It appears that John Atta Mills has been elected President of Ghana. Although the result will not be declared until tomorrow, it now appears in practice impossible for Nana Akuffo Addo to close the gap.

There remain a number of concerns about the count which puzzle and worry me. In particular the swing ti Mills in the final fifteen constituencies to declare appears to be three times the average swing over the rest of the country. Constitutencies which together delivered a net majority to Nana Akuffo Addo of over 150,000 in the first round have yielded him a majority of only about 40,000 in the second round. Looking at each in turn and the swings in the surrounding constituencies, there is no readily available explanation that occurs to me. For example Bantama and Kumawu in Ashanti region, both in the final batch of results, showed substantial falls in Nana Akuffo Addo’s vote whereas all the other seats in Ashanti Region had shown a sufficient increase. Beyond doubt the last twenty constituencies to declare have been much better for Atta Mills than any rational amalysis would lead you to predict.

However I understand that the Electoral Commissioner, Kwadwo Afari-Gyan, is inclined to accept the result as genuine. He was personally out and about during voting in some of the Volta constituencies which particularly concerned me earlier.

I would trust Kwadwo Afari-Gyan with my life. I personally witnessed him, just the two of us in the early hours of the morning, refuse to budge when soldiers held his wife and children at gunpoint and threatened them unless he falsified the result of the 2000 election. If Kwadwo accepts the result, so will I, and I urge Ghanaians to do so too.

Alternation of power is a healthy feature of democracy, and Mills is a good man. But the elephant in the room is the ex-dictator, multiple murderer and half (at least) mad Jerry Rawlings. Does he still control his protege Mills? We have no choice but to wait to find out.

I am becoming very concerned about the electoral process in Ghana. With 207 results declared, John Ata Mills has a lead of 200,000 votes, but in the first round Nana Akuffo Addo had a majority of 170,000 in the constituencies yet to declare – and has been substantially increasing his lead in his strongholds in the second round, while falling back elsewhere.

I hve already mentioned the extraordinary leaps in the NDC vote since the first round in some Volta constituencies. And now we have this extraordinary result declared:

Evalue Gwira (Central Region)

Nana Akuffo Addo 10,818 (minus 36,182 on first round)

John Atta Mills 9,094 (minus 5,906 on first round).

Elsewhere we have the extraordinary appearance of 50% more NDC voters in just three weeks. Here we have the disappearance of 75% of NPP voters in the same period.

In my book The Catholic Orangemen of Togo I introduce the concept of the margin of cheating.. There is regrettably cheating in elections in every country in the World. The problem becomes acute where the amount of cheating exceeds the margin of victory. That is the suspicion which will always hang over the Bush election win in 2000. It looks like this Ghanaian election is going to be won by a figure within the margin of cheating.

Incidentally, in Blackburn where I stod against Jack Straw in the last general election, almost a third of votes were cast by postal ballot – three times the national average. The postal ballots favoured Jack Straw by a margin far higher than the “normal” ballots. I would estimate that Labout boosted its vote by cheating in Blackburn by some 20%.

My last post mentioned that an advantage of statistical psephology is that it highlights anomalies as pointers to possible abuse.

I am concerned by some quite extraordinary figures of increased voting for the NDC in certain parts of Volta Region, which are difficult to believe can be genuine. This is particularly so as they occur in districts where there was no or negligible third party vote.

It is very hard to believe that in Hohoe South, for example, the NDC managed to increase its vote by a full 50% after the first round three weeks ago. Increases of 20 to 25% in Anlo and Avenor also seem extraordinary and out of line with what is happening in general.

These large apparent increases in voter interest have resulted in apparent voter turnouts in excess of 90%. There is a natural friction on election registers, due to death, people moving, being ill or away at election time, forgetting or not wanting to vote, etc. Voting levels in the areas mentioned are apparently well above the average for Ghana and at levels I am not sure I believe to be practical – another flash of a warning signal.

Further Projection

With 135 constituency swings now calculated, a run of very large swings to Mills has increased his projected majority ti 33,000.

Having now calculated the exact swings between the parties in 115 of the 230 constituencies, and applying the average swing across those constituencies which have not declared, we now project a win for John Atta Mills and the NDC by 14,000 votes, or by 50.08 to 49.92%.

It remains a fact that our projections have been remarkably consistent; and that the methodology proved extremely accurate in predicting the results of the first round. But again it must be stated that this is so close that it could yet go either way.

I am making a projection based on sound psephological principles and a methodology used worldwide to project election results. The calculation is then purely mathematical. This is an exercise in prediction largely for fun, but it also has a use in that, if the methodology throws up any anomalies, they could represent fraud. In fact in general the consistency of results within regions in terms of swing trend tends to support the idea that these are fair and genuine elections.