It is common knowledge that in Uttar Pradesh there is very high correlation between certain parties and the caste of their vote base. For example, as a group, Dalits mostly vote BSP in the state. Their next choice is possibly the INC. Their third is likely to be BJP and they will then probably vote anyone else but a Yadav dominated SP. Similarly, the Muslims are likely to vote SP or INC as their first two choices, followed by others which may include BSP. Just as the Dalits are unlikely to vote SP, the Muslims are unlikely to vote BJP. This is something that almost all politicians intuitively “get”.
The other part of this equation is the relative strength of each formation. Uttar Pradesh has about 22% of its population as Dalits. Muslims form about 18% while Yadavs are estimated to form 9% of the population. Together that’s 49%; if one were a Data Analyst for the BJP, that’s a staggering mountain to climb. The party is essentially reduced to competing in the 51% bloc of the population where it isn’t relegated to third choice or less. This section of the population is mostly made up of non-Yadav OBCs and upper castes. While different upper castes usually don’t have as much of a problem voting together given their relative position in the social order, many OBC groups are often vying for the same position in a zero-sum game and hence rarely vote together. That puts some part of this 51% bloc beyond the reach of any single party.
Given that the SP and BSP have been the dominant players in the state, they must have some ability to gain the support of some sub-sections of this group — though in a far less dominant way compared to their own base — forming one part of the game discussed above. So, let’s assume the SP and BSP, put together, can lay claim to 20-25% of this 51% bloc which they must have won to have formed governments in the past. Maybe the SP managed to find some non-Yadav OBCs who’re not in direct conflict with the Yadavs; or the BSP did the same among upper castes. However they did that, that basic arithmetic leaves 38% of the population in a contest between BJP, smaller OBC parties and the INC.To reach the winning threshold of 25-30% total vote share in UP therefore, it simply follows that the BJP has to have a strike rate of over 75% in this section of disparate caste formations. That’s such a tall order that it’s practically impossible except under extreme situations where the identity of the caste is subsumed by a bigger threat.
A trick that any pragmatic BJP politician who is interested in winning more than ethical considerations of propriety would therefore try will be to consolidate this 51% bloc by any and all means. It also does not hurt the SP to play this game given most of these votes might anyway not be coming their way while the Muslims consolidate in its favor from the other end. The parties to lose significantly in this conflict maybe the INC and the BSP and even they may pick a side cleverly to cut losses and thereby exacerbating the social situation. In other words, the structure of the first past the post in a deeply caste-ridden society with a four cornered contest has made engineering conflict a first choice tool of consolidation. It’s just that BJP and SP now stand to gain from this variety of conflict — it’s equally likely that the BSP will be tempted to foment OBC-Dalit conflict or the INC an intra OBC conflict.
While appeals to tolerant and liberal values or long term self preservation are an option, they are still unlikely to deter a politician who has short term electoral victory as the motive from trying to engineer conflict. Therefore an easier solution is to look for a system where it isn’t sufficient if 25% of the population vote for a party while an actual majority actively dislike the party.
There are two ready options. The first is Alternate Voting. Where you as a voter are allowed to simply rank your choices in the order of preference. You can choose to just vote one party in the first preference and no one else at all or vote every one contesting in the order of your preference or something in-between: where a Muslim in UP probably votes SP as first choice and INC as second and then stops. Once votes are cast, for counting, the following algorithm is applied.
1. Count all first-preference votes not yet counted.
2. If some candidate has over 50% of first-preference votes, then HALT. (That candidate then wins.)
3. Take the candidate with the smallest number of first-preference votes and change each vote for that candidate by removing the first-preference votes and turning kth-preference votes into (k-1)th-preference votes for each k.
4. Remove that candidate from the list of preferences of all other voters.
5. GOTO 1
The benefit of this method is simply that hatred is not sufficient to win. Let’s say the intended result of the current conflict in UP is achieved. Then all Muslims vote their first choice as SP. But they will possibly also vote INC as second choice. Similarly the upper castes and non-Yadav OBCs who may vote BJP as first choice under this circumstance may end up voting INC as second. In which case the two main protagonists of the conflict end up not reaping the rewards of hatred but the third party may indeed win. This scenario is of course being charitable to the INC; an equally likely INC instigated intra-OBC conflict may end up helping the BSP for instance. Point being those who aren’t affected by the hatred directly will have a say even if it isn’t their primary concern.
Another important advantage of this method is that there is no explicit tactical voting required on the part of the voter. For example, Muslims in UP often vote SP or INC and base that on their judgement of who is likely to win; given they don’t want the BJP to win. Similarly, Brahmins may choose between BJP, INC or even BSP to keep the SP out. The alternate voting simply takes care of that by allowing the voter to rank choices. Even if a Muslim in UP wants to vote INC but thinks the INC may not win, then voting INC in first choice followed by SP achieves the intended result of helping the SP win and not the BJP. Same is the case with the Brahmin who wants to vote BJP.
A simpler second option maybe to give everyone the choice of voting for as many candidates as they like. Here there is no need for an algorithm to count — the candidate with the maximum votes, wins. This also achieves almost all the advantages of the alternate voting and eliminates the need to rank or for a complicated counting process.
 – Algorithm as used in the British referendum that was defeated.