THE GAME-CHANGING OVERSEAS VOTE

Published February 27, 2022
Illustration by Areeshah Qureshi
Illustration by Areeshah Qureshi

Despite the misgivings of the opposition, overseas voting is now officially the law of the land in Pakistan. With one of the biggest diasporas in the world and over nine million potential overseas voters, votes coming in from around the world could possibly entirely swing Pakistani elections. A group of researchers looks at what the impact could be on the country’s politics…


On November 17, amid protests by the opposition, the government and its allies in Parliament approved amendments to the Elections Act, 2017, allowing the use of electronic voting machines and granting voting rights to overseas Pakistanis. This legislation is a landmark event for multiple reasons: it fulfils a key promise made by the Pakistan Tehreek-i-Insaf (PTI) to its overseas base, which has fervently supported the party through thick and thin. It is also the culmination of a decades-long journey to enfranchise our overseas brethren. And it represents the start of what is likely a new era for Pakistan’s politics.

This development also fits in with a growing global trend. The Supreme Court of Canada recently affirmed the voting rights of overseas Canadians. In the UK, following a long and heated campaign by expats, government efforts are now underway to ease overseas voting restrictions. New legislation in Greece also permits Greeks overseas to vote from their place of residence. And a parliamentary committee in New Zealand is calling for review of eligibility requirements for overseas citizens.

Closer to home, in India, the Election Commission recently approached the government to permit overseas Indians — the largest diaspora in the world — to vote through postal ballots. In Nepal, the Supreme Court in 2018 directed the government to enfranchise citizens residing abroad. And the Election Commission of Bangladesh has begun issuing identity cards to overseas citizens which will eventually enable them to vote from overseas.

This trend is best understood in the context of globalisation, as part of an ongoing diffusion of democratic norms and culture. A study even suggests that the chances of a country introducing overseas voting are almost double if a neighbouring country has done so already.

Permitting overseas voting has faced challenges and resistance in countries around the world. Like in Pakistan, elsewhere too the question of divided loyalties comes up, specifically when talking about enfranchising individuals who have taken other nationalities or are dual nationals. An argument put forward by the opposition in Pakistan is that allowing overseas votes would amount to citizens of other countries determining the future of Pakistan.

As per our estimates, in up to 186 constituencies out of the total 272 in the National Assembly, overseas votes are material, i.e. they can potentially change outcomes. Overseas votes, if cast in large numbers, have genuine political significance...

On the other hand, proponents in various countries, including Pakistan, argue that overseas citizens are a key asset and make significant economic contributions in the form of investments and remittances.

As the arguments for and against enfranchising continue (discussed in greater detail below), overseas Pakistanis have been granted voting rights.

Now that overseas voting is officially the law of the land here in Pakistan, we come to what is arguably the most important and also, surprisingly, the least discussed aspect of this entire exercise: the impact. In the current scheme of things, overseas votes represent a huge swing factor that could prove a game changer.

A couple of preliminary estimates have already hinted at this.

An earlier study examined bye-election numbers of 32 constituencies in 2018 and estimated overseas voters could swing 1 in 5 National Assembly (NA) seats, a significant 20 percent. A Geo News investigation concluded that overseas voters could swing 20 hotly contested NA races, in Rawalpindi, Gujranwala, Faisalabad and Sialkot.

We did a deep dive into the numbers and found that the overseas factor could be far, far more dramatic. In the polls of 2018, a considerable number of National and Provincial Assembly seats had razor thin margins of victory, which could easily be swung by votes from overseas.

As per our estimates, in up to 186 constituencies out of the total 272 in the National Assembly, overseas votes are material, i.e. they can potentially change outcomes. Overseas votes, if cast in large numbers, have genuine political significance: in the short term, they can massively influence the composition of the government; in the long term, they can alter the political landscape and change how our political parties operate.

TO ENFRANCHISE OR NOT TO ENFRANCHISE

Article 17 of our Constitution grants all adult citizens the fundamental right to vote. This article has generally been interpreted to acknowledge that this right extends to all citizens, irrespective of place of residence. Overseas Pakistanis have raised calls for enfranchisement since the general elections of 1970. This push has accelerated in the last decade.

The classic argument against enfranchising overseas citizens is very straightforward: overseas voting rights “rupture the essence of democracy by breaking the link between citizenship and residence”, as political scientist Theresa Reidy observes in a 2021 article.

As per Aristotle, “the governors must also be the governed”. Giving decision-making power to those who do not have to bear the consequences of their decisions is fundamentally unfair. Moreover, for a very long time, emigration has had a negative stigma associated with it, the image of people who have left their homeland for greener pastures.

On the other hand, pro-enfranchisement arguments have accumulated significant weight in recent years. Migration is now the norm and overseas citizens routinely act as de facto ambassadors for their homeland in their country of residence.

While giving overseas Pakistanis the ability to vote can be welcome news, these immense numbers also intensify the risks and logistical challenges of the remote voting exercise.

In our particular case, some 55 percent of our migrant population consists of semi-skilled and unskilled labourers, primarily concentrated in the Middle East. Many leave families behind in Pakistan who they visit often. Most have plans to eventually return. Moreover, their economic contribution is immense: in the last fiscal year we witnessed record-breaking remittances of over 29 billion dollars. This amounts to 9.9 percent of Pakistan’s GDP (compared to a global average of 4.9 percent across 165 countries). So in a sense, these people are a critical part of the national economy and have very real stakes in this country.

This conclusion is not really surprising if we consider actual numbers. For instance, the UK had about 233,000 overseas voters participating in their general elections of 2019. Greece has an estimated 300,000 citizens overseas. New Zealand has up to 750,000 eligible overseas citizens, of which 67,000 are enrolled to vote. Canada, the world’s second largest country in terms of landmass, merely has 36,000 Canadians listed on their international register. In stark contrast, Pakistan has over 9 million potential overseas voters, one of the largest diasporas in the world.

Our primary goal in this article is to work through the numbers and explore some thought experiments which emphasise this reality. While giving overseas Pakistanis the ability to vote can be welcome news, these immense numbers also intensify the risks and logistical challenges of the remote voting exercise. Guaranteeing the secrecy and integrity of votes cast in a foreign jurisdiction in such numbers is very daunting. Our second goal is to unpack the key challenges we face on the road ahead.

The overall debate for or against enfranchising overseas citizens still awaits a conclusive resolution. One important reason perhaps is that it hasn’t really been done yet. Remote voting experiments to date have typically involved small numbers of voters, with marginal effect on election outcomes. There is very little data on what happens when overseas voters in very large numbers amount to an influential bloc with decisive impact. Can election systems in developing countries handle the logistics of large-scale remote voting exercises? How would immense overseas numbers impact local politics and governance on the ground? We do not know yet.

Overseas voting for countries like ours is still something of a brave new world. To get a sense of potential impact, we turn to the actual numbers.

THE MASSIVE OVERSEAS SWING FACTOR

To explore what overseas voting could mean for Pakistan, we first studied the margins of victory for all National and Provincial Assembly seats from the general elections of 2018, i.e. the difference in number of votes between the winner and the first runner-up. We compared these margins with the current numbers of eligible overseas voters in those specific constituencies. In essence, we asked, are there sufficient overseas voters in those constituencies who could have swung those seats by voting a certain way?

This exercise took considerable effort. Election results for 2018 are not available in processable format. We had to write code to download results from the ECP website and third-party websites. Moreover, there was considerable discrepancy between these sources. We had to consult ECP as to which numbers were correct. We then undertook a massive reconciliation exercise and manually verified almost 38,000 data points.

It is worth noting here that many prominent elections analyses we have seen thus far have actually relied on incorrect data.

We also faced issues determining the exact numbers of eligible overseas voters. There is no publicly available data source and the mainstream discourse relies mostly on estimates. We considered two primary sources: first, we contacted the National Database and Registration Authority (Nadra) for district-wise Nicop data, i.e. those migrants who possess a Nicop, the National Identity Card for Overseas Pakistanis. This data, dated November 2021, does not include constituency-wise breakdown, so we estimated the number of eligible voters based on the percentage of local voters in each constituency.

This data also has inherent limitations: it does not include overseas Pakistanis who do not have Nicops; and it includes citizens who may have returned to Pakistan but still possess a Nicop.

Second, we contacted ECP to access details of voters holding Nicops in the electoral rolls. This data gives us a precise constituency-wise breakdown of voters. However, there are significant discrepancies between this dataset and Nadra’s Nicop data, with some 1.2 million Nicops unaccounted for.

We, therefore, give separate results here for each dataset. For a sanity check, we also accessed district-wise migration numbers for overseas Pakistanis from the Bureau of Emigration and Overseas Employment. Overall, we filed 7 Right to Information (RTI) requests to various bodies to obtain this necessary data. We are indebted to the Pakistan Information Commission for facilitating this process.

Even a cursory glance at the margins of victory from 2018 is eye-opening. Fifteen NA seats, listed in table 1, were won with razor-thin margins of less than 1,000 votes. For instance, Awami National Party’s (ANP) Ameer Haider Azam Khan, ex-Chief Minister of Khyber Pakhtunkhwa, won NA-21 by a mere 35 votes, edging out PTI’s Muhammad Atif. The number of estimated overseas voters for this constituency is more than 50,000. Prime Minister Imran Khan won NA-131 in Lahore by only 680 votes, outpacing Pakistan Muslim League-Nawaz’s (PML-N) Khawaja Saad Rafique. Here, the number of overseas voters exceeds 45,000. Muhammad Faisal Vawda, former Minister for Water Resources, defeated PML-N President Mian Muhammad Shahbaz Sharif by 718 votes, with over 15,000 overseas voters waiting in the wings.

Table 1: National Assembly seats in 2018 with margin of victory less than 1,000 votes
Table 1: National Assembly seats in 2018 with margin of victory less than 1,000 votes

Other notable names in this group of 15 NA seats include Mohsin Nawaz Ranjha, ex-Minister of State for Parliamentary Affairs, and Fehmida Mirza, ex-Speaker of the National Assembly. Of these 15 seats, in every constituency save one, the number of eligible overseas voters, as depicted in the graph, far exceeds the margin by several thousands of votes. Even a fraction of these votes could decisively swing these seats.

Figure 1: Visualising the margin of victory vs overseas votes for 15 NA seats
Figure 1: Visualising the margin of victory vs overseas votes for 15 NA seats

This trend persists as we go higher. Twenty more NA seats have a winning margin less than 3,000. Again, an overwhelming majority of these seats show several thousands of overseas voters. This list also includes a veritable who’s who of prominent names, including Farrukh Habib from PTI, and Khawaja Muhammad Asif and Rana Sana Ullah Khan from PML-N.

HAIL THE NEW KINGMAKERS

In almost 50 seats, the margin of victory is less than 5,000 votes. In 90 seats, it is less than 10,000. Overall, as per Nadra’s Nicop data, in 186 constituencies out of 272, the number of eligible overseas voters exceeds the margins of victory. Using ECP’s Nicop data, we get 169 constituencies out of 272.

This analysis, obviously, comes with several caveats. We are comparing the number of eligible overseas voters from 2021 with election margins from 2018, which is, of course, not a perfect fit. These numbers will be different in the polls of 2023.

We can also be certain that overseas voters will not line up to uniformly cast their votes for the same candidate. Participation may also be very low. When ECP piloted internet voting for overseas citizens in bye-elections in 2018, a mere 6,146 out of 631,909 eligible overseas voters cast their votes.

However, this materiality analysis is a useful barometer and is commonly used in research literature. It captures a sense of the potential impact and produces actionable information for policymakers, political parties and civil society.

In almost 50 seats, the margin of victory is less than 5,000 votes. In 90 seats, it is less than 10,000. Overall, as per Nadra’s Nicop data, in 186 constituencies out of 272, the number of eligible overseas voters exceeds the margins of victory. Using ECP’s Nicop data, we get 169 constituencies out of 272.

Extending this line of thought, we explored how increasing participation by overseas voters might swing election results from 2018. We specifically computed the proportion of overseas voters in a constituency needed to flip the results of a seat. For instance, if the margin of victory is 1000, whereas the total overseas voters number 10,000, then 10 percent of the overseas voters need to vote a certain way for the second place candidate to win.

Figure 2: Mapping the swing factor
Figure 2: Mapping the swing factor

Our results in Figure 2 depict this impact: even a 5 percent turnout by overseas voters could potentially flip 19 NA seats as per our model using Nadra’s data and 18 using ECP’s data. With a 10 percent turnout, this rises to 30 (Nadra) and 33 (ECP) seats, further increasing to 55 (Nadra) and 59 (ECP) seats for a turnout of 25 percent. A 50 percent turnout could potentially swing 116 (Nadra) and 103 (ECP) seats. At 100 percent participation, we get maximum impact, the result we noted earlier, where overseas numbers outweigh margins of victory in 186 (Nadra) and 169 (ECP) seats. The district-wise distribution is depicted below.

The overseas factor is also impactful at the provincial assembly levels. Some 49 seats countrywide out of a total of 593 had winning margins of less than 1,000. Eighty-four seats had winning margins between 1,000 and 3,000 votes. Seventy-two more seats had margins between 3,000 and 5,000. Overall, there are at least 397 provincial assembly seats in Pakistan where the numbers of overseas voters in 2021 exceed the margins of victory in the 2018 elections. The majority of these seats are concentrated in Punjab (224 seats) and Khyber Pakhtunkhwa (101 seats).

SWINGING GOVERNMENTS

Another thought experiment: what sort of impacts could these flipped seats have on the overall composition of government?

If we recall the election outcome from 2018, PTI won 116 seats and 9 independents then joined the party, taking the tally up to 125 seats. PTI then formed a coalition government with five partners — the Muttahida Qaumi Movement (MQM), PML-Q, the Grand Democratic Alliance (GDA), Balochistan Awami Party (BAP) and Awami Muslim League (AML) — who brought in 17 more seats, taking the total to 142.

As per our calculations, for the PTI to form a simple majority government, along with the newly joined independents, would require 12 additional seats, with 19,356 additional votes. Extending this line of thought, for a simple majority government without the 9 independents, PTI would need 21 more seats. These could be flipped with 76,036 votes, giving PTI a total of 137 out of 272 General seats in the National Assembly.

A PTI simple majority government (with independents) + 12 seats with 19,356 votes
A PTI simple majority government (with independents) + 12 seats with 19,356 votes

A PTI simple majority government (without independents) + 21 seats with 76,036 votes
A PTI simple majority government (without independents) + 21 seats with 76,036 votes

On the other hand, how many votes would it take to defeat the PTI?

There are multiple paths for the opposition to join hands to form a government. We considered a simplistic scenario where the PML-N and the PPP form an aggregate bloc with a viable path towards government. An inflection point could happen if PTI were to lose 5 seats. As depicted in Table 1, if PTI loses NA-140, NA-239, NA-114, NA-60, NA-131, 4 of these seats go to the PML-N-and-PPP bloc, thereby equalising their numbers with PTI at 111 seats each.

At this point, PML-N and PPP could announce their intention to form a coalition government. When things are this close, it becomes a race for different sides to rack up the full numbers first, there is a flurry of backdoor communications, negotiations and incentives, offers and counteroffers. If the Muttahida Majlis-i-Amal (MMA) were to hop on board the PML-N and PPP bloc with its 12 seats, it could very well have spelt the end for PTI’s government ambitions. All this would have taken is 2,493 additional votes to swing those 5 seats. The number of overseas voters in these 5 constituencies adds up to 170,476.

MAKING VOTES COUNT

These numbers — and their consequent impact — may seem startling at first glance. Even small amounts of votes can potentially swing the government in a big way. We consider some implications.

For one, the overseas impact depends considerably on the processes and methods put in place for voters. One probable reason for the abysmal turnout by overseas voters in bye-elections of 2018 (less than 1 percent) were restrictive processes, several of which have been acknowledged by the ECP.

For instance, eligibility requirements for voters included a machine-readable passport as well as a Nicop, which automatically excludes large numbers of overseas citizens and appears redundant and extralegal. Internet voting also requires digital skills, which many citizens might not possess. Moreover, citizens could only cast votes for the duration of the election day period in Pakistan; a recent audit report by Spanish experts noted that this amounts to “de facto disenfranchising voters that live in other time zones like Australia or America”. There was a distinct lack of outreach and voter education efforts and support services.

To encourage overseas participation, we hope the ECP considers progressive rules and processes.

We hope these findings also underscore the critical need for structural reforms at home to make every vote count. A very disturbing revelation from the general elections of 2018 was the report that some 1.67 million votes were spoilt and rejected from the total count. These votes had the potential to swing 49 NA seats. And if we go even further back to the 2013 polls, alarming numbers of precinct-level totals simply could not be accounted for. Irregularities in such massive numbers cast a very, very dark cloud on our notions of democracy.

THE INTRACTABLE SECRET BALLOT

The fundamental challenge for any remote voting exercise is the secret ballot. Universally recognised as a core component of democracy, the secret ballot is part of the Universal Declaration of Human Rights and mandated in Article 226 of our Constitution. Our Internet Voting Task Force (IVTF) report from 2018 explains the reason: “The rationale behind this, dating back to ancient Greece and Rome, is that if outside parties become privy to a voter’s choice, it opens the door to bribery and intimidation, thereby ultimately corrupting the electoral process.”

This explains the vote casting ritual: the voter steps alone into the polling booth, with curtains and partitions in place to ensure privacy and then casts his vote. The ballot has no identifying information on it, it is cast in a ballot box where it mixes with the cast votes of other citizens and is effectively anonymised. The voter’s intent is counted in the total result but his individual choice cannot be determined.

But once voting is taken out of the precinct setting, without any oversight, the dynamics of privacy and application of the law change.

This problem is best understood with a real-life example: when current Prime Minister Imran Khan cast his vote in 2018, TV cameras inadvertently revealed his vote. This violated Section 178 of the Elections Act, 2017, which specifically refers to “interference with the secrecy of voting”. The ECP withheld Imran Khan’s victory notification until he submitted a written apology.

A few months later when overseas Pakistanis cast votes in bye-elections using an Internet voting system, some overseas voters posted snapshots of their votes on Twitter. Technically, this is the same offence — the secrecy of the vote has been violated. But how does the Election Commission enforce the law overseas?

The implications are far more severe than just publicising the vote. To quote the IVTF report again: “Casting votes outside a poll-booth environment typically enables vote buying and voter coercion. In our particular case, there is a very real possibility that votes may be bought and sold and coerced overseas in regions where the ECP has no mandate to investigate or prosecute such attempts.”

Writing some 2,000 years ago, the Greek historian Plutarch identified vote buying as the beginning of the end for Rome: “The abuse of buying and selling votes crept in and money began to play an important part in determining the elections. Later on, this process of corruption spread to the law courts, and then to the army, and finally, the Republic was subjected to the rule of emperors.”

Vote buying and coercion, though rare in the West, are unfortunately persistent realities in the developing world. Researchers have identified various reasons: these practices tend to thrive in countries where citizens have a low socio-economic status, corruption is systemic, elections administration is weak and society is patterned along hierarchical lines, as opposed to egalitarian models. In past surveys in African and Latin American countries, some 15 percent of respondents reported being offered goods or money in exchange for their vote.

These risks are significantly exacerbated in a remote setting, where it is extremely difficult to police voting irregularities. These risks have also been observed in countries which allow remote voting to local populations.

Strong family and social networks, which can be a boon for immigrants, can also prove a negative in the sense that they can exert pressure on citizens to vote a certain way. Landlords and employers overseas — referred to as power brokers in the research literature — also exert an influence. A study of the 2018 federal elections in Mexico, which included overseas citizens, found that approximately 32 percent of Mexican immigrants in the US experienced vote-buying.

THE WAY FORWARD

In the larger picture, these two factors — an immense and highly relevant overseas vote count and a remote voting method with inherent secrecy issues — can result in a volatile combination.

Moreover, the method of remote voting we will use has yet to be determined. Convenient and easy-to-use modalities — such as internet voting — encourage voter participation, but carry inherent risks and expose elections to the threat of hacking and manipulation. Relatively safer options — such as postal and embassy voting — are not as convenient and will likely reduce voter participation.

We have to navigate these trade-offs on the road ahead. But this challenge, immense as it is, should not be construed as an argument against enfranchisement. These shortcomings need to be studied and addressed before we embark on any large-scale overseas voting exercise. This is simply Murphy’s Law. However, we can learn significant lessons from international experience and innovate.

It is essential this entire process should be undertaken in a transparent and democratic manner to engender trust among citizens. Thus far, we lack forums and mechanisms to engage stakeholders in constructive dialogue. These need to be set up on an urgent basis. These particular requirements should top ECP’s list of priorities.

These are not easy tasks; there are no ready-made solutions we can simply borrow from other countries. As noted earlier, we are venturing into bold new territory. The next few years can prove transformative for Pakistan’s elections and democracy.


Salman Shabbir is an overseas Pakistani who works on technology transformation programmes and advocates for overseas Pakistanis’ rights. He tweets at @salmanshabbir

Ain Ul Zia and Hina Binte Haq are researchers affiliated with NUST in Islamabad

Taha Ali, also at NUST, has a postdoc in election systems and advises the government and ECP on electronic voting. He can be reached at taha.ali@gmail.com

The authors are also part of the PivotPk collective to promote responsible use of election technology (@pivotpk)

Calculations and source documents used for this article are publicly available. The writers are also launching a public dashboard to access this data

Published in Dawn, EOS, February 27th, 2022

Opinion

Editorial

Breaking the deadlock
09 Dec, 2022

Breaking the deadlock

It is time for PDM and PTI to show flexibility and realise that the future of over 240m people is at stake.
A targeted killing
09 Dec, 2022

A targeted killing

IF there were any doubts about a sinister, transnational plot to kill journalist Arshad Sharif, the 592-page report...
Dog-bite epidemic
09 Dec, 2022

Dog-bite epidemic

AN exploding population of stray canines has fuelled a dog-bite epidemic in Sindh, with the provincial health...
Worsening hunger
Updated 08 Dec, 2022

Worsening hunger

THAT the dollar liquidity crunch has started hurting the import of essential items such as vegetables and raw...
Bannu beheading
Updated 08 Dec, 2022

Bannu beheading

The state must take up the cudgels and neutralise barbarism before it spreads.
Smog misery
08 Dec, 2022

Smog misery

IF 2022 has taught us anything, it is that generations of reckless disregard for Mother Nature has accrued very ...