Written by Katharine Adeney.

Most of what we now know as South Asia has always been governed through the concession of territorial autonomy. This was as true under the British Raj as it was under the Mughals.

The area was too religiously, linguistically and territorially diverse for any other solution to have been adopted. In the formal constitutional negotiations in the early twentieth century, federalism was formally adopted as part of the Government of India Acts of 1919 and 1935. Federalism was adopted after independence in the successor states of India and Pakistan, although not without contestation, particularly concerning the boundaries of the federal units, the language(s) that the federation would operate in, the provinces’ representation in governing institutions as well as the powers that they would receive.

This short contribution focuses on the boundaries of the federal units, of the creation of what is known in the federal literature as ‘ethnofederalism’ when the boundaries of at least one of the units of the federation corresponds to those of the group within it. The boundaries of the units of both India and Pakistan at independence bore little correspondence to the various groups that lived within those units.  Before independence, demands had been made for the redrawing of provincial boundaries around group identities. This had been achieved in some cases, such as Sindh and Orissa, but many decisions were deferred until after independence.

Federal Solutions after Independence

After independence however, both India and Pakistan’s new leaders were reticent to undermine national unity through recognising ‘subnational’ identities. This was partially the result of the violence of partition, but both Nehru and Jinnah had favoured a centralised state before independence. Nehru favoured the model of centralised planning, while in Pakistan, the recognition of regionally concentrated language groups potentially undermined the unity of the Muslim homeland. Both leaders were worried that ethnofederalism would lead to the weakening of the centre, and potentially, the Balkanisation of their states. They shared this in common with other critics of ‘ethnofederal’ solutions.

Despite their leaders’ common concerns, India and Pakistan diverged in their constitutional solutions for their diversity. India, after initially recognising the rights of its multilingual provinces to choose their own languages, allowed for the territorial reorganisation of the country into more homogeneous units. The central Congress leadership did so under protest, after elements within the Congress Party vociferously protested against their leaders’ failure to countenance the redrawing of the political map.

The internal borders of India were redrawn in 1956 along de facto linguistic lines, although the States Reorganisation Commission recommended a balanced approach between language, economic viability and administrative convenience. This process of ‘right sizing’, has continued, in India with other tranches of reorganisations in the 1960s, 1970s and 2000s. Most of the later reorganisations were undertaken along non-linguistic lines, for example the tribal recognition of the 1970s or the caste or development narrative of the states created in the 2000s. The process continues, with the creation of the 29th state of the Union – Telangana – in 2014. It is unlikely to be the last.

The case of India demonstrates that demands for “ethnic” provinces, such as in the Seraiki region of Pakistan and in the Madhesi regions of Nepal are likely to increase rather than decrease affinity with the central state.

In Pakistan, a state with a much weaker political leadership after partition, the protracted constitutional negotiations finally (in 1956) came up with a federal formula that reorganised the territorial boundaries of the country. The internal reorganisation was in a different manner to that of India however, and merged all the units and princely states of the western wing of Pakistan into one province: West Pakistan. The so-called One Unit Plan was a device to counterbalance the demographic dominance of its Eastern wing: renamed East Pakistan. Both provinces received equal weighting in the National Assembly, despite the majority of the population (55 percent) of Pakistan residing in the eastern wing.

This constitutional arrangement only lasted two years, with martial law declared by Ayub Khan in 1958. In 1970, his successor, Yahya Khan nullified the One Unit Plan resulting in the restoration of the three western provinces (albeit with their boundaries altered to include the princely states), and the creation of a new one, Baluchistan). The restoration of democracy in Pakistan precipitated its breakup, after the leaders of the western wing refused to recognise the democratic mandate of the politicians of the eastern wing. After the secession of Bangladesh in 1971, although the constitution of Pakistan was rewritten, the opportunity was not taken to reorganise the political map and Pakistan’s federation continued with only four provinces. In 2009 the area of Gilgit Baltistan was given semi-provincial status but has yet to be fully integrated as a fifth province of Pakistan, with representation in the National Assembly.


The legacies of the original decisions were profound.  In the case of India its willingness to continually reorganise its internal boundaries accommodated many groups and enhanced the representativeness of the Indian state. It has also accommodated other demands for linguistic recognition. Thus, not only were provinces (then states) allowed to choose the language(s) that they operated in, the Indian state retained English as an official language in addition to that of Hindi.  This was essential, as many of the states in the south and the northeast of the country did not speak Hindi (spoken by only 30 to 40 percent of the population) and resented the assumption that it was the ‘national’ language of the country.

This accommodative framework, far from leading to the Balkanisation of India, ensured that multiple identities were encouraged to develop.  Evidence for this can be found by reading Moreno surveys on the allegiance of Indians living in different areas of India to their national or regional identity, or a combination of both. Although there are differences between regions, with the South and the East more likely to report feeling ‘regional’ than those in the North or West, there is a clear majority in all regions for feeling either more national than regional or equally national and regional. The inclusion of Indians from all over India in core central institutions, including that of the cabinet has also promoted this unity.

National versus Regional Identity in India 

Onlynational Morenational Equallynational and regional Moreregional and less national Onlyregional NoOpinion
North 41 7 20 5 10 17
East 26 19 15 13 17 10
North-East 20 7 32 9 15 17
West 42 7 27 5 10 10
South 23 22 15 20 13 7
India 32 12 21 10 12 13

Data taken from the State of Democracy in South Asia Survey (2008).

Of course, there are areas of India, particularly in its non-Hindu peripheries, where it has only managed to maintain its territorial integrity through the use of extreme force. There are several reasons for this. First, although India reorganised its units along ostensibly linguistic lines, almost half of the states of India retained significant heterogeneity. In those cases the locally dominant group felt threatened – as witnessed in Assam, Nagaland and Punjab – often leading to the violent targeting of minorities within the units. Second, where democracy or effective autonomy has been undermined, as in the case of Punjab, Jammu and Kashmir and most of the North-eastern states where state governments have been regularly dismissed, tensions with the centre have increased.

It is notable that in Kashmir, where electoral manipulation was commonplace, insurgency did not develop until the late 1980s, after the rigging of the 1987 election. The securitisation of the response from the centre through the use of mechanisms such as the Armed Forces Special Powers Act also increased conflict. Although ten people lost their lives at the hands of police bullets in the Patidar protests in 2015 in Gujarat, the situation is incomparable to the use of pellet guns in Kashmir in 2016. Within six months 100 people were estimated to have been killed and 6000 injured. Therefore, violent conflict cannot be divorced from the fact that these states have seen their effective autonomy being reduced.


In contrast, the unwillingness to make compromises over language alienated many groups from the Pakistani state.  This included Bengalis whose language was not recognised as State Language on par with Urdu until 1954. It was only belatedly accorded this recognition as a quid pro quo for giving up its demand for a majority of seats (to which they were entitled on the basis of their demographic majority) in the National Assembly. Language policy also alienated other groups within Pakistan, notably Sindhis. It was only after the constitutional redrafting in 1973 that provinces in Pakistan were able to choose to operate in a language other than Urdu. This alienation was compounded by exclusion from the core institutions of the state. Bengalis, Sindhis and Balochis all suffered from underrepresentation in institutions such as the army and the bureaucracy. In addition, their provinces suffered from a lack of investment, or, in the case of East Pakistan, under-development, as the resources of the East were extracted to finance the development of the West, particularly that of Punjab province.

National versus Regional Identity in Pakistan

Onlynational Morenational Equallynational & regional Moreregional & less national Onlyregional Noopinion
Urdu 60 8 9 4 16 3
Hindko 51 19 7 9 11 3
Punjabi 47 17 12 4 15 5
Pushto 33 16 17 5 22 7
Seraiki 32 10 13 10 29 6
Sindhi 23 11 8 9 36 14
Balochi 18 18 16 3 17 28
Pakistan 40 14 12 6 20 8

Data taken from the State of Democracy in South Asia Survey (2008).

The refusal to redraw provincial boundaries ensured that Pakistan’s federation exited with a very low number of provinces. This has not only exacerbated conflict between provinces (e.g. the tension between East and West Pakistan) but also meant that the larger provinces in terms of population – East Pakistan before its secession in 1971 and Punjab after 1971 – threatened the other provinces by their demographic majority. In the case of the Punjab, its domination was compounded by the over representation of Punjabis (or sections of Punjabis) in the core institutions of state such as the army and the bureaucracy.

As Henry Hale has argued, ‘ethnofederal states are more likely to collapse when they contain a core ethnic region – a single federal region that enjoys dramatic superiority in population’.  Given that many of the units of the western wing were linguistically heterogeneous, the basis for a reorganisation of provinces along linguistic lines exists – although it must be conceded that parties supporting particular reorganisations (such as those agitating for a Seraiki province (out of Punjab) or a Hindko speaking province (out of Khyber Pakhtunkhwa)) do not receive much electoral support.  In addition, any reorganisation of provinces would have to face up to the thorny issue of the city of Karachi, and demands for it to be separated from the province of Sindh, which would be explosive.

Lessons for Other Federations

Federal (re)design continues apace in the region and elsewhere. In the South Asia region, federal discussions continue in Myanmar and Nepal.  The case of India demonstrates that demands for “ethnic” provinces, such as in the Seraiki region of Pakistan and in the Madhesi regions of Nepal are likely to increase rather than decrease affinity with the central state.  It also prescribes that these territories should be made as homogeneous as possible.  The states of India that have continued to experience violent conflict after territorial reorganisation along ‘ethnic’ lines have been those in which sizeable pockets of diversity remain e.g. Nagaland and Assam. Where such diversity remains, non-territorial power sharing is necessary in addition to territorial models.

However, this comes with a caveat: such autonomy should be part of a wider accommodation of groups within central power structures. Access to central power is important and Pakistan’s failure to include all of its provinces within central power structures has undermined the affinity of many of its groups to the states. In states such as neighbouring Myanmar and Nepal it is important not to pursue a majoritarian-led democratisation. A truly representative democratisation is vital for federations to accommodate territorially concentrated groups successfully.

Katharine Adeney is Professor of Politics at the University of Nottingham and Director of the Institute of Asia and Pacific Studies. This piece was first published on the blog 50 Shades of Federalism and can be found here. She has published widely on both Pakistan and India. She tweets @KatAdeney. Image credit: CC by Pixabay.

Leave a Reply

Your email address will not be published. Required fields are marked *