Rizal — a province represented only by two representatives. The province of Davao Occidental was just created in 2013, with a single representative.
Districts versus Political Boundaries
There are more contemporary and practical problems. Consider that legislative districts do not necessarily follow established political boundaries of provinces, and towns (cities/municipalities). This confusing feature ensures the disconnect of local governance and national representation on issues.
Of course, this is not a problem for provinces or towns large enough to have one representative in the Lower House. The legislative district is also formally governed by a local executive body which can then be technically compelled to facilitate consultations. Moreover, the representative’s constituency belong to a formally governed territory which provides material and social basis for common positions on issues.
However, the structure is different for towns with large population base (which house a large segment of the Philippine population). It is possible for a town to have a single mayor yet 6 representatives to the Lower House. “Legislative districts” are then created as groupings of barangays within the town.
But these districts do not necessarily correspond to any uniformly governed structure, and with the exception of few cases, they do not have a shared history of one. Who will support mechanisms for them to come up with common stand for their representatives to bring to Congress (apart from the small, “district offices” they set up)? On what basis can the representative form positions on issues to begin with?
But perhaps the most important reason is that there is no mechanism to link the two; upon elections, they can go on their separate ways, except when the chief executive has something to lobby in Congress (e.g. approval of an economic zone in the locality). The result is national legislative largely divorced from local concerns, or worse, legislators with no real constituencies except themselves once they win.
Winner Take All: Single Member Districts and the Failure of Representation
There is another problem. The political science literature is already replete with studies on the disadvantages of the “winner-take-all” systems, a system wherein all a candidate must do to govern or represent a political territory, no matter how many votes she ends up getting in proportion to total votes, is simply to have more votes than any other candidate. For one, we end up with plurality incumbents — voted by a minority yet governing over the rest.
For smaller, rural communities, this might not be a problem. Their concerns and issues might be homogenous enough to be processed and represented by a single solon in Congress. But denser communities, and their urban issues, are far more complex to be embodied by one politician. Dividing them into artificial districts doesn’t diminish the complexity of collective problems urban citizens are facing — like traffic congestion, or access to housing.
We know that representatives do not usually campaign in their districts based on issues. We can blame this, partly, to the fact that it doesn’t make political sense to do so. A candidate just has to win the most votes, so he has to appeal to the broadest possible set of voters. It makes sense to be a polarizing figure if one has a substantial base to rely on — but then another candidate can contest that base, forcing the original candidate to reach out to other voter bases.
But let us assume that they can, and for that matter, let us assume that a candidate won because of an issue-based campaign, say championing the extension of a railway line. Suppose that the first runner up has the support of groups asking the legislation of women-friendly policies. The winning candidate can focus all her energies on the platform the won her the seat, ignoring women’s issues and all other issues in the district. Suppose that this is the case also on an adjacent legislative district, then we will have two champions for the railway and none for women’s issues.
A multimember district system would have made it possible for one candidate to champion the railway issue and another for women’s issues.[3] This is apt for metropolitan areas, like Quezon City — which anyway has 6 districts. Instead of having six representatives elected in a winner-take-all manner in their artificial districts, why not let the entire Quezon City electorate vote for six legislators at large?
This brings us to my proposal.
Synced LGU, Regions, and Districts, Multi-member Districts in Large Settlements
Instead of shoehorning the electorate to inflexible legislative district boundaries, we can stick to existing local government unit (LGU) and regional boundaries and just assign representatives proportional to their population share. My scheme allocates seats to towns, provinces, and regions in a hierarchical manner, making sure that as much as possible, there is a uniform number of legislators per a number of individuals at all levels of the hierarchy.
But before I elaborate on the proposal, let’s address the problem of the number of representatives. I do not prescribe a number here. But if we are going to choose a number, it should automatically move together with the population and not set by the Congress.
One rule we can adopt is the “cube root rule”: that the “optimal” size of the parliament should be the cube root of its population.[4] Every new census (5 years), we update the size of the Congress for the next election.
Our population as of 2015 census is 100,979,303. The floor of the cube root of that number is 465, which gives us the “optimal” size of our Congress for the 2019 elections. We take floor of the quotient of the population and Congress size and get 217,159 — the number of people represented by each representative. This is close to the size of Congress post-World War II (see Chart below). Thus, 1 legislator will be assigned for each 217,159 individuals. I emphasize the “floor” rule here: a settlement with 651,476 (217,159 times 3, minus 1) citizens is still entitled to 2 legislators, not 3.
So let’s start. At the national level, we have to ensure that the number of legislators is 465, corresponding to a population of 100,979,303. We then look at the regional level. Calarzon has 14.414 million, which means it deserves a total of 66 seats in the parliament. NCR has 12.877 million, earning it 59 seats. And so on. Observe that if we take the sum of the allocated seats of regions, this will just add up to 455–10 votes short of the total 465. This means that 10 legislators will be voted at large at the national level. I’ll discuss later how this will be done.
We proceed the same way for provinces. Zooming in on provinces of Calabarzon, we observe that Cavite has 3.678 million, and so deserves 16 seats. Laguna, with its 3.035 million, deserves 13. Another 13 for Rizal’s 2.884 millon. 12 for Batangas’ 2.694 million, and 9 for Quezon’s 2.122 million. All in all, we have 63 seats. But as I mentioned earlier, Calabarzon’s 14.414 million deserves 66 seats — which means Calabarzon citizens will be voting an extra 3 representatives at large.
Finally, we have towns — cities and municipalities. All towns exceeding 217,159 individuals (usually highly urbanized cities and independent cities) are entitled to become legislative districts, with one legislator for every 217,159 people. Consider towns of Cavite province. Given 659,019 citizens, Dasmariñas directly elects 3 parliament seats. Another 2 will be allocated to Bacoor and 1 each for Imus, General Trias, Silang, and Tanza. The rest of the towns don’t make it to 217,159 individuals, so they won’t get seats. We have 9 in total for these towns, and yet Cavite deserves 16 seats. Thus, the 7 seats will be elected at large within the Cavite province.
So how will the City of Dasmariñas’ 3 seats be allocated within the town? Do we split the city into three districts? No need for that. As I said earlier, the logic of large settlements dictates the need for multi-member districts. These representatives will be voted at large within the settlement, with winners selected by plurality.
Note that some towns will end up with a large number of representatives. Quezon City has 2.936 million people, so it should get 13 representatives. Manila, with 1.78 million, can vote for 8 legislators. Davao City, with 1.632 million, can vote for 7 — same with Caloocan City with 1.584 million. Cebu City can vote for 4 while Zamboanga City, Taguig, Antipolo, Pasig, Cagayan de Oro, Parañaque, and Dasmariñas can vote for 3 each.
Consider some more illustrations to drive home the point. Let us take Metro Manila’s case. The area (which I consider a province, for simplicity) contains 12.877 million people — which means that, given a parliament of 465, it is entitled to 59 seats. But we already distributed a total of 51 seats to its constituent legislative districts that reached at least 217,159 people. This leaves us with 8. Thus, we allow all citizens of Metro Manila to vote for 8 representatives at large — on top of citizen of cities considered as legislative district voting for their own representatives.
Cebu province has 4.632 million people, which means it should have 21 seats. 7 of those votes were already distributed in constituent towns with at least 217,159 people — leaving it with 14 seats to be voted at large within the province. Cavite province has 3.678 million and 16 seats. 9 seats were already allocated, which leaves it with 7 seats to be voted at large. And so on.
Weighted Votes for At-Large Reps based on those yet to be represented
At this point, one might want to raise the issue of fairness. A citizen of Pateros doesn’t get to vote for her own representative (for Pateros), but both citizens of Pateros and Quezon City get to vote for 8 Metro Manila representatives at large even as the Quezon City citizen already voted for 13 representatives. Same with the citizen of Danao, Cebu versus that of Cebu City — both get to vote 14 provincial seats at large but Cebu City was already able to vote 4 representatives since Cebu City is a legislative district.
To address this, I introduce a weighting scheme. This is where the “floor rule” introduced earlier really comes in handy. Consider Metro Manila again. Quezon City was assigned 13 representatives. This means that a total of 2,823,067 people are already represented. But Quezon City’s actual population is 2,936,116. 113,049 people from Quezon City are not yet represented. Manila has 8 seats, representing 1,737,272 people. But its actual population is 1,780,148. 42,876 Manileños are not represented.
We total all unrepresented people of towns within the unit of the next aggregation — Metro Manila in this case, then we get the share of each town. That’s the weight of their at-large votes. Quezon City has 6.27% of unrepresented voters in Metro Manila, so the votes of QC citizens for Metro Manila representative-at-large will be weighted 6.27%. San Juan has all of its population unrepresented yet, but in total, it only commands 6.78% of the unrepresented people, so its votes for representative-at-large will be weighted accordingly. Pasay, with one representative already, corners 11.06% of the unrepresented within Metro Manila — so their Metro Manila at-large votes will have the weight of 11.06%. And so on.
Consider Cebu province. Cebu City, with its 4 seats, still has 53,975 of its people unrepresented. This is 1.73% all unrepresented people in Cebu province, which means its votes for Cebu province at-large representatives (14 solons), will be weighted at 1.73%. Lapu-lapu City (Opon), with 1 seat and 190,953 people unrepresented, corners 6.14%. Toledo, the largest town in Cebu province which failed to get at least 1 seat, will have its Cebu province at-large votes weighted at 5.47%.
As with before, at-large representatives win by plurality. For Cebu province, the top 14 candidates in terms of weighted votes get to represent Cebu province at-large.
In summary, we will have 122 representatives directly representing 63 out of 1,634 towns, 299 representatives representing at-large 70 out of 81 provinces, 34 representatives representing at-large 16 out of 17 regions (Metro Manila plus constituent towns take up all the votes for NCR), and 10 representatives voted nationally, at-large. All in all, we will have 465 representative in Congress.
Check out the calculations in the excel workbook here: https://www.dropbox.com/s/d8g3qmkarcrz3t8/districting_new.xlsx?dl=0
Linking Local Legislators to National Legislators
There are additional benefits. As I mentioned earlier, one major problem in the Philippine legislative system is the disconnect between local legislation and national legislation. One reason is that we elect them separately, but as we’ve seen, the more logical explanation is the unsynchronized areas for legislative districts and LGUs.
In a better setup, national legislature should be a “promotion” for local legislators. They practice first doing local legislation in Provincial Boards and City/Municipal Councils, before they attempt to be elected to represent their district in the House of Representatives. That way, they’ll have a sense of . Moreover, there should be mechanisms requiring national legislators to consult the local legislators of their districts before voting on a bill or a resolution.
Having synchronized LGUs, regions, and legislative districts in my proposal, then we can put all of these mechanisms in place. We can require that candidates for national Congress must have served at least one term in the local legislature of their area of candidacy (Town Council, Provincial Board, and if the PDP-Laban federalism proposal pushes through, Regional Assembly). We can require a local legislature resolution supporting the representative’s vote in Congress for the vote to be valid. We can give the local legislature powers to recall a national legislator[5], if there is a perceived failure of representation. We can institutionalize legislator’s caucuses in towns, provinces, and regions.
[1] In his dissenting opinion in Aquino vs. COMELEC (G.R. 189793) (See http://sc.judiciary.gov.ph/jurisprudence/2010/april2010/189793_carpio.htm), Justice Antonio Carpio insists that for creating legislative districts, the “constitutional standards, as far as population is concerned, are: (1) proportional representation; (2) minimum population of 250,000 per legislative district; (3) progressive ratio in the increase of legislative districts as the population base increases; and (4) uniformity in apportionment of legislative districts in provinces, cities, and the Metropolitan Manila area.”
[2] Maybe the function for computing legislative seat is a linear combination of land area and population. I doubt that this is the case, but if it is, it still violates the precept that representation in a unitary republic should be on a per capita basis. We will have to wait for a federal system to see how much constitutional framers value territories.
[3] In effect, a multimember district system would promote party-politics, as opposed to personality-politics, at the local level.
[4] For explorations on the cube root rule, see this study: https://ecpr.eu/Filestore/PaperProposal/3bc100be-56fe-4efc-8d8c-a1f0b85e7f24.pdf.
You can also visit this politics stackexchange page for more information: https://politics.stackexchange.com/questions/364/is-there-an-optimal-size-of-a-parliament
[5] We might have to be stringent though and require a supermajority vote, say 2/3s, for recalls to happen. A special elections will then be conducted.
