Systems and flaws


Institutional design for deepening democracy

A recent article in a national daily talked of making voting compulsory in Pakistan, as one way of deepening democracy and ensuring participation in the democratic process. Apart from practical issues, related to implementation, there are questions whether ‘democracy’ and ‘compulsory’ go together, but even more importantly, if we are thinking of deepening democracy and ensuring wider participation of the public, and institutional change is possible, we should be thinking of a number of options.

One of our professors, an acknowledged and well respected expert in constitutional theory and design of political systems, especially voting systems, and one who had helped write a number of constitutions for states that had become free in the 1950s/60s, and had designed their political systems, used to begin his lectures by saying ‘When states/governments come to me for advice on political systems and their design I always begin by asking them: “what sort of results do you want”.’

Institutional design matters and is not neutral. To take one issue only, voting rules that we set up have an impact on how aggregation will work and thus how democracy will function. Though a run up to an election is not the right time for such thinking, but since these issues are more medium term it is not a bad time to start debates on some of these issues. Here I will focus on some issues with vote aggregation systems.

We have the ‘first past the post’ voting system where, irrespective of the number of candidates for a seat, whoever gets the most votes cast is declared the winner. Consider some implications. Suppose a constituency has 100 eligible voters. 10 candidates stand for elections. A gets 11 votes, B gets 9 votes and the other 8 candidates get 10 votes each. Even assuming 100 percent participation, we have someone winning the election with 11 percent of the vote and with 89 percent voting against her. If only 40 percent of eligible voters vote, roughly the norm for Pakistan, the winning candidate could win with even 5 percent of eligible voters.

The first past the post system also favours parties that have a geographic concentration of voters. So MQM and ANP are likely to do better in first past the post system as their supporters tend to be more concentrated in specific areas. And the system tends to work against newer parties as it requires a threshold concentration of votes for a party to win even a single seat. For example, imagine PTI had 15 odd percent of the cast vote across the country. But if the winning percentage was around 30 percent of the cast vote, with votes split between 4-5 candidates, then PTI, despite polling a whopping 15 percent of the vote across the country, might end up with no seats in the Parliament. Should the 15 percent who support PTI have no representation in state at all?

Proportional representation, used in many countries, gets round the two problems mentioned above. Parties get seats according to the proportion of votes they get. It seems a fairer way of creating aggregation and its relationship to representation, and it allows new comers an easier entry. But it has its own limitations. Proportional representation weakens the constituency-politician relationship as parties offer a roster of candidates and these candidates are no longer tied to specific geographic constituencies. If national and provincial representatives are to work for the people of an area, proportional representation systems will need tweaking before that connection can be made. Similarly the phenomenon of independent candidates, contesting for one seat alone, is also less viable under the proportional system.

Proportional representation also makes outcomes of coalition governments more likely. In polities that are very divided, as Pakistan is, on lines of religion, sect, ethnicity, clan and class identities, the vote is likely to be quite dispersed as well. But this can be seen as a strength too. Where first past the post system can work against the inclusion of smaller groups, proportional representation system allows them to be a part of the system and the larger parties, to form coalitions, would need to woo smaller parties: this might be an important means, for divided polities, of ensuring the inclusion of the disenfranchised into the mainstream.

There are many other ways we can deepen participation. We could have stipulations on the minimum percentage of vote caste or actual registered vote in a constituency that the winning candidate should have. But this does increase the burden on the Election Commission and complicates the voting process. If no candidate has secured the minimum required in the first instance, we might need more rounds of voting to allow candidates to get the required percentage. This raises costs and the administrative burden. But it has benefits of broader and deeper representation.

There are many more sophisticated voting schemes that are also available, each with its own tradeoff of how it aggregates voter choice, the costs and the administrative burden involved. For example we can have runoffs in top candidates from round one to ensure majority support for winning candidates. We can have preference voting and aggregation by preferences. This allows us to cater for strategic voting on the part of voters but the system gets quite complicated and it is not possible to aggregate votes without some computing ability. But even if systems need to be kept simple and cost effective, there are a number of options that can be explored. The issue of which system cannot be answered apriori: it depends on what the objectives of the polity are.

Institutions and institutional rules matter in now democracy functions or can function. We do want our democratic system to become deeper and garner more citizen involvement and support. Having free and fair elections, and ones that reflect and aggregate voter choice faithfully are important parts of how these goals should be achieved. Though it might be difficult to have a deeper discussion on these issues in the run up to this election, this is a more longer term debate, linked to the health of our democracy, that should continue post elections as well. But on all such issues, debate we must have.

The writer is an Associate Professor of Economics at LUMS (currently on leave) and a Senior Advisor at Open Society Foundation (OSF). He can be reached at [email protected]


  1. Making more uneducated to vote wont solve the governace issues;
    Democracy has pre-requisites whic must be known

  2. In Pakistan, the uneducated mostly rural voters have more rational choice than the urban educated ones. Democracy has its own pitfalls, of which one is corruption. But it is preferred because it is relative in nature. Power corrupts but the absolute power corrupts more absolutely.

Comments are closed.