Opposition to Ethnic Territorial Autonomy in the Indian Parliament

ABSTRACT

Why do central governments resist regional ethnic autonomy? Is it nationalism? A fear of lost revenues? Scholars rarely investigate these questions with fine-grained data on elite behavior. The author analyzes new data on Indian parliamentarians during that country’s federal reorganization in the 1950s. Legislators debated dozens of proposals for ethnic autonomy. What kinds of proposals met the most opposition? Contrary to existing theory, central legislators were not opposed to autonomy for revenue- or resource-rich areas. Religious and linguistic nationalism were influential but they did not create a direct relationship between ethnic differences and MPs’ stances on autonomy. Regional ethnic rivalries were a key driver of opposition to autonomy. Parliamentarians rebuffed proposals that threatened to undermine their constituents’ position relative to neighboring non-coethnics. These findings suggest new insights for the comparative study of ethnoterritorial politics.

Supplemental Data

Supplemental data for this article can be accessed at https://doi.org/10.1080/17449057.2020.1734900.

Notes

1 ‘Tribes’ are hereditary groups. ‘Caste’ refers to two concepts. Varnas are tiers in the ritual hierarchy and not specific to an ethnolinguistic group. Jatis are ethnicity-specific. Government reservations are based on jatis. Both tribal and caste schedules can be tailored to favor particular ethnicities.

2 The SRC’s mandate did not extend to Jammu and Kashmir.

3 Eastern and Western Hindi do not have a common origin (Shapiro, 2003; Singh, 2000).

4 Summary statistics for all variables are in Table A1.

5 Information on MPs from Election Commission of India (2015).

6 Not taking a stance on a particular autonomy demand is similar to an abstention in roll call data. Poole and Rosenthal (1997) suggest abstention is correlated with indifference and uncontroversial issues. MPs who spoke in the Lok Sabha had a (loosely enforced) time limit, making it likely that not speaking on a particular autonomy proposal indicates indifference. My results are similar if all non-mentions are coded as cases of non-opposition (Table A8), except that MPs were especially unlikely to mention out-of-region autonomy proposals compared to in-region proposals.

7 Based on Central Statistics Office (1951) and Election Commission of India (1951).

8 Based on Central Statistics Office (1951); Grierson (1903); and Office of the Registrar General and Census Commissioner (2004).

9 Slated to join under the SRC’s recommendations or a slightly modified version of those recommendations circulated by the Congress leadership prior to the Lok Sabha’s debate. Given the Congress parliamentary majority, either plan might plausibly have been enacted. MPs needed to take the implications for their constituency into account.

10 Results are similar if these categories are kept separate (Table A14).

11 Cross-tabulations of key independent variables and MP opposition to autonomy are displayed in Tables A3–A6. Appendix Table A7 reports models using independent variables separately and without controls.

12 I present regressions without MP fixed effects in Table A16.

13 Appendix Tables A15 and A16 add another proposal characteristic: whether the SRC recommended self-rule. That variable is tested alone and in interaction with membership in the Congress party. Including an SRC recommendation variable could disguise the importance of factors that the SRC and MPs considered in forming their opinions. For example, after including the SRC variable, the natural resources dummy has a positive and insignificant correlation with MP opposition to autonomy. The non-significant result suggests the SRC and MPs had similar preferences with regard to resource rich areas.

14 Hindu right parties were less electorally important in India in the 1950s than they are today and had relatively few MPs. Hindu right MPs were more likely than other MPs to oppose religious minority autonomy. They were no more or less likely to oppose minority language proposals (Tables A15 and A16).

15 In Table A8 MPs’ non-remarks are coded as lack of opposition. Economic development has a positive and statistically insignificant coefficient.

16 See pages A.19–A.25. There was no pro-autonomy violence in a Hindi-dominated area prior to the 1955 debates. Thus, it is not possible to model an interaction between language and violence.

17 E.g. Parliament of India (1955, pp. 63–64).