The Nobel Prize-winning economist Amartya Sen brought the phenomenon of “missing” women to the attention of the world in the 1990s. His argument is simple: if the laws of demography hold women should make up a majority of the world’s population. They do not in many parts of Asia. Mr Sen calculated the number of “missing women” to be more than 100 million at the time.

A question he did not ask is how many women are “missing” from the electoral rolls in developing polities. This is a pertinent question to ask in the run-up to the general elections on May 11. If the laws of demography hold, we should as Mr Sen did, expect the gender (male-female) ratio to be below 1. In Pakistan, we find this ratio to be far in excess of one, nationally and in every province (Figure 1).

A crude conservative estimate of missing women voters can be found by comparing their actual numbers with the numbers that would have been, had the gender ratio equalled that of the 1998 census. This gives us an estimate of nearly eight million “missing” women voters or an astonishing one-fourth of the existing number of registered women voters.

The electoral system is not only missing registered women voters; in 2008 it was missing the votes of women who were registered to vote.We can crudely estimate gender differences in turnout in 2008 by comparing male-female differences in gender-segregated polling stations, which account for 46 per cent of all stations. The results show that in Punjab and Sindh, the average turnout among women was 11 percentage points less than the turnout among men. This difference exists even if we do not adjust for the high level of ballot stuffing that allegedly takes place in female polling stations (Figure 2).

The largest differences in male-female voter turnouts were in Khyber Pakhtunkhwa and Fata (Figure 2). In KP female turnout was 2.5 times less than Punjab and Sindh even though its male turnout was only slightly lower than the other two provinces. It turns out that KP’s lower than national average turnout is largely explained by the low turnout among women. The important takeaway for parties competing in Punjab and Sindh is that increasing turnout among women promises significant electoral gain given the narrow margins with which contests are decided in these provinces.

It is common knowledge that the vast majority of missing or unregistered voters are also missing from the Nadra database. In a recent (2011) representative survey of four poorer districts in south Punjab, researchers at the Centre of Economic Research, Pakistan found a huge gender gap in CNIC possession among adults, with only 60 per cent of women reporting having a CNIC compared to 80 per cent of men. This suggests that acute levels of gender inequality continue to exist in spite of the significant gains made by Benazir Income Support Programme in this regard. It is not that the state does not know that women voters are missing; it is simply not making enough effort to find them.

The big puzzle is why political do not parties mobilise women voters in spite of the incredibly high electoral returns this promises. BISP is a novel intervention by a political party to register and mobilise women voters. It would be interesting to evaluate the electoral returns that the PPP gets from this strategy in the upcoming elections. The historical experience of other countries tells us that mobilising this huge demographic group has been a game-changer for parties that are able to appeal to women - a lesson that Obama’s campaign team got right in the recent US election. There are lessons here for political parties in Pakistan.

Mobilising women is not easy because of the deep social and political divide that exists regarding the question of the public and political space for women. Mobilising women will not be possible, in this context, unless all the main political parties make a public, non-partisan and unequivocal commitment to uphold the constitutional rights that guarantee dignity, freedom and equality to all citizens; including the right of political participation in terms of holding office and voting and forbid discrimination on the basis of gender. If low turnout exists among women it ought to be a result of the exercise of free choice and not because of violence and coercion. Unfortunately the manifestos of many political parties are extremely lacking in this regard. Parties have to realise that, if for no other reason, it is in their electoral interest to mobilise these citizens to vote.

Mobilising women will not be possible without developing campaigns around issues that matter to women voters. We know from varied contexts that household spending priorities are very different when women control household income instead of men. There is also evidence from the region, which suggests that public policy priorities also tend to vary by gender. This suggests that effective mobilisation will not be possible without getting to know what women want. However, the existing structure of voter mobilisation in Pakistan, the dhara, systematically excludes women and makes it hard for parties to elicit their preferences.

Therefore, what can be done? Evidence from varied contexts tells us that the increased representation of women on general seats has led to substantial reprioritisation of public spending towards public investments (such as education and public health) that are highly valued by women voters. Therefore, an effective mobilisation strategy is to increase the proportion of general election seats awarded to women members of political parties. This has been a long-standing legislative demand of the Women’s Parliamentary Caucus. At present, there is huge under-representation of women among candidates contesting general seats. The share of women candidates on a general seat in 2013 is less than four per cent of the total candidate pool, according to the Fair and Free Election Network.

It is important to realise that the effective mobilisation of women voters has resulted in tremendous social policy and spending gains across a range of democracies. For example, recent research by Professor Grant Miller of Stanford has shown that the passage of the right to vote for women in the US was followed by immediate shifts in legislative behaviour and large, sudden increases in local public health spending. This growth in public health spending led to a decline in child mortality by 8-15 per cent (or 20,000 less annual child deaths nationwide) as a result. Few public policy changes promise such large gains in such a short span of time.

As we stand today, in spite of the promise of substantial development gains, it appears that the electoral system is missing a lot of women voters, while political parties do not appear to miss this pivotal vote bank much.

Faisal Bari is visiting Fellow at the Institute of Development and Economic Alternatives (IDEAS) and Associate Professor of Economics, Lahore University of Management Sciences (LUMS). Ali Cheema is Senior Research Fellow at IDEAS and Associate Professor of Economics and Political Science, LUMS. Syed Ali Asjad Naqvi is Research Director, Center for Economic Research in Pakistan (CERP).