By Potpher Mbulo
DATE: 08-04-2011
As a follow up to my note entitled “Who will win 2011 Elections?” I have taken it upon myself to put into account observation of critics of my work and I have also incorporating my detailed second opinion.
ANALYSIS:
To start with, though the total registered number of voters (RV) and the average voter turnout (VTO) at national level does not affect verdict of election outcome but affects total votes cast, however, RV and VTO at provincial level affects electro outcome. Therefore it’s not just my opinion but it is a fact that statistical calculations based on summed up presidential election results at national level are not very close to reality. A better analysis is one that takes into account the country’s demography at provincial level. This is what I aim at doing in this presentation. In this regard, data of cardinal consideration are:
1. Total Registered voters at provincial level
2. Voter turnout at provincial level
3. Distribution of votes based on past election trends at provincial level
4. Popularity growth at provincial level
5. The Patrick Levy Mwanawasa (PLM) factor
6. The Charles Milupi (CM) factor
7. The Mongu demonstration factor
8. Possibility of rigging
9. New young voters
Lets ride through how we incorporate the above considerations into the analysis:
1. Total Registered voters (RV) at provincial level:
As earlier assumed in the first presentation, that 2011 election will register more than 5.5 million voters. Now based on past election registration trends, I have relatively apportioned registered voters to provinces according to share ratios trends. As of the 14th January 2011 ECZ’s register of the on-going registration process showed:
CB 832,116
LSK 755,333
NOR 626,682
SOU 625,190
EAS 614,373
CEN 462,643
LUA 394,872
WES 375,821
NW 297,685
TOTAL 4,984,715
Extrapolating the share ratio trend while limiting the total number of registered voters to 5.5 million, we calculate that the following holds:
Copperbelt: 914,812 registered voters
Lusaka: 834,682 registered voters
Northern: 723,391 registered voters
Southern: 667,746 registered voters
Eastern: 645,488 registered voters
Central: 556,455 registered voters
Lwapula: 500,809 registered voters
Western: 389,518 registered voters
North Western: 267,098 registered voters
TOTAL: 5,500,000 registered voters
Just a brief comment: The reason why PF is so prominent at national electro outcome is because it’s popular in the three prominent provinces in terms of registered voters (RV) i.e. Copperbelt, Lusaka and Northern provinces have the most registered voters. Notice that RV is highest in these regions and actually, the list of provinces here above was made in accordance with descending RV with the most on top. Most importantly, this data has been reviewed in light of new raw data of 31-03-2011 close date of ECZ register. (See notice at the end of this presentation)
2. Voter turnout (VTO) at provincial level:
To ascertain voter turnout (VTO), I had to refer to past trends as well as a discussion with friends. For instance, it is expected that duel to the Mongu fracas, there will be apathy (VTO = 55%) towards voting in Western province. The apportioned VTO for each province is as follows:
Copperbelt: VTO = 70.00%
Lusaka: VTO = 70.00%
Northern: VTO = 65.00%
Southern: VTO = 75%
Eastern: VTO = 93.66%
Central: VTO = 75.10%
Lwapula: VTO = 65.00%
Western: VTO = 55.00%
North Western: VTO = 65.00%
3. Distribution of votes based on past election trends at provincial level:
In general, the distribution of votes to candidates per province is calculated by considering registered voters (RV) of the province, voter turnout (VTO) and popularity growth. The formula I used is here below:
P = rtw(1 + g)/100
Where: P = postulation of 2011 election result
r = number of registered voters in year 2011
t = percentage voter turnout
w = latest poll score i.e. 2008 election in percentage
g = popularity growth or gain in poll score of year 2006 to 2008 in percentage
Notwithstanding the rule’s generalization, this rule has been disregarded rather overridden by the Patrick Levy Mwanawasa factor at Copperbelt and Central provinces. Again the rule has been overtaken by the Charles Milupi factor in Western province. In these three instances I have apportioned sharing ratios as percentages.
4. Popularity growth at provincial level:
The formula for determining popularity growth is as follows:
g = (w1 – w2)/100
where w1 = latest percentile poll score i.e. 2008 election
w2 = percentile poll score in 2006 election
As earlier mentioned, even this rule has been disregarded rather overridden by the Patrick Levy Mwanawasa factor at Copperbelt and Central provinces. Again the rule has been overtaken by the Charles Milupi factor in Western province. In these three instances we have apportioned sharing ratios as percentages.
5. The Patrick Levy Mwanawasa (PLM) factor:
The Patrick Levy Mwanawasa (PLM) factor is presumably a belief that some members (Lamba/Lenje) within MMD such as Shakafuswe, George Mpombo, Mike Mulongoti, etc have fallen out of favor with the MMD. In view of this, a sizable number of votes will swing back from MMD to UPND in Central Province and Rural Ndola of Copperbelt Province. We have apportioned sharing ratios as percentages.
Copperbelt province:
With PLM factor Without PLM factor
PF = 65.40% 65.08%
MMD = 20% 29.24%
UPND = 14% 5.10%
Others = 0.60% 0.57%
Central Province:
With PLM factor Without PLM factor
PF = 27.6% 27.57%
MMD = 33% 51.19%
UPND = 38.6% 20.42%
Others = 0.80% 0.81%
6. The Charles Milupi (CM) factor:
An influential new entrant on the political scene in Western province has emerged. This is no other than Charles Milupi. Though he might not stand in the elections in 2011 at presidential level, with his exit from MMD, he definitely will swing a significant number of votes from MMD to the opposition mostly to UPND.
With CM factor Without CM factor
PF = 9.99% 10.00%
MMD = 50% 65.43%
UPND = 40% 23.43%
Others = 0.01% 1.15%
7. The Mongu demonstration factor:
The issue of Barotseland Agreement visa vie Mongu demonstrations early this year and failure by MMD to do damage control should not be overlooked in Western province. Is it termed “damage control” or is it “damage repair”? Anyway, whatever you say, the effect will tend to alienate voters in Western Province. A few of the disgruntled elements will vote for the opposition but mostly of these angered individuals will might likely choose to stay away from casting their vote. For this reason, I have set the Voter Turnout at 55% in Western Province. For those that will vote for the opposition, I have incorporated their grievance in the Charles Milupi factor shown above. In any case, this low voter turnout ratio tends to caution the Charles Milupi factor’s contribution to national summation and thus there is no need for skeptics to cry foul play on my part.
8. Possibility of rigging:
Rigging can only be to an extent. If an incumbent loses too much, it’s difficult to rig. Moreover, if MMD did rig before, say in 2006 and 2008, that rigging component is already factored in the data as the analysis looked at the final ECZ results.
However I know that Voter Turnout Ratio in Luapula in the years 2001, 2006 and 2008 were as follows respectively: 71%, 36% and 30%. I can explain why there was a low voter turnout in 2008 as it was a by-election with a lot of relocation of voters since voter registration in early 2006 making it difficult for electorates to cast their vote after two years. However, I was just wondering why there is a big disparity between year 2001 and 2006 in terms of voter turnout. FTJ who hauls from Luapula was supporting LPM in 2001 and perhaps GRZ inflated the VTO ratio in Luapula as it was very difficult to monitor elections in the light of the fact that many roads were impassable in 2001 in Luapula. If, I’m not mistaken, ZAF Helicopters were used to ferry ballot boxes from remotes to totaling centers back then in Luapula. In this regard it was easy to rig. However, I’m not saying that MMD did rig.
Well you may ask: how can GRZ manipulate Voter Turnout Ratio (VTO)? I’m glad you asked. Simply by slowing down the voting process and by making sure ballot papers are delayed in reaching polling stations where the opposition is strong. That way GRZ can insure voting starts at say 11:00 hrs instead of 7:00hrs and stop the exercise at 18:00hrs in Lusaka, Copperbelt and Northern Province. In the mean time, GRZ can commence voting at 6:00hrs and end at 19:00hrs in the incumbent’s stronghold such as Eastern Province. Deliberate creation of apathy or high voter turnout is a great tool for smart rigging.
I have deliberately put the VTO ratio in Eastern province at all time high of 93.66% to cater for rigging. Without this, say we reset the VTO ratio to 70% in Eastern province, we see PF win at national total of 39.69% followed by MMD at 36.60% and with UPND at 22.91% of total votes cast translating into 1,551,267 votes for PF followed by 1,430,452 votes for MMD while UPND has 895,502 votes. In such a scenario, PF leads MMD by over 120,000 votes.
9. New Young Voters:
There are over a million new young voters on the register this year 2011. Most of these want change of government and they are mainly pro PF. I don’t want to appear to be harsh to MMD. I have not assigned any factor to the computation. I assume this is taken care of in the usual various political parties’ popularity growth rate. If critics are not satisfied, they can argue with statistics on which the popularity growth rate is determined and I must hasten to say these statistics on which I calculate popularity growth rate are not mine but is raw data from ECZ.
Now we are through with the cardinal considerations. Let’s go to the results that show the final outcome. The results of this comprehensive analysis that takes into account the issues discussed above yields the following final tabulated results:
SAMMARY OF RESULTS:
Jan 2011 Registered Voters = 4,942,000
2011 Projected Registered Voters = 5,500,000
Total Votes Cast = 4,094,965
Average Voter Turnout = 74.45%
2001 2006 2008 2011 Votes Cast
PF: 3.35% 29.37% 38.13% 38.60% 1,580,792
MMD: 28.69% 42.98% 40.09% 38.60% 1,580,817
UPND: 26.76% 25.32% 19.70% 21.99% 900,542
Others: 39.59% 2.33% 0.76% 0.80% 32,815
TOTAL 98.39% 100.00% 98.68% 100.00% 4,094,965
DISCUSSION:
We should not be deceived by the results in the table that shows that MMD and PF have a tie at 38.60% while UPND is trailing at 21.99%. What makes MMD to equal PF is the setting of voter turnout in Eastern Province at an all time high of 93.66%. Since MMD is popular in that region, Eastern Province contribution to total national count becomes very profound.
CONCLUSION:
Only if and if only the voter turnout trigger point of 93.66% in Eastern Province is achieved, MMD’s presidential candidate will not pass the post else PF’s presidential hopeful will carry the day. Like earlier mentioned, I have deliberately put the VTO ratio in Eastern province at all time high of 93.66%. This is the breakeven point for MMD to bypass PF all things being equal. Without this, say we reset the VTO ratio to 70% in Eastern province, we see PF win at national total with 39.69% followed by MMD at 36.60% and with UPND at 22.91% of total votes cast translating into 1,551,267 votes for PF followed by 1,430,452 votes for MMD while UPND has 895,502 votes. In such a scenario, PF leads MMD by over 120,000 votes.
NOTICE:
As the ECZ Register closed on 31 March 2011, the unverified data is as follows:
CB 869,406
LSK 776,866
NOR 658,902
SOU 655,380
EAS 654,204
CEN 482,571
LUA 409,754
WES 389,609
NW 310,398
TOTAL 5,207,090
I did plug in this raw data in my excel sheet & against my postulated RV values. I’m glad to report that I calculated average discrepancy of 0.0001098%. This proves that my postulation based on past election trends was accurate within reliable range & thus my conclusion is valid.
However I admit that I had overestimated the total registered voters i.e. 5,500,000 registered voters yet it’s supposed to be a total of 5,207,090 registered voters. I had overestimated my data by 292,910 translating into 5.63% overestimation. Does this have any consequence on the final result and the eventual conclusion? Absolutely not as you will see.
Using this new raw data which was captured by close of ECZ Register on 31-03-2011 and placing VTO in Eastern Province at 70%, PF still wins scoring 38.42% (1,421,376 votes) while MMD follows closely at 37.33% (1,381,000 votes). This means PF wins the election with over 40,000 votes ahead of MMD. Notice that in 2008, PF lost by 35,000 votes against MMD
With this raw data, for MMD to breakeven with PF, the trigger point is achieving a VTO of exactly 77.80% in Eastern province against an average national VTO of 72.24%
