Thread by @DawieScholtz

1/ So next stop: LIMPOPO!

This is one gets way too little attention and analysis. It has been a vital ground zero for ANC losses in the last decade.

High-level expectation, compared to 2014:
ANC losses
EFF gains
DA minor gains

But still an easy ANC hold.

2/ History shows that the ANC obviously comes off a very, very high level here:

Previous Limpopo election results (ANC - EFF - DA):
2011: 83-0-7
2014: 79-10-7
2016: 69-17-8

Special note: That's -14% for the ANC in 5 years (!) and all going to the EFF. LM is crucial for EFF.

3/ The ANC has historically been incredibly dominant in Limpopo because of demographics. The black electorate is roughly 2.45 mn of the 2.55 mn voters in Limpopo.

In the past, when the ANC was almost uncontested in the townships, it therefore won enormous majorities in Limpopo

4/ But things have changed in the last 3 elections. We've observed two key trends:

- Weak turnout in Limpopo's townships
- A deflation of ANC support in Limpopo's townships

Limpopo township turnout vs Gauteng township turnout:
2011: 49%-53%
2014: 61%-74%
2016: 49%-53%

5/ The township voting pattern across all of Limpopo (roughly 2.4 million voters):

ANC-EFF-DA (%)
2011: 86-0-3
2014: 84-10-3
2016: 74-16-4

That's -12% for the ANC in 5 years. And this excludes a special pocket of the township electorate where the EFF is especially strong.

6/ As an aside, it should be noted how important Limpopo is to the EFF's overall tally. In 2016, they won 213000 of their 1.2 million votes (17%) in Limpopo.

Cyril Ramaphosa's - yet untested - popularity in Limpopo could therefore be a significant risk for the EFF.

7/ So what to expect for May (1/2)!?

Polling suggests the EFF is on the up and up. And if it's true nationally, it's going to be true in their heartland. By-elections have been uneven, but in the aggregate good for the EFF in Limpopo and bad for the ANC.

8/ Expectations for May (2/2):

So highly likely that the EFF outperforms their 2014 numbers in Limpopo's townships. Crucial questions is whether they hit the 2016 high again (17%) given Ramaphosa's entry. There's significant uncertainty about this.

9/ Ultimately, what I'm expecting based on what we know now:

ANC: Closer to 70 than 80
EFF: Closer to 20 than 10
DA: 5-10

Disclaimer: This is based on unscientific 'best-guesses' for various unknown variables, informed by polling and by-elections. Not a formal prediction.