Monday, February 18, 2013

The myth of 'Tyranny Of Numbers' bebunked and exposed as a hoax

The projection that Uhuru Kenyatta starts off with an unbreakable numerical lead is pure myth. That 43 per cent Mutahi Ngunyi speaks of may actually represent the best that Uhuru can hope for. (Remember Mitt Romney's 47 per cent?)

Nothing has provoked as much controversy in this election cycle, as the narrative about a “tyranny of numbers” put out by the veteran political commentator Mutahi Ngunyi a few weeks ago.

The basic points of this narrative are as follows: historically, Kenyans always vote on a clear tribal pattern, irrespective of what the issues are in any election. As such, if you look at the combinations of key tribal leaders who have agreed to work together in a political coalition, you can more or less ascertain how many votes each coalition will get.

Arising from this comes two key points: that (as one source puts it) “Jubilee Coalition and Its Tyranny of Numbers; if history is a reliable reference point, we assume that the Gema and the Kalenjin nations will vote for this coalition as a bloc. If they do, the combined vote for this bloc is 6.2 Million or 43.2 per cent of the total vote according to the IEBC December 18 voter registration tallies.”

This source adds that “The CORD coalition is primarily composed of the Luo nation and 80% of the Kamba nation. The combined starting point for CORD coalition is 19.2 per cent or 2.74 million votes…arrived at this figure by adding the Prime Minister’s ethnic support (Luo) and that of the Vice President (Kamba).”

This, of course is what caused so much consternation among those who support CORD and so much rejoicing among the supporters of the Jubilee Coalition. By any standards, a political coalition starting off with 43 per cent of the total Kenyan vote locked up in its strongholds would be more or less invincible. Especially if the principal rival coalition has only 20 per cent roughly, as a starting point. Indeed Mr Ngunyi actually went so far – in his TV interview – as to call the presidential election in Jubilee’s favour, announcing that the outcome of this election had been predetermined at the voter registration exercise, and CORD was bound to lose.

This kind of wild generalisation would be merely amusing, for very many reasons, if not for the fact that it has brought back the spectre of massive rigging into the upcoming election.

When speaking in private, the hard core of Jubilee Coalition supporters – who base their expectations entirely on the ‘tyranny of numbers’ theory – claim that they are sailing to victory with the greatest of ease. And that nothing short of phenomenal rigging could possibly deny Uhuru Kenyatta the presidency in the very first round of voting.

On the other hand, the partisans of CORD are willing to come out quite openly to allege that the whole purpose of this deeply-flawed theory of the ‘tyranny of numbers’ is to create the right psychological framework for massive rigging by Jubilee. And that the whole point of this theory being put out in the first place, is to provide a narrative which will be used to explain why and how Uhuru Kenyatta won – after he has been rigged in on March 4.

Thus, it is worth considering what merit, if any, this theory has, and whether it is in any way an accurate rendering of what we should expect in the elections.

Perhaps the most effective refutation of the ‘tyranny of numbers’ theory in the results of the recent opinion polls. Virtually all the reputable pollsters are unanimous in predicting, as of this point, a “dead heat” between the two leading presidential candidates, with the CORD candidate, Raila Odinga considered to have a slim lead. In general, Uhuru Kenyatta is rated to have somewhere between 42 per cent and 44 per cent of the vote at this point in time, and Raila Odinga is rated at between 44 per cent and 46 per cent. Given the margin of error in such polls, this amounts to a situation in which either candidate could just as easily win.

Interestingly enough, this is more or less what the pollsters had as Kenya went into the 2007 general election: Raila had a slight lead over the incumbent President Mwai Kibaki in that election, which was also generally judged to be “too close to call”. To this we may add the fact that Kenyan pollsters have accurately predicted the results of the 2002 presidential elections, the 2005 referendum, the 2007 presidential elections and the 2010 referendum.

This is a track record of reliability which would be acceptable in any nation and at any time since the advent of opinion polling.

So, the first question which could be posed to those who uncritically accept the theory of ‘the tyranny of numbers’ is, If Raila Odinga starts off with only 20 per cent of the vote, then where does he get the additional 25 per cent of the vote that pollsters are predicting he will get?

Or put another way, since Uhuru Kenyatta’s supposed ‘tyranny of numbers’ only amounts to 43 per cent - the same percentage of support that the pollsters give him – does this mean that only the Gema and the Kalenjin will vote for Uhuru, while almost everyone else will vote for Raila? Are we facing the spectre of a clear voting pattern of '40-against-2'?

Oddly enough, there is nothing unusual about people saying to pollsters that they will vote for one candidate, and then turning round and voting for the rival candidate. It is actually quite common in an election where deep (but often unspoken) racial or communal rivalries exist, which is considered to be ‘politically incorrect’ to discuss openly. In the US, this is known as ‘the Bradley Effect’ among political scientists and newspaper columnists.

In a 2012 article in The Daily Caller, political columnist Mickey Kaus explained, “The Bradley Rule holds that voters will be reluctant to tell pollsters they are voting against an African-American for fear of being labelled racist. It allegedly hurt Tom Bradley in 1982 and Douglas Wilder in 1989.”

Indeed, thanks to what happened to Doug Wilder (i.e. he too seemed to lead in the polls, but ended up losing) this phenomenon is sometimes called the Wilder Effect too. And it is always a consideration when there is a black candidate on the ballot for a significant political office like governor or senator.

The New York Times columnist Nicholas Kristof noted in 2008, for example, that “Democrats are wringing their hands in nervousness over the 'Bradley Effect', in which black candidates poll well but win fewer votes than expected. That’s usually attributed to voters lying about racism — telling pollsters they’re going to vote for a black when they actually plan to vote for a white — but some scholars believe it has less to do with lying than with unconscious racism that leaves voters with growing doubts that lead to a last-minute mind change in the voting booth. In any case, though, the Bradley Effect has been diminishing in recent years, and in many caces, has disappeared.”

In the Kenyan context, what we are faced with at this time is not an election result which paints a completely different picture from that which the polls had led us to expect. Rather, it is a conflict between two narratives, based on more or less the same voter registration figures.

Nobody seriously doubts that the great majority of Kenyan voters tend to go along with a clear tribal preference when it comes to presidential elections. So Mr Ngunyi’s point of departure – that the Kalenjin and the GEMA ‘nations’ together amount to a huge basket of votes – is not in doubt.

The unanswered question though, is, what has Mutahi Ngunyi missed? For if his projections were accurate, then there should be no question of a ‘dead heat’ between Uhuru and Raila. It should have been equally obvious in the opinion polls that Uhuru had a huge (though by no means insurmountable) lead. Or at the very least, the numbers given by the leading pollsters should vary so widely as to create serious doubts about their credibility.

One pollster may be wrong; but they cannot all be wrong. So how do you reconcile the pollsters being unanimous that this election is a dead heat, while Mutahi Ngunyi’s theory – and a very influential theory it is – argues that Uhuru Kenyatta has such a huge lead, that Raila cannot hope to catch up?

Just as the Amercian political scientists turned to psychology to explain what they thereafter termed as “the Bradley Effect” so too, must we now turn to psychology to explain this divergence between the two sets of statistics: Mutahi Ngunyi’s reliance on the registration numbers from the Independent Electoral and Boundaries Commission (IEBC) to project an unassailable lead for Uhuru Kenyatta, and the combined results of reputable Kenyan pollsters, which suggest that either Uhuru or Raila could just as easily win, with Raila – at this point – having a slightly better chance of winning.

But before we even consider the psychological factors, we should note that the ‘tyranny of numbers’ theory is based on just 62.4 per cent of the vote. That alone makes it suspect. If this theory were to be taken seriously as in any way reflecting final outcomes, it should at least have accounted for 95 per cent of the vote, thus giving some indication as to why one candidate had effectively won already; and the other was doomed to fail.

In this case, there is a clear 37.6 per cent which is not accounted for. This is far too large a number to be referred to as a ‘swing vote’. The swing vote factor only comes in when 90 per cent or more of the vote has been accounted for. Using yet another American example, here is an analysis published in 2001 in the Jerusalem Center for Public Affairs, and written by Jeffrey Helmreich. It is titled ‘The Israel swing factor: How the American Jewish vote influences US elections’ and it made this argument:

“There are two clashing myths on the political power of American Jewry. One claims that the community is too small to affect national elections; Jews make up less than three per cent of the US population. A contrasting view holds that US Jews play a disproportionately large role in national politics thanks to their campaign donations and media influence.
According to evidence confirmed in the most recent elections (2000) however, American Jewish voters maintain the potential to be the decisive factor in national election results. Yet, this ability does not emanate from any financial or public relations clout, which is overestimated. Rather, American Jews wield power through their high concentration in key states, and their tendency to behave as a swing vote in ways that set them apart from virtually all other groups in American politics.”

So if a population as small as three per cent counts as a decisive swing vote in close elections, then how can any calculation that leaves out a clear 37 per cent of the vote – and confidently predicts a precise outcome from this calculation – possibly be taken seriously?

A different interpretation of the IEBC registration numbers then – when reconciled with the results of the opinion polls – may well be that most of the country is united against the Gema-Kalenjin political coalition that goes by the name Jubilee. And that perhaps a significant number of Kalenjin and Gema community members have also decided to vote against their tribal “wave” and support Raila Odinga instead. And that this claim of a solid 43 per cent of the vote having been locked in by the Jubilee coalition is a hoax, which will be exposed when the final tally is made.

There could be any number of reasons why “the rest of the country” might unite against Jubilee: it could be a growing awareness that the Jubilee candidates for president and deputy president will not be here to preside over anything if they should win; that they will have their hands full defending themselves at The Hague.

It could also be a fear of economic sanctions if the two refuse to obey the summons from The Hague.

It could further be a simple sense of resentment that only those two communities – the Kikuyu and the Kalenjin – have ever had “one of their own” occupy State House, and so someone from some other tribe should now have an opportunity to lead the nation.

Or it could be that – contrary to appearances on the surface – there are many within the Kalenjin community who are not yet reconciled to voting for a Kikuyu presidential candidate, given the extreme bad blood between the two communities as revealed in the 2008 post-election violence.

It could be that when William Ruto gave up his own presidential ambitions to support Uhuru Kenyatta, he did not really take the entire Kalenjin vote with him into the Jubilee coalition, much as the DPM Musalia Mudavadi evidently did not carry all of Western Kenya with him into the Amani Coalition.

This, of course, would be a classic example of the psychological phenomenon defined as the ‘Bradley Effect’: that Kalenjin voters may currently appear to be reconciled to voting in a Kikuyu president, but then when they get to the ballot box, we will find that – owing to unresolved historical differences – they are, ultimately, not willing to do so.

In short, it may well be that the true “tyranny of numbers" will be expressed in a 57 per cent vote against Uhuru in the widely-anticipated runoff between him and Raila; that although he is a strong candidate with massive grassroots support in Central Kenya, Uhuru Kenyatta will lose solely because it goes against the national sentiment to have a Kikuyu succeeding another Kikuyu to the presidency.

And it is to be noted that such a perspective does not argue that Mutahi Ngunyi cooked up those numbers from the IEBC. Those numbers – on the face of it – appear to be real enough. What is however revealed is that Ngunyi had no real idea what those numbers signified; and that he lacked the sophistication to go beyond a simple tribal calculus which, in any case, he based on less than 65 per cent of the total vote.

We can conclude then, that the ‘tyranny of numbers’ narrative is little more than an empty hoax. The alleged ‘starting point’ of the Jubilee Coalition is basically just the 43 per cent that all the polls estimate Uhuru Kenyatta will get. And the only ‘tyranny’ possible, is that the rest of the Kenyan voters will unite to deny Uhuru victory in the second round of balloting.

The projection that Uhuru starts off with a huge numerical advantage is pure myth. That 43 per cent may actually represent the best that he can hope for. It is an ‘end game’ number, not a starting point. And unless all the pollsters are wrong – despite having been right for the last five electoral events in Kenya – this presidential election is going into a runoff which distinctly favours Raila Odinga as the eventual winner.

No comments: