CODESRIA Bulletin Online, No. 6, May 2023

Sierra Leone’s Voter Registration Data Discredits the Midterm Census Data: What are Implications for the Presidential Election of June 2023?

Yusuf Bangura

Nyon, Switzerland 

Statistics Sierra Leone carried out a midterm population and housing census in December 2021. The declared purpose of the census was to review the country’s development status since the 2015 census and update the base maps and enumeration areas. Opposition parties challenged the procedures by which the census was conducted, and argued that its real purpose was to redraw the constituency boundaries for the June 2023 general elections. They urged their supporters to boycott it. Donor agencies, such as the World Bank and the European Union, were also critical of the exercise. The World Bank even withdrew financial support from the project, citing ‘insufficient time ... to satisfactorily address all pending actions’ (Reuters 2021) and deliver a credible outcome.

Questionable Population Figures
The statistics agency ignored the criticisms and boycott and went ahead at full throttle. However, the data it released to the public in May 2022 was met with widespread disbelief and condemnation. It showed high levels of depopulation in some regions or districts and overpopulation in others when compared with previous censuses and datasets.

There were four main criticisms. First, the census data overturned the lopsided population distribution recorded in previous censuses between the two ethnoregional voting blocs—the North-West and the South-East—by making them almost equal. In the 2015 census, the population distribution or ratio was 56.2:43.8 in favour of the North-West. The 2021 census recorded instead a ratio of 50.01:49.99 between the two blocs. This new ratio is also at odds with the lopsided distribution of 54.1:45.9 recorded by the 2004 census carried out by the Sierra Leone People’s Party (SLPP) government of Ahmad Tejan-Kabbah (Bangura 2022a, 2022b).

Second, the 2021 census massively reduced the population of Freetown, the capital city, which has experienced high levels of inward migration over the years. It recorded a figure of 606,701 as the city’s population whereas the 2015 census recorded 1,055,964 people. The 2021 census figure for Freetown is even lower than the figure of 772,873 recorded in the 2004 census. Other datasets, such as those for primary and secondary school enrolment, assessable properties (formal houses) recorded by the Freetown City Council, and slum settlements, also cast the accuracy of the 2021 census data in doubt. For instance, there were 413,407 students aged between six and eighteen years enrolled in Freetown’s primary and secondary schools in 2020 (Ministry of Basic and Senior Secondary Education, 2021). But the majority population is constituted by children below six years and those aged eighteen and above.

In her letter to the Statistician General, Osman Sankoh, challenging the census results, the Mayor of Freetown, Yvonne Aki-Sawyerr, highlighted the number of assessable properties in the Council’s database, which in 2022 was 107,526 (The Sierra Leone Telegraph, 10 June 2022). With a recorded average of 9.2 people per household, she concluded that 989,239 people could be living in formal, assessable properties in the city. The data for 56 slum communities in Freetown show that there were 344,147 slum dwellers in 2022. The combined figures for people living in formal housing and those in slum dwellings produce a population of 1,333,386 for Freetown (The Sierra Leone Telegraph, 10 June 2022).

Third, the 2021 census reported that there are more people in Western Area Rural district (662,056) than in Western Area Urban (606,701), hereafter Freetown. Anyone familiar with the settlement patterns in the two districts would agree that the data does not make sense. There were 1,055,964 people in Freetown and 444,270 people in Western Area Rural in the 2015 census. In other words, Freetown had two and half times more people than Western Area Rural.

The fourth point concerns the elevated population numbers of Kenema District in the 2021 census, which make it the most populated district in the country—even more populated than Freetown, which in the 1985, 2004 and 2015 censuses had more people than any other district. In the 2021 census, Kenema is allocated 772,472 people and Freetown 606,701. Indeed, three districts in the East and South (Kenema, Kono and Bo) each recorded more people than Freetown. As we have seen, in the 2015 census the population of Freetown was 1,055,964; that of Kenema was 526,379. In the 2004 census, Freetown accounted for 15.76 per cent of the country’s population, and Kenema 10 per cent.

These anomalies indicate that there were strong grounds for the rejection of the 2021 midterm census by large sections of the informed public. When confronted with these irregularities, the head of the statistics agency made the laughable claim that people migrated from Freetown on the night of the census and were counted in their hometowns and villages (Hashim 2021). A migration of 449,263 people, or 42.5 per cent of the population, in one night must be the biggest migration in per capita terms in the world outside conflict zones. It is doubtful whether Sierra Leone even has the logistics or required number of commercial vehicles to perform such a feat.

The voter registration figures for the 2023 elections released in December 2022 should settle the controversy over the accuracy of the midterm census data. The distribution of registered voters for the two regional voting blocs (North-West and South-East) mirrors those of the 2018 and 2012 voter registers and gives the lie to the recorded parity of the 2021 census. The North-West accounts for 59.19 per cent of the registered voters for the 2023 elections, and the South-East 40.81 per cent. Amazingly, this 59.19:40.81 ratio is almost the same as the 59.35:40.65 ratio between the two voting blocs in the 2012 biometric voter register, which I described as lopsided bipolarity when analysing the relative voter inflexibility in the two regions (Bangura 2015). The ratio in the 2018 voter register was 60.46:39.54.

So, we now have registered voter data for three elections generated under two different governments producing similar outcomes. The parity in population distribution between the two regional voting blocs recorded by the 2021 census data therefore should be seen as spurious unless we assume that many people in the South-East refused to be registered for the elections. However, non-registration cannot be a strong argument since the South and East are the putative strongholds of the ruling party, which would have been highly incentivised to get people in those regions to register.

We encounter a more serious problem when we compare the number of registered voters in Freetown and the recorded population of the city in the 2021 census. Figure 1 compares the census data and voter registration data for Western Area Urban (Freetown), Western Area Rural, Kenema and Pujehun.


Figure 1: A comparison of the 2021 population census data and the 2023 voter registration data for Western Urban (Freetown), Western Rural, Kenema and Pujehun

Source: Statistics Sierra Leone (2022) and Electoral Commission for Sierra Leone (2022)

The number of registered voters for the 2023 elections in Freetown is 598,022, and the recorded population in the census data is 606,701. This means that almost everyone (98.56 per cent) in Freetown—including babies—registered to vote, if we are to believe the census data. About 50 per cent of Sierra Leoneans are below the voting age of 18. A registered voter population of 598,022 should equate to a population of about 1,200,000 in the city—almost the same number as Aki-Sawyerr’s estimates.

The 2023 voter register shows also that there are more voters in Western Urban than in Western Rural. The number of registered voters in Freetown is 598,022 compared to Western Rural’s voter population of 341,764. These figures are at odds with those of the 2021 census, which rank Western Rural as the more populated district. There is no reason to believe that people in Western Urban were more keen than those in Western Rural to register for the elections. The registration data for all previous elections reveals that more people registered and voted in Freetown than in Western Rural.

A similar problem is encountered when we compare the registered voters in Kenema District and Freetown. Despite the higher number of people recorded for Kenema in the census data (772,472 for Kenema and 606,701 for Freetown), a lot more people registered to vote in the 2023 elections in Freetown (598,022) than in Kenema (317,978). Indeed, as in previous voter registers, there are more voters in Freetown than any other district. Only 41.16 per cent of the 2021 census population of Kenema registered to vote. If we assume that the number of people in Kenema below the age of 18 (who are not qualified to vote) does not deviate much from the national average of about 50 per cent, and the ruling party is highly incentivised to mobilise voters in Kenema to register, we should have a much higher percentage of voters per population (about 50 per cent) than the 41.16 per cent who were registered.

The ratio of registered voters to recorded population in the census data seems bizarre in a number of districts in the South-East and North. Pujehun stands out in this regard in recording the lowest ratio in all sixteen electoral districts. According to the 2021 census, Pujehun has a population of 429,574 but only 115,789 registered voters for the 2023 elections. This means only 26.95 per cent of the people in Pujehun registered to vote. It is doubtful that 73.05 per cent of the people in Pujehun are below the age of 18. Either the census data is terribly wrong or many people refused to register. Again, this is a stronghold of the ruling party, which should be highly incentivised to get people in that district to register. It is clear from this four-case comparison that the 2021 midterm census was money down the drain. Its data is woefully unusable. 

Implications for the 2023 Presidential Election
The lopsided ratio of registered voters between the two regional voting blocs has implications for the presidential election scheduled for 24 June. Apart from the presidential election of 2002, in which the Sierra Leone People’s Party’s (SLPP) Ahmad Tejan-Kabbah won 70 per cent of the votes, a bipolar ethnoregional cleavage has defined Sierra Leone’s electoral politics since the reintroduction of multiparty politics in 1996. In the run-off presidential election of 2018, 67.3 per cent of Maada Bio’s SLPP votes were from the South and East and 89.2 per cent of Samura Kamara’s All People’s Congress (APC) votes were from the North and Western Area. This ethnoregional bifurcation explains why census and registered voter data are always closely watched as they give some indication of how parties are likely to fare in presidential elections.

If the ratio of 50.01:49.99 recorded in the 2021 census was repeated in the voter register, this would massively boost the chances of the incumbent president, Maada Bio, to win the 2023 presidential election. But, the registered voter ratio of 59.19:40.81 between the two voting blocs sends a negative signal to Bio and his ruling party and elevates the main opposition APC’s chances of returning to power.

However, a favourable voter register does not necessarily guarantee electoral success, as can be seen from the defeat of the APC in 2018. In that election, the APC lost ground in all sixteen electoral districts. Significantly, its vote share declined in the Western Area, from 72 per cent in 2012 to 60.5 per cent in 2018, and in the North, from 88 per cent in 2012 to 82 per cent in 2018. The National Grand Coalition (NGC), which drew 60 per cent of its support from the North (88 per cent of its votes were from the North-West), contributed to the six per cent decline of the APC’s share of votes in that region. The APC’s vote share in Kono in the East also declined, from 58 per cent in 2012 to 27.4 per cent in 2018.

In ‘The Humbling of the All People’s Congress: Understanding Sierra Leone’s March 2018 Presidential Run-off Election’ (Bangura 2018, 2021), I characterised the APC’s failure to retain power in 2018 as a protest vote against the party. I traced it to the party’s two-region strategy (fielding its standard-bearer and running-mate from the same voting bloc—North and Western Area) and poor record in office, especially during its second term when the key economic indicators sharply deteriorated and its leader, Ernest Koroma, became an all-powerful and unaccountable president.

Bio, on the other hand, won the 2018 presidential election on the strength of a four-region strategy. In the South and East he maximised his vote share in the six Mende-speaking districts of Bo, Moyamba, Bonthe, Pujehun, Kenema and Kailahun to stratospheric levels (securing an average of 89 per cent of the votes in those districts). He flipped Kono in the East (which was facilitated by the APC’s expulsion of Sam Sumana, who hails from that district, from the party and as vice president of the country). He made substantial inroads in the Western Area, where he increased his vote share from 25 per cent in 2012 to 39.5 per cent in 2018. And he almost tripled his vote share in the North, albeit from a low base—from six per cent in 2012 to 17.8 per cent in 2018.

The two alternations in power that Sierra Leone experienced after the reintroduction of multiparty rule—in 2007 and 2018—occurred after an incumbent party had served two terms in office. However, this should be seen as largely fortuitous rather than ordained. Tejan-Kabbah secured a second term in 2002 with a 70 per cent majority because of the postwar dividend (the economy rebounded strongly, recording the highest-ever post-civil-war growth in 2002 after the regression of the war years) and his commendable effort in reaching out to the other half of the ethnoregional divide, including in his ministerial appointments. He even secured one-third of the votes in the North during the 2002 presidential election. Similarly, Ernest Koroma’s first tenure as president (2007–2012) coincided with a global raw materials boom, which boosted the country’s GDP, helped in the construction of roads and electricity supply, and gave the impression of progress. He was rewarded with a 57.8 per cent majority in the first round of the 2012 election.

Bio, we should recall, won the 2018 election by a margin of only 3.6 per cent, or 92,235 votes. A vote swing of only 1.8 per cent or 46,118 votes will cause him to lose the 2023 election. In the forthcoming election, he is confronted with serious headwinds, chief among which is the poor state of the economy. Inflation currently stands at 26.81 per cent (compared to 16.03 per cent in 2018), and food inflation is at a record post-civil-war high—registering an astonishing 43 per cent in the fourth quarter of 2022. The national currency has depreciated against the dollar almost three-fold, from SLE 7,664 in 2018 to SLE 21,632 in 2023, and the GDP has grown at an average rate of only 2.8 per cent between 2018 and 2022—not enough to generate meaningful jobs for the large number of unemployed or underemployed youths trapped in poverty. We do not have poverty data for the period 2018–2023 (the World Bank’s most recent value for poverty in Sierra Leone is for 2018). However, a study conducted jointly by the World Food Programme and Sierra Leone’s Ministry of Agriculture in August 2022 found that 81 per cent of households were food insecure and 14.9 per cent were extremely food insecure (World Food Programme, 2022).

A popular refrain in Freetown on the high cost of living during Bio’s first two years in office was ‘di gron dry’ (literally ‘the ground is dry’—or times are hard). The new words on the streets are ‘sufferness’ and ‘sufferation’ (or suffering). On the plus side, the government has spent substantial sums of money on its flagship Free Quality School Education programme (accounting for 22 per cent of the budget), which has raised enrolment levels tremendously across the country. However, parents are still responsible for about 25 per cent of the resources or finances received by schools, in the form of levies. Non-fee expenditure, such as for non-basic textbooks, uniforms, bags, food and transportation, is still a burden on households, and many parents prefer to send their children to private schools as they do not consider state-funded, free-tuition schools good enough.

Bio’s anti-corruption drive has also run out of steam as there does not seem to be much difference between his government and the APC’s, whose corrupt officials he has hounded and tried to discredit. Corruption continues to be the bane of development. Sierra Leone gained three points in Transparency International’s Corruption Perceptions Index during the first year of Bio’s administration, moving from 30 points in 2017 and 2018 to 33 points in 2019. However, performance has stalled at 34 points since 2021. A score of 34 out of 100 is still very low.

To compound the problem, the well-respected and diligent Auditor General, Lara Taylor-Pearce, was arbitrarily suspended in November 2021, a few weeks before the release of the annual audit report for 2020. The report contained damning revelations of fraudulent practices, questionable use of public funds, and systematic failure to comply with established rules and procedures in financial transactions across government ministries and agencies. The Office of the President was implicated in these wrongdoings, which included the falsification of hotel receipts during the president’s travel to Lebanon, double-dipping by staff on the president’s travel team, inaccurate retirement receipts totalling USD 110,000 racked up by the State Chief of Protocol and the Personal Assistant to the First Lady (both of whom were asked to refund the money) stemming from the president’s overseas trips, and direct cash payments of USD 170,489 (which may expose holders of such a large amount of cash to charges of money laundering) to settle medical bills (Audit Service Sierra Leone, 2021).

Like his predecessor, Ernest Koroma, whose cabinet largely consisted of people from the North-West (81.8 per cent in 2007 and 75 per cent in 2010), Bio’s appointments to top government jobs have been highly ethnoregional, which has alienated those who do not trace their origins to the South-East. Even though he campaigned on a four-region strategy in 2018, he has governed in the last five years on a two-region ethnocratic platform. According to an Institute for Governance Reform report in 2018, Beyond Business as Usual: Looking inward to change our story, 58.6 per cent of ministerial appointments were held by individuals from the South-East, which accounts for markedly less than 50 per cent of the population.

The ethnoregional bias worsened over the next two years. A new book by Umu Tejan-Jalloh (2023), The Early Policies of President Maada Bio: My Thoughts and Analysis, shows that by 2021 the South-East accounted for 64 per cent of cabinet appointments. Shockingly, twenty-eight (or 90.3 per cent) of the thirty-one heads of parastatals and core government agencies were people from the South-East. Only one Northerner and two Westerners headed a parastatal or core government agency.

The totality of these economic problems and ethnic biases suggests that all is not well in the polity. A survey by Afrobarometer and the Institute for Governance Reform in 2020 (Afrobarometer and IGR 2020) found that only 32 per cent of Sierra Leoneans believed the country was ‘going in the right direction’—a 13-point drop from 45 per cent in 2018. Disturbingly, there was a stark divide in opinion in the two ethnoregional blocs: only 14 per cent of respondents in the North and 16 per cent of those in the West believed that the country was moving in the right direction—against 57 per cent of respondents in the South and 53 per cent of those in the East who believed the country was on the right track.

Given Bio’s small margin of victory in 2018 and the current unfavourable voter registration ratio between the two regional blocs, one would have expected a more inclusive policy in top-level appointments, similar to that of Tejan-Kabbah, in order to improve his prospects of re-election in June 2023. As I showed in my article (Bangura 2018, 2021), Bio would not have won the 2018 election without support from voters in the North. Let me quote that section of the article to illustrate the point:

Relying on the South and East would have given him only 34.85 per cent of the votes; and including the Western Area would have raised his vote share to 46.32 per cent. It is only when his votes in the North are added that he is able to get to the 50 per cent+1 mark. The interesting point about Bio’s Northern votes is that reliance on only his votes in the districts with strong minority presence (Kambia, Koinadugu, Falaba and Karene) would have given him only 2.92 extra per centage points, which would have raised his overall vote share to 49.24 per cent. He needed his votes in the predominantly Themneh-speaking districts of Port Loko, Tonkolili and Bombali (which gave him 2.57 extra per centage points) to get him across the victory line.

Bio’s governance record also is not very different from Koroma’s in terms of how key state institutions, such as the police and judiciary, serve the interests of the party in government. In 2018, just a few months after the elections, the judiciary upheld the SLPP’s petitions against some APC parliamentarians and created 10 SLPP members of parliament, who lost the election to those APC parliamentarians in constituencies that were strongly pro-APC. This produced a spurious parity in parliamentary representation between the two parties. The judiciary has also played an activist role in the affairs of the APC: slamming the party with several injunctions, dissolving its national executive, constituting an interim transitional governing committee and dictating the category of people that should be nominated into it. The most bizarre injunction was the one granted to an aggrieved member of the party on the eve of the party’s convention to elect its standard-bearer on 17 February 2023, after the court had previously given the greenlight for the convention. There were fears that the APC would not be allowed to field a candidate for the presidential election. Western donors met the judiciary, the electoral commission and representatives of the government in an Election Steering Committee meeting on 17 February (EU in Sierra Leone, 2023). The judge reversed her decision on the same day after the Election Steering Committee meeting and the convention was held on 18 February.

Basic freedoms, such as the rights of free speech, assembly and movement are still fairly respected. Bio’s government repealed the obnoxious criminal libel and seditious law (Part V of the Public Order Act of 1965) in 2020, which improved the country’s scores on the World Press Freedom Index from 69.72 in 2020 to 70.39 in 2021 and 71.03 in 2022 and moved it up the global ranking from 85 to 75 and 46 respectively, out of 180 countries. However, like many countries that are experiencing democratic backsliding or incomplete democratisation, these freedoms are not well protected and are periodically abused. Police high-handedness and targeted harassment or assault of journalists continue—such as against Standard Times journalist, Fayia Amara Fayia (who was beaten up by soldiers in Kenema); Salieu Tejan Jalloh, editor of The Times newspaper (who was forced to flee the country after being charged with defamation for an unpublished story); and the editor of the US-based Africanist Press, Chernoh Alpha Bah, who has received death threats (CPJ, 2022) for his hard-hitting investigative work on corruption by government officials in Sierra Leone.

The Media Foundation for West Africa (MFWA 2022), a press freedom watchdog, recently concluded that ‘the state of press freedom remains an admixture of the good, the bad and the ugly’. Arbitrary arrests, refusal of permits for public demonstrations, detentions, unlawful killings, and use of excessive force against civilians and those considered to be anti-government activists persist. Many of these infractions and abuses are documented in the US State Department’s 2022 Country Reports on Human Rights Practices (US State Department 2022) and Amnesty International’s 2023 report on excessive use of force by Sierra Leone’s security forces in 2022.

These shortcomings in human rights or freedoms may explain why Sierra Leone and other countries with similar experiences are not listed as democracies in The Economist Intelligence Unit’s Democracy Index (EIU 2022). These countries are classified instead as ‘hybrid regimes’—one level above the category of ‘authoritarian regimes’ but two levels below the categories of ‘full democracies’ and ‘flawed democracies’. Sierra Leone is also listed as ‘partly free’ in Freedom House’s Freedom in the World Index (Freedom House 2022). Of course, these global indexes are not perfect because many of the issues they measure and the categories for ranking countries require subjective evaluation. However, they are widely consulted by the general public and exert pressure on states to improve governance practices.

A worrying development is the use of violence by the two main parties in pursuing political objectives. The Armed Conflict Location and Event Data Project (ACLED 2020) shows that the number of violent incidents has risen dramatically—from an average of two events a month before 2017 to twenty in the election year of 2018, and has stayed at a very high level. The spike in violence was expected to dip after the election, as happens in many countries. However, between September 2018 and 2020, there were ten events of political violence every month, with occasional peaks of almost thirty events. Clashes between parties, internal party violence, and targeted violence against civilians or political party agents and supporters have all increased in recent years. It is not surprising that Sierra Leone dropped fifteen places in the Global Peace Index, from thirty-five in 2018 to fifty in 2022. Fears of an upsurge in electoral violence have been cited by the Political Parties Regulation Commission (2023) as the reason for its controversial decision on 3 April 2023 to ban political rallies during the election campaign period.

Sierra Leoneans seem to be stuck with the SLPP and APC, which have governed the country since independence in 1961, even though neither of these parties has been able to move the needle on development and improve the lives of citizens in any substantive way. Unfortunately for voters, the party that promised an alternative, the NGC, is in its death throes after its leader, Kandeh Yumkella, called for a ‘strategic partnership’ with the ruling party and struck a so-called ‘progressive alliance’ with it on 14 April 2023. Even well before its fracture in January 2023, the NGC had become a poor shadow of its 2018 version that secured 6.86 per cent of the national vote and four parliamentary seats in Kambia. It may well end up as a small ethnic minority party if it contests the June elections separately or in alliance with the SLPP.

The merger or alliance of the NGC and SLPP reminds me of the joke during the collapse of the Soviet Union in the early 1990s, that ‘socialism is the longest route to capitalism’. It is now clear that for all its talk about offering Sierra Leoneans an alternative governance and development path, the NGC is, essentially, the longest route to membership of the SLPP.

The demise of the NGC as an electoral force will make it easier for one of the two main parties to win the June presidential election in the first round of voting. The other minor parties, such as the Coalition for Change, the Alliance Democratic Party, the People’s Movement for Democratic Change, the Revolutionary United Front Party, the United National People’s Party, the Citizens Democratic Party, the United Democratic Movement and the Unity Party, which accounted for 7.2 per cent of the votes in the 2018 presidential election, seem to have fizzled out or lack traction. If the APC wins the election, it will not be because it has put up a robust opposition in the last five years or advanced credible policies that will transform the economy or improve the governance regime. The 2023 presidential election is instead likely to be a referendum on Bio’s economic development and governance record—just as Bio’s victory in 2018 was a referendum on Koroma’s record.

The instinct for winner-takes-all outcomes and lack of a civic culture in terms of how parties behave in and out of office are likely to make the forthcoming elections perilous. Leaders and supporters of the incumbent party, who talk about an existential threat if they lose the election, may seek to cling to power; those of the main opposition party, who are equally determined to win back power, complain about targeted harassment and raise the spectre of revenge if they get back into office. This kind of atmosphere is unlikely to guarantee free and fair elections. There is a real danger of voter suppression, targeted violence and falsification of results. When will Sierra Leone get its act together and rise above the politics of ‘yu du mi, ar du yu’, or tit for tat?