As things stand, Pakistan is only a couple of months away from a general election, which means there’s a lot more election related analysis in the national press these days. There was an IRI poll a few weeks back, and then Herald and SDPI launched their own election poll just last week. Both have different samples and fairly different methodologies, and naturally, both need to be taken with much more than just a pinch of salt. This, by the way, is the first of two posts on elections, and covers the findings of IRI’s latest opinion survey . The second post will cover how pre-election reportage has developed a standard template, which in turn relies on far too many anecdotes and stereotypes.
Predicting elections is always tricky business, and more so in Pakistan where you have a range of factors that no survey can possibly account for (‘angels’, for example), and where election-day activity plays a large part in driving turn-out, voting patterns, and hence, outcomes. More importantly, there’s a substantive difference between expressing approval of a party, and voting for a candidate who’s running on that party’s ticket. I think that’s where even the most rigorous survey data fails us on the prediction front: the disjoint between candidate appeal and party appeal is just far too immense – a function of lack of party institutionalisation – especially in rural areas, patronage-based voting, and largely whimsical allocation of electoral tickets.
Think about it this way: Voting is an incredibly localised affair for the vast majority of the citizenry. For most people reading this post, who coincidentally are not the vast majority of Pakistan’s citizenry, the act of voting will most likely be about asserting a principle or voicing an opinion. For everybody else, a vote is political currency. You can exchange that currency for something in return – most likely some meagre form of patronage, or in drastic cases, avoiding the oppressive cost of NOT voting. A party will figure into this calculus for a much smaller portion of the voting public, which is why we have such high degrees of anti-incumbency, as well as equally high-degrees of candidate switching. So when a person expresses approval or disapproval of a party, his or her notion is most likely based on a media-made-and-consumed image of that political party. The PPP has a 14% approval rating, according to an IRI poll, but does that mean it’ll get 14% of the vote or 14% of the seats? Obviously not. The PPP will get a lot more than that, and by most reasonable guesstimates, it’ll be in the fray as far as post-election government making is concerned
Personally, I’d like to see an approval survey that’s representative at the constituency level, but that’s clearly not happening any time soon. So we make do with what we have, and look at survey data and the reactions this data induces from supporters, opposers, and party office-bearers (‘PML-N has rigged the IRI survey’ seems to be a popular refrain). The latest IRI poll, released last month, was carried out from the 2nd to the 21st November 2012 and had a sample size of 4,997 – representative of the adult population at the provincial level.
The following table gives the break-up of respondents on a range of demographic indicators:
Since detailed results are only released to political parties and not to the general public, we have to rely on watered down press-releases to look at the standing of each party nationally, and within the provinces. Here’s what IRI reported as the approval percentages for each party at the national level:
PML-N’s overall appeal is a function of its consolidation in Punjab over the last year or so, where they’ve finally managed to get some semblance of an act together. Very obvious (and very messy) infrastructure projects, such as the Metro Bus System, seem to have a degree of resonance within the public imagination, and convey a sense of ‘something is being done’ – regardless of how superficial, financially unsustainable, and short-term it might be.
The most glaring thing about the released results from Balochistan is the lack of smaller/nationalist parties in the results. The BNP-M and the PKMAP both boycotted the 2008 election, but have given an indication of contesting this time around. Balochistan, as expected, remains the most difficult province to call simply because of the exogenous factors (army, nationalists, and the secessionist movement) impacting political activity.
Another interesting aspect is that a provincially representative sample captures the reflection of Punjab’s internal politics on a statistically calculated national mood, which is why a comparison of IRI’s last survey – held in August – with the latest one confirms the upward trend in PML-N’s popularity:
As I wrote earlier as well, PTI’s strategy of public mobilisation during 2011-2012 was slightly mistimed, given how the election was still a full year (and more) away. Fast forward to 2013, and PTI’s gone into an extensive intra-party election process, which while being an incredibly positive step for party democracy, has allowed fractures to emerge on the eve of a general election. Party leaders like Masood Sharif Khattak have openly questioned the logic of holding internal elections at a time when the party should be allocating tickets and cementing candidates in their constituencies. This has had a direct bearing, in my opinion, on their approval ratings over the last 6 months. The more the party focuses on internal matters, the less it features on national media outlets.
The PPP, on the other hand, has maintained a 14% approval rating over a 6 month period, which probably just feeds into the theory that the party has a loyal support base and no real competitor in rural Sindh. The PML-F, which interestingly doesn’t figure in IRI’s results for Sindh, is busy cobbling together an anti-PPP alliance in the province that might just have more than a marginal impact on electoral outcomes.
Finally, I’m not one for making election predictions (especially not ones that the internet will remember) but there are a few things which appear to be near-certainties: 1) a hung parliament, 2) PML-N’s stronger-than-expected showing in Punjab, 3) PTI’s emergence as an independent bloc in parliament; the size of this bloc, however, remains difficult to predict.
(For more on election related stuff, be sure to follow Saba Imtiaz’s election watch blog)