News Headlines
Home » Gen. News » Where did President Mahama’s Votes go in 2016?: Voter Apathy vrs Vote-Rigging
The Writer :J. Atsu Amegashie

Where did President Mahama’s Votes go in 2016?: Voter Apathy vrs Vote-Rigging

Download the full pdf version by clicking on this link!

By: J. Atsu Amegashie
December 19, 2016
Allegations of vote-rigging abound in Ghana’s presidential elections. Even in victory, a member of the NPP, Hackman Owusu Agyemang, alleged that the NDC rigged the 2016 election and the NPP’s presidential election campaign manager, Peter Mac Manu, revisited allegations of vote-rigging in the 2012 election. The NDC’s Afotey Agbo and Koku Anyidoho have also alleged that the 2016 presidential election was rigged. In response to the argument that Mahama’s loss of votes in 2016 supports the allegation that the 2012 election was rigged, I undertook an examination of the 2012 and 2016 presidential elections. Voter apathy partly accounted for president John Mahama’s poor performance in the 2016 presidential election. In five regions (Brong Ahafo, Central, Northern, Upper East, and Upper West), Mahama’s losses of votes were not unusual; they were negligible (very small) or almost equal to Nana Addo’s gains. Mahama’s losses of votes in regions like Greater Accra region and Volta region were unusual, although voter apathy explains part, but not all, of his losses. Compared to turnout after a party has been in power for 8 years, the decrease in turnout of about 7 percentage points from 2012 to 2016 is not surprising. After 8 years of the NPP in power, the turnout in the first round decreased from 85.10% in 2004 to 69.52% in 2008, a decrease of 15.58 percentage points. I argue that the Electoral Commission’s conduct in 2016 gave credence to the allegation of an intent to rig the 2016 election and, in retrospect, this conduct and Mahama’s significant decrease in votes in some regions in 2016 cast some doubt on the credibility of the 2012 presidential elections. However, the NPP’s non-disclosure of its collated results in 2012 equally casts some doubt on its claim that Nana Addo won the 2012 election. In 2012, it is conceivable that, during electronic transmission of constituency collated results, there was vote inflation in Mahama’s favor. Yet the NPP’s arguably strategic non-disclosure of its collated results and inconsistent theories/claims of vote-rigging meant that Mahama’s victory in 2012 reflected the true will of Ghanaians.
Allegations of vote-rigging are very common in Ghana’s presidential elections. It seems that the only free and fair presidential elections are those won by candidates of opposition parties. When a candidate of an incumbent party wins the presidential election, the opposition always complains of vote-rigging. This happened in 1992, 1996, 2004, and 2012. Even in 2008 when the election was won by the candidate of the opposition party, John Atta Mills of the NDC, there were allegations of vote-rigging (see appendix A below). And after the NPP won the 2016 presidential election a member of the NPP, Hackman Owusu Agyemang alleged that had president Mahama not rigged the 2016 presidential election, its candidate, Nana Addo’s margin of victory would have been more than 3 million votes instead of almost 1 million votes.1
However, these allegations of vote-rigging are not without merit. In a country where politics is one of the surest paths to almost instantaneous wealth, the private sector is not very vibrant, and corruption is arguably rife, the incentive to rig elections is strong. However, there is also the tendency to instinctively allege vote-rigging, without evidence, because a candidate or party lost an election.
In the presidential election held on December 7, 2016, president John Mahama lost about 900,000 votes relative to the votes he obtained in the 2012 election. This poor and puzzling performance has been attributed to the following factors:
1                 Mahama’s economic mismanagement and numerous allegations of corruption.

2                 Voter apathy (i.e., Mahama’s supporters did not turn out to vote in 2016).

3                 Vote-rigging: in 2012, NPP agents were bribed by the NDC to certify false results at      polling stations (i.e., pink sheets)

4                 Vote-rigging: in 2012, during the electronic transmission of constituency collation sheets to the EC’s strong room, the figures were altered in favor of John Mahama.

5                 Vote-rigging: in 2012, ineligible voters (i.e., non-Ghanaians in Togo and minors) were allowed to vote. This favored John Mahama.


In this short article, I shall attempt to shed some light on some of these claims.  
1He was reported to have said that “The NPP should have stretched them by over three million votes. …The votes they got, I believe there was rigging.”:­2016-poll-should-ve-lost-by-3m-votes-Kofi-Jumah-494862
ANALYSIS Table1: regional net gains of votes between 2012 and 2016 for Mahama and Nana Addo
Region Mahama (NDC) Nana Addo (NPP) 2016 votes minus 2012 votes 2016 votes minus 2012 votes
Notes: Computed on December 12, 2016 based on 272 constituencies out of 275. The figures for the computations were obtained at The symbol * refers to constituencies for which the data are incomplete. For Eastern Region, the 2016 figures are based on 31 out of 33 constituencies and for Northern region, on 30 of 31 constituencies.
Based on 272 out of 275 constituencies in 2016 and 275 constituencies in 2012, table 1 shows that Mahama lost 891,786 votes while Nana Addo gained 364,689 votes.2 Candidates other than Mahama and Nana Addo gained 14,242 votes. The fact that Mahama lost votes in every region is consistent with dissatisfaction by voters with his performance. But it is perhaps puzzling that a significant number of these votes did not go to Nana Addo and that there appears to be (891,786 − 364,689 − 14,242) = 512,855 missing votes. This may suggest that Mahama’s votes in 2012 were inflated (vote-rigging) but is also consistent with voter apathy (i.e., some people did not to vote). In 2012, the number of votes cast (including rejected ballots) was 11,246,982. Therefore, the apparently missing 512,855 votes account for 4.56% of votes cast in 2012.3
In October 2015, the Electoral Commission set up a committee chaired by Professor V.C.R.A.C. Crabbe to hold a two-day public hearing on the Voters’ Register. In its report, the committee noted that the names of the deceased remained on the register.   The
2It is very likely that after the votes in the two outstanding constituencies in the Eastern Region are taken in account, Nana Addo’s net loss in the Eastern region will fall and indeed may become a net gain. Mahama’s net loss in the Northern region may also fall. 3Of course, I am not suggesting that the same set of voters in 2012 also voted in 2016. There were new eligible voters in 2016; some people who did not vote in 2012 may have voted in 2016 and some people who voted in 2012 may not have voted in 2016. Other voters died between December 2012 and December 2016. However, we do not know the identities of voters in 2012 and 2016. We can only rely on aggregate data. This does not weaken the analysis here. The claims about vote-rigging or voter apathy are based on trends in aggregate numbers.
committee estimated that as many as 600,000 voter deaths would have occurred cumulatively between the last election (in December 2012) and 2016. The 2016 register was bloated by at least 600,000. Therefore, although the EC claimed to have registered about a million more voters since the 2012 election, the number of voters in the 2012 and 2016 registers may have been close.4
In the Northern and Upper East regions, the votes lost by Mahama were negligble (less than 2000 in each region) and in the Upper West, Mahama’s loss was almost equal to Nana Addo’s gain. The absence of a one-for-one relationship or a zero-sum relationship between Mahama’s loss of votes and Nana Addo’s gains or gains by other candidates is not necessarily evidence of vote-rigging. Therefore, the results in the cases of Central region and Brong-Ahafo region need not be puzzling. The loss in votes in, for example, the Volta Region and Greater Accra Region may be puzzling. I return to these regions in subsequent analysis.
Why were Mahama’s votes not inflated in the NDC’s strongholds of Upper East, Upper West, and Northern regions? One would have thought that to make vote-rigging difficult to detect, the inflation would have been spread over the NDC’s strongholds instead of concentrating all of it on a stronghold like the Volta Region.
Rigging at polling stations and also during electronic transmission of results does not make sense. If the NDC and EC wanted to rig the election by altering figures during transmission, they would not also have rigged the votes at the polling stations. And if they had rigged the election at polling stations by altering “Statement of Poll and Declaration of Results” Forms (i.e., pink sheets, there would have been no need to also alter the votes during electronic transmission because the constituency collation sheets being transmitted would have been a collation of figures already altered on pink sheets. In 2012, rigging during transmission made more sense because the parties did not have copies of the collation sheets that were transmitted to the EC’s head office in Accra. They had copies of the pink sheets but not the collation sheets.5 But a puzzle remains. Were there no NPP representatives in the strong room of the EC (national collation center) to verify collated constituency results?6 In any case, one of the NPP’s major rigging theories was that the votes were altered (by an Israeli firm) during electronic transmission.
Even if the votes were altered during transmission, the NPP could still have collated the figures on the pink sheets given to its agents at polling stations — just as it collated pink sheet votes in 2016 — and told Ghanaians if Nana Addo had won the election. Four months after the election when the NPP presented its case in the Supreme Court, the party did not disclose its collated results. Its collated results for all 26,002 polling stations were never disclosed. This suggests that the NPP lost the 2012 presidential election because had its collation of votes on the pink sheets
4Following a Supreme Court order in May 2016, the EC deleted the names of 56,772 persons who registered as voters with National Health Insurance cards. According to the EC, 29,550 of the affected people re-registered.5A constituency is made up of polling stations. Pink sheets have a record of the votes cast in a polling station and copies of these sheets were given representatives of candidates at polling stations. A constituency collation sheet aggregates or tallies the votes cast in ALL the polling stations in a constituency.
6Representatives of candidates are allowed in the EC’s strong to endorse results. For example, watch the 2011 documentary titled “An African Election”. This was about the 2008 presidential election:
in 2012 given a favorable outcome, as was the case in 2016, it would have announced its collated figures to Ghanaians. That’s exactly what the NPP did on December 8, 2016, a day after the 2016 presidential election.
Instead, in 2012, the NPP selected 11,842 pink sheets out of the 26,002 pink sheets and argued that if the votes on those pink sheets were annulled and subtracted from the EC’s collated figures, Nana Addo would be the winner.7 This meant that about 14,160 pink sheets, in their judgement, had no electoral irregularities. On the basis of only those implicitly self-selected pink sheets, the NPP argued in the Supreme Court that Nana Addo won election and should be declared the president-elect. This is a classic case of what is known in statistics as self-selection bias, selection bias, or selectivity bias.
(a) Survey/Anecdotal evidence of voter apathy
There is some evidence to support the claim that Mahama lost votes due to voter apathy. On election day December 7, 2017, JoyNews reported that:
“There is low turnout in the Anlo constituency in the Volta Region as some residents protest poor living conditions.
Joy News’ Ivy Setordzie said most of the residents in the constituency refused to take part in the ongoing presidential and parliamentary elections.
As at 5:00 pm only 260 out of 447 voters had cast their ballots, at Fuveme, Setordzie reported.”:­anlo-polling-station.php
Note that the Anlo Constituency is not close to the Ghana-Togo border and the JoyNews reporter spoke to actual voters (not ghost voters) – most residents in constituency – who refused to vote. The turnout in Fuveme (in the Anlo constituency) was 260/447 = 58%. In the Anlo constituency, valid votes cast in 2016 was 34,601. I do have not the figure for the number of registered voters in 2016 in the Anlo constituency. But it was 50,836 in 2012. Given that the number of registered voters in 2016 is likely to be higher than the number in 2012 and using a conservative increase of 10%, it follows that the turnout in the Anlo constituency was not more than 61.87%, close to the overall turnout of 61% in the Volta Region.
Prior to the 2016 election, residents and chiefs of the Volta region expressed anger over the NDC’s neglect of the region. Here are some examples:
7This was in Dr. Bawumia’s affidavit to the Supreme Court in April 2013, four months after the December 2012 election. At a press conference in December 2012, Dr. Bawumia said the NPP had examined more than 24,000 pink sheets.
December 12, 2014: “Chiefs in the Volta Region have expressed anger over government’s failure to fix the deplorable roads in the Region. Residents in the region on Thursday embarked on a demonstration to express their grievance against the Mahama-led administration for “neglecting the region.” According to the chiefs, the government had on numerous occasions failed to fulfill its promises made to the region, especially with issues regarding the poor nature of roads in the region.”:­mahama/#sthash.oZagjFU4.dpuf
April 15, 2016: “Fisher folk in the Ketu South Constituency of the Volta Region, have threatened to vote out the incumbent National Democratic Congress (NDC), come November 7 2016, if the party fails to fulfill its numerous promises to improve their livelihood. …. The general concerns expressed by the fisher folk were echoed by one Delali Tetevi, a fisherman from the Ketu South District, who … warned that if their concerns are not met, then there would be no motivation to vote for a party that will not meet their needs. … Many of the fisher folk who spoke to Citi News’ King Norbert Akpablie, expressed similar sentiments saying promises made over the years including the construction of a fishing Harbour have all not been fulfilled.”:­if/#sthash.OHFBsFla.dpuf
September 13, 2016: “The Agbogbomefia of Asogli and President of Volta Regional House of Chiefs, Togbe Afede XIV has bemoaned the sore state of the Volta region as far as developmental projects are concerned. According to Togbe Afede, it is sad that the region has been consistent and kept faith with the NDC since the inception of the fourth republic but there is little to see in terms of development.”:­togbe-afede/22913
The voter apathy hypothesis, at least, in the Volta Region is not without merit. Table 2 below shows that John Mahama lost votes in all 26 constituencies in the Volta Region.
(b) Comparing turnout in constituencies near to and far from the Ghana-Togo border
According to an article in Newsweek: “… in the Ho Central constituency in the Volta region, considered the ‘world bank’ of the governing party on account of the massive votes that it garners in the region, the president won 53,100 votes on Wednesday compared with 62,400 in 2012.
This decline in turnout was replicated in other areas: in polling stations near Ho, turnout dropped by anything from 5 percent to 18 percent as compared to the 2012 elections.”:
This was also true in constituencies like Ketu North and Ketu South in the Aflao district (a district which is located very close to the Ghana-Togo border) and in constituencies like Ho and Anlo that are not close the Ghana-Togo border. Aflao is about 110 kilometers from Ho.
In Ketu-South constituency (close to the Ghana-Togo border), Mahama’s votes decreased by (81,880 – 50,648) = 31,232, a decrease of 38% and in Ho-Central constituency (farther from Ghana-Togo border), it decreased by (62,363 – 24,363) = 38,000, a decrease of 61% which is 23 percentage points more than the decrease in Ketu-South.
There is, however, evidence that some Togolese (who may not be dual citizens of Ghana and Togo) cross the border to vote in Ghana. The NDC does not deny the claim that people cross the border from Togo to vote in Ghana’s elections. In fact, it claims that these voters are Ghanaians who live in Togo. In August 2015, the NPP disclosed that by comparing the Togolese and Ghanaian voters’ registers, it had discovered about 76,000 Togolese whose names were in both register.8 This number is not big enough9 to have a significant effect on the number of registered voters in all 26 constituencies of the Volta Region. In fact, the NPP claimed that these Togolese voters were concentrated in the Ketu-South constituency. On the basis of the NPP’s figure of 76,000; a comparison of turnouts in Ketu-South and Ho-Central; and JoyNews’ Ivy Setordzie report in the preceding subsection, it appears that the voter-apathy effect was strong in the Volta Region.10 Of course, the conclusion that the voter-apathy effect was strong in 2016 does not necessarily imply that there was no vote inflation in 2012.
(c) Comparing performance of NDC parliamentary candidates in 2012 and 2016  
Altering parliamentary votes during electronic transmission was not necessary in 2012 because the MP-elect (winner) of a constituency was — as is always the case — declared at the EC’s constituency head office before the results, if at all, were transmitted to the EC’s head office in Accra. No significant objections were raised in the 2012 parliamentary elections. No result was challenged in court. Therefore, it reasonable to assume that even the NPP believed that the 2012 parliamentary elections were fair. In general, parliamentary elections are much less contentious compared to presidential elections. In fact, the NPP used the difference between the votes obtained by Mahama and the votes obtained by the NDC’s parliamentary candidates in 2012 as one of the grounds for alleging that the 2012 presidential election was not fair.
The NDC lost almost 50 parliamentary seats in the 2016 election. In the Greater Accra region, the NDC’s parliamentary candidates lost 140,850 votes in 2016 while Mahama lost 213,535 votes, a difference of 72,685 votes. In the Volta region, the NDC’s parliamentary candidates lost 109,095 votes in 2016 while Mahama lost 144,543 votes, a difference of 35,448 votes.11 These differences in votes are about 4.0% and 3.4% of total votes cast in the 2012 elections in the Volta region and Greater Accra region respectively. This need not be puzzling because, as president, voters were likely to hold John Mahama more accountable than NDC’s MPs for the poor
8­voter-roll-375993 9 However, it is possible that some Togolese on Ghana’s register were not detected because they were not registered voters in Togo. I do not expect a significant number of voters in this category.10 Of course, foreigners should not vote in any country’s presidential election and in a close election every vote counts. John Atta Mills won the 2008 presidential election with about 40,000 votes.11For Greater Accra region, I excluded Ayawaso North constituency and for Volta region, I excluded Adaklu constituency because data for the 2016 parliamentary elections in these constituencies were not available.
performance of the economy. A president has a higher level of responsibility and commensurate powers than an MP, and so his/her electoral gains and electoral losses are likely to be bigger than the corresponding losses and gains for an MP.
An MP may be a stronger parliamentary candidate in his/her constituency than the presidential candidate of his/her party. In 2016, the NPP’s parliamentary candidates in Ablekuma North and Dome Kwabenya constituencies (in Accra) got 32,123 votes more than Nana Addo. In 2004, the number of votes cast in the parliamentary elections was 417,161 more than the number of votes cast in the presidential election. And in 2012, the number of votes cast in the presidential election was 20,630 more than the number of votes cast in the parliamentary elections. It is tempting to conclude that there are anomalies in some of these statistics. But this may be a false conclusion.
The bottom line is the NDC’s poor performance in 2016 was not limited to president Mahama. Its parliamentary candidates also performed poorly.
(d) Comparing Kufuor (NPP presidential candidate) in 2004 to Nana Addo (NPP presidential candidate) in 2008
In 2004 John Kufuor, the NPP’s presidential candidate, got 4,524,074 votes in the first round; in 2008, Nana Addo got 4,204,073 votes in the first round, 320,000 votes less than what Kufuor got in 2004. This was a 7% reduction in votes obtained by the NPP’s presidential candidate. Yes, the candidates were different but, as in 2016, the 2008 presidential election was a protest against the corruption and mismanagement of the incumbent party. Nana Addo, as the NPP’s presidential candidate, ran on Kufuor’s record.
Mahama’s record in terms of managing the economy was worse than Kufour’s because Kufuor introduced social intervention programs and managed them better (relative to Mahama): school feeding program, metro buses, NHIS, etc. Thus, it may not be surprising that in the 2016 presidential election, Mahama lost more votes in both absolute and relative (15% decrease) terms than Nana Addo (7% decrease) in the 2008 presidential election.
In 2004, the turnout in the first round was 85.10%. Kufuor was re-elected in the first round. In 2008, the turnout in the first round was 69.52%.   Note that the turnout in 2012 was about 7 percentage points higher than the turnout in 2016 while the turnout in 2004 was 15.8 percentage points higher than the turnout in 2008.12 Is the significantly lower turnout in 2008 necessarily an indication that Kufuor rigged the election in 2004? Not necessarily.
12 The Electoral Commission reported a turnout of 68.62%. However, adjusting the 2016 voters’ register by taking into account the estimated number of 600,000 deceased voters, the turnout was 72.4%. The turnout for 2012 was 79.4%. The turnout in 2016 of 72.4% is based on 272 constituencies. When the total number of votes cast in the three outstanding constituencies are taken into account, the turnout will be higher. In either case (2004-2008 and 2012-2016), a statistical test shows that the difference in turnout was significant. But this shed no light on the vote-rigging versus voter apathy theories. Statistical significance in this case only means that the populations of registered voters were different in terms of their motivation to vote. It says nothing about the reasons for these differences and the test is not applicable if the population of registered voters is not accurate (i.e., bloated register).
(e) Inconsistent claims of vote-rigging in 2012
December 12, 2016: “New Patriotic Party (NPP) campaign manager Peter Mac Manu has explained his party’s triumph in the 2016 general elections was down to a vigilant campaign to ensure, ‘foreigners’ do not get the chance to vote. … He believes that if these illegal groups were to participate in the December 7 general elections, they could hurt the party’s chances as it happened in 2012 when the party lost to the National Democratic Congress (NDC).”13
In this case, the implication is that the main channel for vote-rigging in 2012 was not through the manipulation of figures during electronic transmission.  
Prior to the announcement of the 2012 presidential election results by the EC, the NPP’s General Secretary, Kwadwo Owusu Afriyie (aka Sir John) announced at a press conference on Saturday Dec 8, 2012 that the NPP had won the election with 51.3% of the total valid votes cast for Nana Addo and about 45% for John Mahama. Then six days later, December 14, 2012, Kwadwo Owusu Afriyie said that “… the Electoral Commission (EC) rigged an excess of one million votes for John Mahama.”
Note that he was a very senior official of the NPP (its general secretary). He did not make a general claim of vote-rigging. He specifically stated, on behalf of the NPP, that Nana Addo won the election by 51.3%, John Mahama got about 45%; and that EC had rigged an excess of one million votes in favor of John Mahama. He could only have known this information on the basis of the votes on pink sheets given to the NPP’s agents at the 26,002 polling stations. In paragraph 79 of his April 2013 affidavit at the Supreme Court, the NPP’s 2012 vice-presidential candidate, Dr. Bawumia swore that “… in some instances, votes obtained by the 1st Petitioner (Nana Addo) were unlawfully reduced, whilst at the same time votes of the 1st Respondent (John Mahama) were illegally padded with the sole purpose of procuring the victory of the 1st Respondent in the 2012 December Presidential Elections.”
On December 14, 2012, I undertook an algebraic analysis of Sir John’s two statements (i.e., the statements on December 8, 2012 and December 14, 2015). It clearly showed that they were inconsistent with each other. Therefore, one of Sir John’s statements was false or both were false. The full analysis is in appendix A below.
To be fair to the NPP, the NDC is also guilty of making such baseless claims. The NDC’s director of elections, Samuel Ofosu Ampofu, and its campaign spokesperson, Joyce Bawa-Mogtari, made similar but less specific claims of victory a day after the 2016 presidential election. At a press conference, Ofosu Ampofu said the NDC was winning while Joyce Bawa-Mogtari made the same claim in a press release. In retrospect, given Nana Addo’s impressive performance, it appears the NDC knew that Mahama had lost the election when its officials claimed victory. The NDC also alleged, without proof, that there were numerous cases of over­
voting, especially in the Ashanti Region, during the 2016 elections. This allegation was made by Nii Laryea Afotey Agbo, Greater Accra Regional Minister and MP for Kpone Katamanso, and Koku Anyidoho, the NDC’s deputy general secretary.
(f) The EC’s conduct and suspicions/allegations of voting rigging
In the 2012 elections, the EC gave copies of pink sheets to representatives of parliamentary and presidential candidates and political parties but not constituency collation sheets. In presidential elections, constituency collation sheets, not pink sheets, are transmitted to the EC’s national collation center in Accra where the results are tallied and, on the basis of the tallied votes, the winner of a presidential election is declared. An Electoral Reforms Committee set up by the EC recommended, among others, that the EC should give copies of constituency collation sheets to representatives of political parties and candidates. The EC accepted 27 of the committee’s 29 recommendations (including giving copies of constituency collation sheets to parties and their agents) and stated that “… the implementation of this comprehensive list of reforms will make the outcome of the 2016 elections credible and acceptable to all the stakeholders involved.”14
However, when the EC prepared the Public Elections Regulations 2016 (C.I. 94) and submitted it to parliament it included a regulation that mandated the commission to give only the constituency collation sheets for parliamentary elections to parties/agents but not the constituency collation sheets for presidential elections. It took the astuteness of a former Director of Research at Parliament, Kwesi Nyame Tease-Eshun, to expose this anomaly in C.I. 94. He petitioned the Supreme Court and in October 2016, the Court ordered the EC to give constituency collation sheets (for presidential elections) to agents of presidential candidates.
Given this lack of transparency by the EC and the need to coerce it to be transparent, the suspicion of an intent by the EC to rig the presidential 2016 election was reasonable.
Voter apathy is likely to have played a role in president John Mahama’s poor performance in 2016 election. Voter apathy is not independent of the argument that Mahama’s poor economic management and allegations of corruption accounted for his poor electoral performance. An incumbent’s mismanagement could lead to voter apathy especially in his stronghold where voters do not want to vote for him but will not also vote for any of his rivals.
Mahama’s losses of votes in five regions (i.e., Central, Brong Ahafo, Northern, Upper East, and Upper West) do not appear to raise any red flags. In these regions, his losses were either negligible or almost equal to Nana Addo’s gains. His losses in the Volta Region and Greater Accra Region were unusual. I have rationalized his losses in these regions on the basis of voter
14Retrieved on December 17, 2016:­
apathy and economic mismanagement. However, I cannot definitively rule out vote inflation in Mahama’s favor in the 2012 election. As argued above, the EC’s conduct in 2016 did not engender trust.
Yet given the NPP’s non-disclosure of its collated figures in 2012 and its inconsistent theories of vote-rigging, it is conceivable that even if Mahama’s votes were inflated, this did not affect the outcome of the election because there is reasonable cause to believe that the NPP’s own collation of votes on the pink sheets showed that John Mahama won the election and this was why the party did not reveal its collated figures in 2012 (but did so in 2016). In other words, if rigging occurred by artificially inflating Mahama’s votes, it may not have had a material effect.15
To the NPP’s credit, it has set a standard for Ghana’s election. Based on its action in 2016, the standard is that a claim of electoral victory must be based on an announcement/disclosure of the results of a candidate’s or party’s collation of the figures on pink sheets. Where there is a discrepancy between a candidate’s collated figures and the EC’s collated figures, it will be a matter of checking the authenticity of pink sheets and adding up the numbers. For example, the NPP intends to challenge the 2016 parliamentary election results in five constituencies in the Western region because its collated results are different from the EC’s declared results. “According to the Regional secretary of the NPP, Charles Bissue, the party was certain it had won in the constituencies based on its collated results, but the official results declared by the Electoral Commission (EC) subsequently proved otherwise.”:
15Or the NPP’s collated figures for all 26,002 polling stations showed that Mahama and Nana Addo each got less than 50% of the vote. But the NPP suppressed this information because it did not want a run-off. This was not likely because the NPP left room for the possibility of a run-off based on the annulment of votes. In his 2013 affidavit on behalf of the NPP, Dr. Bawumia stated that “… if the only violations complained of were absence of the signatures of the presiding officers in the 1,826 polling stations and the votes in these polling stations were annulled, the 1st Respondent, John Dramani Mahama, would have obtained 49.45% of the valid votes cast, and the 1st Petitioner, Nana Addo Dankwa Akufo-Addo would have obtained 49.03% of the valid votes cast.” Similar scenarios were presented for other alleged irregularities.
Note: The total number votes reported at for Mahama (in 2016) is higher than what is in the table above. This is because, I think that the votes for Mahama in Ho-West constituency is incorrect. recorded 53,117 votes for Mahama in Ho-West constituency in the Volta Region while the NDC’s parliamentary candidate in the constituency got 27,204 votes. Peacefmonline has the same figure for the NDC’s parliamentary candidate but has a much lower number for votes (i.e., 28,016) for president Mahama. I think peacefm’s figure is the right figure.
I do not have the 2016 figure for the number of registered voters in Ho-West. In 2012, it was 48,891. Including’s votes of 53,117 for Mahama, the total number of valid votes cast in 2016 was 59,466. This will require at least an increase of 21.6% in the number of registered voters from 2012 to 2016 (and a turnout of 100%) for over-voting not to have occurred in this constituency.  
APPENDIX A: The EC versus the NPP: an algebraic evaluation of Sir John’s statements in
Written on December 14, 2012 Prior to the announcement of the 2012 presidential results by the EC, the NPP’s General Secretary, Kwadwo Owusu Afriyie (aka Sir John) announced at a press conference on Saturday Dec 8, 2012 that the NPP had won the election with 51.3% of the total valid votes cast for Nana
Addo and about 45% for John Mahama:­addo-51-3-mahama-45-3/ From CNN: “NPP General Secretary Kwadwo Owusu Afriyie had said Saturday night, (that) …
Akufo-Addo had won the presidency with 51% of the vote.”: Six days later, December 14, 2012, Kwadwo Owusu Afriyie said that “… the Electoral Commission (EC) rigged an excess of one million votes for John Mahama”:
Let’s examine both claims by Sir John: According to the EC: Total Number of valid votes = 10,995,262, Nana Addo’s votes = 5,248,898, John Mahama’s votes = 5,574,761. Let Nana Addo’s votes be reduced by X and Mahama’s votes be increased by Y. Therefore, the
alleged rigging in favor of Mahama was by (X + Y) votes. Then based on Sir John’s statements, the following equations must hold: X + Y = 1,000,000 + α,         (1) and
5,248,898  X
 0.51,       (2)
10,995,262  X  Y
where α ≥ 0.
It is important to note that we require X ≥ 0 and Y ≥ 0. Otherwise, X < 0 means that Nana Addo’s actual number of votes, as claimed by Sir John, was smaller than the figure reported by the EC and Y < 0 means that John Mahama’s number of votes was greater than the number reported by the EC. This will be rigging in favor of Nana Addo. This was certainly not Sir John’s claim.
Solve equations (1) and (2) to get:
X = 7,565,719 + 25.5α and Y = − 6,565,719 − 24.5α.     (3)
Suppose α = 0, so that Sir John’s claim boils down to the EC rigging the election by exactly 1 million votes in favor of Mahama. Then X = 7,565,719 and Y = − 6,565,719. But Y = − 6,565,719 < 0 contradicts the requirement that Y ≥ 0. In fact, it is clear from (3) that Y < 0 for any α ≥ 0.
Therefore, Sir John’s statements on December 8, 2012 and December 14, 2012 were inconsistent with each other. One of them was false or both statements were false. Using 0.531 (51.3%) instead of 0.51 (51%) in equation (2) does not change this conclusion.
Finally, equations (1) and (2) are linear equations. Therefore, the solution in (3) is unique (i.e., the two straight lines in (1) and (2) can intersect only once). Therefore, we do not have to worry about other solutions.
APPENDIX B: similar allegations of vote-rigging in 2008 when the NPP was in power (4 years before 2012)
The Successful Ghana Election of 2008 – a Convenient Myth? Ethnicity in Ghana’s Elections Revisited
Authors: Heinz Jockers, Dirk Kohnert, Paul Nugent
GIGA Research Programme: Transformation in the Process of Globalisation
Turnout in five of the constituencies in Ashanti region – Bantama, Kwadaso, Manhyia, Nhyiaeso, and Suame – also demonstrated unusually high numbers of votes cast, which was over 95 per cent of registered voters.
At Manhyia where the recorded turnout was 95.7 per cent, the NPP vote rose from 66,116 in the first round to 95,281 in the second (+24 per cent). At Nhyiaeso, where the alleged turnout was
98.3 per cent, it rocketed from 37,043 to 54,545 (+26 per cent) and in Suame it climbed from 47,768 to 67,790 on a 95 percent turnout (+24 per cent; cf. Table 1). The total turnout figures for Ashanti (83.3 per cent) were out of line with the figures for the rest of the country (72.7 per cent). Constituencies that declared late included the same ones where there was an implausible turnout, notably Nhyiaeso (98.3 per cent), Kwadaso (94.5 per cent), Manhyia (95.7 per cent) and Suame (94.7 per cent).
In the 2000 presidential elections – the first democratic polls which brought about an alternation of power (from the NDC to the NPP this time) and which were also highly contested – the turnout was relatively low. This was attributed at the time to a bloated voter register. On the national level, valid votes accounted for only 60.8% in the first run and 59.7% in the sec‐ ond one. But even here, Ashanti and the Volta Region scored highest, with 64.3% (65.0%) and 59.7% (67.8%) respectively (cf. Smith 2002: 629, also for a detailed discussion of differences of voter turnout in Ashanti and Volta in the 2000 elections).
Money was allegedly used to get registration officers ‘on board’. The deal was said to involve the registration of anyone provided by the party with no questions asked, including ‘minors’, and to place potential opposition voters on a wrong list of voters. This was precisely what happened in Hohoe during the registration process where two employees of the Regional Electoral Commission (REC) were arrested by the police for manipulating voters’ lists. According to information released by the police, the accused confessed having taken money from the NPP. However, no complaint about the registration process was brought to the knowledge of the EU observers. Other fraudulent practices concerned ballot papers meant for NDC strongholds that would be rendered useless by printing mistakes and subsequently be rejected by NPP party agents. This happened in a few places (e.g. in Nkwanta District), but its net effect was only to delay the voting process, not to hinder it. In addition, the NPP organized so‐called ‘Party Youth’
in a number of constituencies like Krachi East, with the aim of harassing opposition and election officials. This happened in Dambai.
What is open to speculation is the much higher turnout in the second round, namely a total of 739,669 votes as compared to 679,396 in the first round, amounting to an increase of 60,600 in the number of votes cast. Some of the increase in specific constituencies would warrant closer investigation by the REC or EC. 11,167 more votes were cast in Ketu S., 4,792 in Anlo, 5,073 in Avenor, 4,786 in Tongu N., 4,495 in Ho C and 4,954 in Ho East, all of which are outstanding examples.

Leave a Reply

Your email address will not be published.