Part 2: Are constituencies cast(e) in stone in Bihar? | Number Theory

The second part will take this analysis further back in time

The first part of this series explained the problem of analysing Bihar’s assembly constituencies (ACs) from a historical caste-stronghold perspective and identified 107 out of the 243 current ACs which have elected an MLA from the same sub-caste since 2010. The second part will take this analysis further back in time.

AFP picture

• 
  There are 79 ACs in our database which have never seen an MLA from a different sub-caste

  As described in the first part of this series, ACs do not appear uniformly in the HT database of MLAs in Bihar. This is because successive delimitation exercises have changed the nature of AC boundaries in the state. If one were to count the number of unique AC names which have not elected an MLA from a different sub-caste in Bihar it comes out to 79 or 23% of the total 343 ACs in the database. The largest share of these ACs is in the cohort which appears in the database exactly once. If one were to exclude the 44 ACs which have appeared only once – which leaves 299 ACs which have had at least three elections – the number of such ACs decreases to 35 or 12%. As can be seen from the chart below, the share of same sub-caste-MLA ACs increases as one dilutes the historical longevity threshold for an AC.
• 
  Even in the 151 ACs which appear in all elections, caste concentration is far from insignificant

  While only six out of the 151 ACs (4%) which appear in all 15 elections in the database have elected an MLA from the same sub-caste, this does not negate the extent of long-term caste concentration in Bihar at the AC level. Of the 151 ACs described above, 41 saw their reservation status change during this period, which is likely to have disrupted the nature of caste competition at the AC level. In the remaining 110, the share of ACs which have elected an MLA from the same sub-caste is slightly higher at 5.5%. If one were to gradually dilute the caste-concentration criterion for an AC in this cohort, more than half of them have elected an MLA from the same sub-caste in half of the elections. Controlling for change in reservation status increases this share once more.
• 
  The share of ACs with MLAs from just two sub-castes is even bigger in Bihar

  While tracking ACs with an MLA from the same sub-caste is an extreme way to measure caste dominance at an AC level, looking at ACs which have elected MLAs from just two sub-castes is also a useful way to track the relationship between caste and electoral competition. Of the 151 ACs that appear in all elections, there are 30 ACs (20%) which have had MLAs from just two sub-castes. As is to be expected, the share of such ACs increases as one dilutes the historical threshold criteria or if one looks at only ACs whose reservation status remained unchanged.
• 
  Sub-caste dominance at AC level has shifted away from upper castes to Yadavs in the database

  As can be expected from the analysis above, it is difficult to look at changes over time in the caste composition of strongholds. One way to do this is to look at stronghold ACs in each period of stable AC boundaries: 1967-72 (when three elections were held); 1977-2005 (when eight elections were held); and 2010-2020 (when three elections were held). Expectedly, ACs held by the same sub-caste in all elections is higher in the first period (86 ACs) and the last period (107 ACs) than in the period with eight elections in the middle (27). However, each period also shows a change in the composition of such ACs held by the same sub-caste. Yadavs had the second highest share in strongholds in the first period and the highest share in the next two. However, it is not just the top stronghold position that has changed hands among different sub-castes. Three of the top five positions were held by the Rajputs, Bhumihars, and Brahmins in the first period. This decreased to two of the top five positions (Bhumihars and Rajputs) in the second and third period.