In last year’s general elections the Election Commission of Pakistan collected gender disaggregated data for the first time. The effort, mandated under the Elections Act, 2017, allows for an analysis of differences in voting choices of men and women.

The Free and Fair Election Network (FAFEN) followed a multi-phased methodology that involved analysis of ECP-furnished data from35,988 male-and female-specific polling stations across 246 NA constituencies (88 percent) and generated some ground-breaking findings. These identify multifarious issues that impede women’s participation in elections and politics and also provide evidence to guide future strategies to improve women’s political autonomy in mainstream politics.

Some of these are presented here. The analysis is based on data from across 52,565 (32 percent) census blocks for which the results from both the male and female polling stations were collated and analysed. While some of the gender-related findings have been reported earlier, the trend of women’s choices deviating from those of men’s in the same area is an important indicator of increased political agency and exercise of autonomy.

Missing women

Women constitute 49 percent (101.314 million) of Pakistan’s population. Despite a high-paced citizen registration process prior to the 2018 elections, the number of registered women voters last year was 46.731 million, which constituted 44 percent of the electoral rolls.

The gap between the number of men and women registered as voters has been growing. The primary cause for this gender imbalance is believed to be that many eligible women do not possess CNICs which is a legal requirement for voting. Despite concerted policy, campaigns and technological and pragmatic efforts to close this gap, the gender electoral roll imbalance actually increased from 10.97 percent in the 2013 elections to 12.49 percent in 2018 elections. This gender gap is the largest in Balochistan (15.65 percent), followed by Khyber Pakhtunkhwa (KP) (13.65), Sindh (11.02) and Punjab (11.095).

The 10 percent women

FAFEN’s collation reveals that around 65 percent National Assembly constituencies (173 out of 272) had a gender gap of more than 10 percent. Specific to provinces, 83 constituencies in the Punjab, 47 in the KP, 28 in Sindh and 15 in Balochistan have more than 10 percent women missing from their rolls. A similar pattern is observed in the Provincial Assemblies where 381 out of 577 constituencies have a gender gap of above 10 percent.

While the overall gender gap of enrolled voters remains under 20 percent, the troubling aspect is that women’s absence from electoral rolls has steadily increased by 10 percent (or more) in 80 out of 127 districts across the country (63 percent). Very few districts (just 17) have managed to be shielded from a large increase in the widening gap between registered men and women voters.

Contesting women

The Elections Act, 2017, introduced a requirement for allotting 5 percent of party tickets for general seats to women, albeit without legal consequence for its non-compliance. The benefit of this provision remains unclear since no significant improvement in the number of women candidates was seen. Just 463 women ran for national and provincial assemblies in GE-2018 which is only 18 more candidates compared to general elections 2013 (though a marked growth over 192 candidates in general elections 2008). Women winning on general seats in 2018 remained at 16 as in 2013 and down from 26 in 2008.

No zero turnout

Of the registered women, 21.746 million cast their votes in 2018 for the National Assembly. This made up 40 percent of the 54.657 million votes cast. In general elections 2013 there were 13 polling stations where no woman had turned out to vote on polling day but none such in 2018. Credit for this is attributed to the ECP’s Gender Affairs wing and collaborative efforts civil society drives to increase women’s participation.

Women voter’s turnout was 47 percent of the registered voters, with Punjab, Sindh and Balochistan showing an average voter turnout gap of about 8 percent between men and women while Khyber Pakhtunkhwa had a turnout gap of almost 19 percent. Several metropolitan cities in each province showed a larger gender gap in turn out compared to other parts of the provinces – Peshawar 25 percent; Karachi West 14 percent; Lahore 13 percent; Quetta 13 percent.

According to the FAFEN analysed data, female polling stations generally showed a lower turnout with 73 percent of the polling stations hosting less than 20 percent turnout (mostly KP).This raises serious questions of efficacy at the expense of making ‘cultural concessions’ or overestimating mobility issues for women voters. A more extensive study of all polling stations will allow for more accurate policy on this matter.

Over-and-underestimat-ing women’s agency

A gender analysis of electoral choices exercised at the 2018 polls challenges misconceptions about lack of autonomy amongst women voters and the myth that they always vote for the same political party as male family members. Men and women’s voting choices were found to be similar in 82 percent of the electoral areas analysed by FAFEN(that is, 43,178 out of 52,565 in number) but the ‘deviancy’ of women’s voting from men was found in 18 percent – up from 11 percent in 2013.

Clearly, women voters are ‘deviating’ considerably and parties stand to gain by directly engaging and being responsive to the demands of women voters.

‘Deviant’ women

The voting choices of men and women were more homogenous in ICT than any other area (86 percent voted for the same winning party but in 14 percent of electoral areas women deviated). Balochistan has two districts where there is no difference between the electoral choices of men and women in any of electoral areas studied, whereas five provincial districts had 31 to 40 percent electoral areas where men and women voted differently.

The difference in voting choices was starker in Khyber Pakhtunkhwa and Sindh, where men and women voted for the same winning party (77 percent and 79 percent respectively, of the electoral areas) but women deviated in around a quarter (23 percent and 21 percent) of the electoral areas and voted for a different winning candidate/party.

The ‘deviancy’ patterns in KP and Sindh are more vibrant than for the Punjab where women’s autonomous voting agency is flatter and less variable (85 percent voted for the same winner and in 15 percent electoral areas women chose a different one). In Dadu and Badin in Sindh, the differential in women’s choices from men’s for winning parties were recorded in 41 percent and 35 percent of electoral areas respectively, whereas in the Punjab, higher women’s voting ‘deviancy’ was noted in Muzafargarh (26 percent), Khushab (25 percent) and Nankana Sahib (22 percent).

Findings on gender-based comparative voting for specific parties may be summarised as follows with some broadly derived observations;

ICT: The PTI and the PPPP won more men’s votes in the capital than from women in the analysed electoral areas. However, for PML-N’s wins in the ICT, women’s votes were found have deviated three-fold over that of men’s choice.

It was predominantly men who voted for the PTI and the PPPP and more women for the PML-N in the ICT.

Punjab: More men voted for the PTI in the areas where the party won in the Punjab while women’s ‘deviant’ votes were found to be higher where the PML-N and PPPP won.

More men voted for the PTI while women preferred to vote for PML-N and PPPP in Punjab.

KP: More men voted for the PTI and the MMAP wins in the electoral areas while more women’s ‘deviant’ votes were found for PPPP and PML-N where these parties were successful.

The highest rate of deviancy of women’s voting overall was observed in KP (80 percent) in favour of the PPPP wherever the party won in the electoral areas.

Sindh: An overwhelming 83 percent of the PPPP’s victory was from Sindh province. In almost every district of Sindh, the choice of women voters for this party dominated that of male voters.

Without women voters the PPPP would struggle to win its majority in Sindh. Women’s deviancy votes were three times higher in urban Sindh for the ANP wins.

Balochistan: Independence or deviancy in electoral choice by women for the PTI was highest in Balochistan’s electoral areas.

Balochistan is the only province (other than some electoral areas in rural Sindh) where women’s voting deviancy went in favour of the PTI, although the party that won most support from women’s ‘deviant’ voting overall in the province was the ANP followed by the PKMAP (although the figures for PPPP are still being analysed).

(Another set of raw figures awaiting cross-checking is women’s (fairly substantial) deviancy vote in KP (urban), Sindh (rural) that suggests these favoured the Tehreek-i-Labbaik Party). 

Inter-provincial deviation

Another interesting point of ‘deviation’ was that smaller parties won in electoral areas according to a similar gender pattern as seen with the PTI, that is, with men’s votes and not from women deviating their votes for them. However, for other bigger parties’ wins (PML-N, PPPP, and ANP) the trend was found to be reverse (that is, more women ‘deviant’ votes).

This provincial level data is being further refined. A detailed analysis will be cross-checked and released soon. However, the main point is that it’s erroneous to think women’s deviations would not impact the electoral results; the overall share of victory for political parties was such that the PML-N, the PPPP and the ANP were victorious at 54 percent, 57 percent and 61 percent of electoral areas respectively, where women’s votes deviated from men’s. 

Cross-provincial deviation

There are other interesting cross-provincial variables; comparatively, more of women’s ‘deviant’ votes were cast in those electoral areas that converted into a win for the success of the PPPP in KP (79.4 percent electoral areas) and the Punjab (61.3 percent electoral areas) than in the party’s own province of Sindh (55.5 percent electoral areas) which enjoys base support.

A full 100 percent of women’s ‘deviant’ votes that led to a win for the ANP were those in Punjab, followed by in Sindh (76.9 percent), Balochistan, and the party’s own province, KP, had the least amount of women’s ‘deviant’ votes for winning parties (60 percent electoral areas) where it enjoy base support.

The PML-N received the most number of ‘deviant’ votes (by women in 172 electoral areas compared to that of men in 52) to win in the ICT (76.8 percent) followed by in KP (54.1 percent) and in Punjab (54 percent) - enjoying base support in the latter. 

Rural-urban results

These results challenge traditional estimates about rural-urban voting behaviour. Cumulatively, the PTI’s wins based on women’s deviant votes amounted to under 50 percent electoral areas in urban and rural areas in the KP, the ICT, and the Punjab, and in 51 percent electoral areas from rural Sindh, and 53 percent urban and 50 percent rural in Balochistan.

With much more variation, the PML-N wins achieved the highest number of female deviant votes from the ICT electoral areas (rural 82 percent and urban 66 percent), followed by those in the KP (65 percent urban and 52 percent rural). In Punjab, the PML-N gained wins where women’s deviant votes amounted to 54 percent in urban and 53 percent rural electoral areas and in under 50 percent of electoral areas in Balochistan.

The most notable pattern of women’s ‘deviant’ voting was observed in the PPPP wins in the KP as found in 79 percent of urban electoral areas and even higher at 80 percent of rural electoral areas. This is followed by women’s votes for the party’s wins in Punjab (60 percent urban and 62 percent rural electoral areas) and less in its wins in the home-base of Sindh where a more balanced division of women’s ‘deviant’ votes were cast in 58 percent of urban and 54 percent rural electoral areas. In Balochistan, the PPPP’s wins showed women’s ‘deviant’ voting in 63 percent of urban electoral areas and 41 percent rural.

Deeper gender analysis is needed since there is rich empirical data available now but clearly, the evidence shows that neglecting women’s electoral agency is a direct and consequential disadvantage for political parties. It also suggests that undervaluing rather than expanding women’s electoral concerns is a losing strategy. More qualitative studies will support and enable a better grasp of the evolving nature of electoral behaviour in Pakistan.

*This does not mean these necessarily translated into overall wins - just the electoral areas based on the polling stations analysed.

 The writer is a board member of the TDEA-FAFEN. Any errors are a result of her own tentative reading of the data which is still being refined and analysed and is available on the FAFEN website.