What if the Elections were Held Today?
The recent Innovision poll provides a very good snapshot of the political state of play with less than 6 months to go before elections
Bangladeshis have traditionally been a highly politicized people. Put two of us together and we would form three political factions, an old joke goes, one for each of us and the third one where we are against the world.
Of course, politics took a back seat under the fallen Hasina despotism. But political debates and arguments, from lounge addas to the social media to TV screen, have been louder than ever since the despot fled last August.
Amid the cacophony of our newly freed political discourse, however, it is important to separate the signal from the noise. And opinion polls are crucial for that, particularly as we head to February’s election.
With its second edition, Innovision’s People’s Election Pulse Survey is becoming an important feature of our incipient democratization process. The full results of the survey of 10,413 persons have just come out, with a treasure trove of data for political animals of all types, from armchair pundits to party leaders.
Of course, no poll is ever a foolproof prediction of the only poll that really matters -- the one in February where 120 million voters will participate. However, the latest poll results do give us a very good snapshot of the political state of play as of the first half of September.
What Do the Numbers Tell Us?
Firstly, fewer than three out of five have made up their mind about their preference -- oddly, perhaps, this is a decline since March (57.8% now vs 62% then). If the election was indeed held in the first half of September, and the turnout was actually 57.8%, it would have been a departure from our recent democratic history, though in line with older experiences.
This is shown in Chart 1, where the green bars are turnouts in historical elections, and the red one is the percentage of "decideds" in the poll.
Chart 1: Election turnout
Source: Election Commission, Innovision. 1996 refers to the June election.
Why then are so many yet to decide? Well, the poll provides a clue. When asked about the importance of the candidate (qualification and experience) versus the party (manifesto and symbol) in voting decision, four out of five respondents chose the former.
We haven’t had any meaningful election, at any level, for well over a decade. Parties are yet to announce candidates formally. In many areas, the veteran candidates who ran in 1991-2008 elections have retired or passed away.
Is it any surprise then that many voters are yet to make up their mind?
One key takeaway for political parties is that candidate selection would be crucial to reach those who are still undecided. Of course, parties know this very well. Perhaps political pundit class should pay more attention to the candidates, and local politics, than grand narratives and ideological fault lines at the national level.
Turning our attention to those who have decided, one in six declined to reveal their preference, a marked drop from over one in three in March. Why might respondents not want to reveal their preferences? One theory is that in a low trust society like ours, still scarred from the Hasina years when one could be jailed for posting political jokes on Facebook, many might be uncomfortable with sharing their political preferences. If this conjecture is correct, then the halving of the share of those declining to reveal their preference is perhaps a sign of the public’s confidence in the overall political process, which would be consistent with the general confidence reported in the first tranche of results.
Let’s now turn to those who have revealed their preference. As shown in Table 1, relative strengths of parties have not changed materially (indeed, the changes are not statistically significant) between March and September, with one exception -- the Awami League.
Table 1: Voting preferences by party
Source: Innovision.
Whether in full sample, or only looking at those who have decided and revealed their preferences, the support for the Awami League has increased since last March, while those of other parties are broadly unchanged (that is, the changes are within margins of error).
Further, when asked about the Awami League’s right to participate in the elections, the respondents returned a statistical dead heat between "All parties should be given the opportunity to participate" and "Awami League should not be allowed to participate in elections before trial."
Putting all of the above together suggests that the Awami support might be under-reported in the poll, and the analysis that follows.
It is, however, the Awami League that is most interesting. Notwithstanding the recent increase in its revealed support, or the potentially "unrevealed" support, the party’s standing among the general public is worse than its historical low of the late 1970s. Back then, the party received the support of around an eighth of total voters, or a quarter of those who voted in the 1979 election. In the four high quality polls that have come out since Hasina’s flight to India, the Awami League support has been much lower than that. This is clear in Chart 2, which also shows that BNP and Jamaat have been solidifying their position as the two top parties in the country.
Chart 2: Party support among all surveyed
Source: Innovision, BIGD. The surveys are not directly comparable because of differences in methodology and sample size.
The sample size and scope of the poll is not conducive for constituency level comparisons. However, it is possible to infer areas of support from the map below, which shows party support among revealed voters in different divisions.
The map shows that Jamaat support is particularly strong in North Bengal and Barishal, which collectively have 93 seats. In contrast, BNP has sizeable lead of 13-20 percentage points in all the other divisions that make up 207 seats. That is, the vagaries of the first past the voting system means that BNP may well have won a comfortable victory had the election been held earlier this month. Further, the above electoral map suggests that a Jamaat victory or a BNP dominated one-party parliament would have been very unlikely.
Of course, Awami League is barred from operating until the trials of the July massacres have run their course. As it happens, the second preference of around half of the revealed Awami supporters is BNP and if the Awami League is not on ballot, BNP stands to gain significantly (Table 2).
Table 2: Voting preferences by party if AL is not allowed to participate
Source: Innovision.
There is a perception among the pundit class that some in BNP are seeking to rehabilitate the Awami League. Senior BNP leaders have said that while they want to hold Awami League accountable through a judicial process for its past crimes, they are in principle against executive bans on parties. However, against this, the poll suggests that BNP stands to gain most if Awami candidates are barred from the hustings, making it unlikely that they would secretly be trying to get them on the ballot.
Of course, the election is still over four months away, and as every political analyst knows, the only poll that matters is the one on the election day. With only three out of five surveyed responding their voting decision has been made, there is a lot that can happen between now and February.
There are at least 20 weeks of campaign ahead of us -- that is an eternity in politics. Let the race begin!
What's Your Reaction?
