What is the Prediction for NDA in 2019?
We will use CrowdWisdom’s various methods to arrive at a single number for the Prediction for NDA in 2019.
What is the Crowd’s record in election predictions for the last 6 elections?
In the last 6 elections, the Crowd’s record has been quite healthy when compared with exit pollsters with an average error of 18%. Axis, the best pollster had an error of 18%. Please remember that the Crowd constitutes of people from across the country accessing a variety of digital media – Whatsapp, Facebook, Twitter and LinkedIn.
We will use 3 different methods from our platform to arrive at a single prediction for NDA in 2019
The Straight Question Method
In this case, we ask voters to fill a form. See the form to predict the 2019 election here
In this case, we arrive at NDA = 296 Seats
Our average error has indicated earlier is 18%. Typically when NDA has been the incumbent, the error has been on the upside. So we correct this by 18%. We arrive at a figure of 251 Seats
The Polling Question Method
In this, we ask a direct question to the predictor. Predictor usually make up for about 2-3% of all the people who visit the web-site.
If one notices carefully
300+ = 33.9%
< 200 is 13.6%
So it is clear that the most preferred number is located somewhere between these two.
If you look at % of people who have chose 250 to 275 and above, it comes to 67.8%. When we look at our performance in the past election, at 65%+, for any choice (for BJP/NDA), the prediction becomes more accurate
So with this method, the sweet spot is 250 to 275 seats.
This is a number that we had arrived using bottoms up calculations(state level to country. Read the NDA prediction for 2019 calculated in March
The Prediction Market Method
This is our stock market for predictions (No real money is traded). You can see the market traded volumes and prices by clicking here
As we mentioned yesterday, the sweet spot for BJP appears to be in the 210 to 225 range and the most bought prediction for NDA is in the 250 to 275 range
There are of course two opposite schools of thought – NDA supporters who believe that NDA is headed for a massive win because of Prime Minister Modi’s popularity and Opposition supporters who believe that NDA is headed for a shock defeat (like 2004) on account of ordinary economic performance. But the chart below suggests that the situation on the ground may be far more complex
When we look at Google Trends and check for Modi-BJP correlation, we find that the correlation was quite high 0.75 in 2014 (1 is the max possible). But when one looks at 2019, the correlation is just 0.05 (negative). Part of the reason is the searches for Modi have increased substantially since 2014. But the national figure in 2019 hides the substantial variances between the States. Further, the correlation appears to be the highest in States where the party is supposed to do well (from opinion poll data) and the lowest where the party is relatively weaker or expected to struggle.
In other words, the Modi strong and weak theories may not be consistent and both statements may be true depending on the State you are visiting.
Our own read of the data is that parties in power in each State seem to be struggling much more than the Opposition party in that State.
All 3 methods appear to converge (after correction) in the 250 to 275 range. This would imply a loss of between 81 to 106 seats for the NDA versus 2014 with primary losses in Tamil Nadu (new ally AIADMK) and in Uttar Pradesh where the losses may occur due to the alliance. The gains of course will come from Odisha and Bengal.
How this plays out on 19th and 23rd will test many of these hypotheses but there is more than strong evidence that NDA is looking good to return to power in 2019.