Do Pakistanis vote on an ethnic basis?

The question of regionalism is as old as the 1940 Lahore resolution where Pakistan was presented as a collection of “sovereign” territorial units or regions. The subsequent episodes of the country’s story are well-known, including the 1971 second partition. More than 40 years later, the coming general election offers a great opportunity to assess the resilience, or otherwise, of regional political cultures.

Pakistaniyat, or the sentiment of “Pakistan-ness”, has made significant progress in the course of time. A major survey published in 2008 (but conducted in 2004-2005), entitled The State of Democracy in South Asia, showed that Pakistanis were proud of Pakistan wherever they lived (the interviewees of rural Sindh were the “least” proud – with 69 per cent – whereas the inhabitants of northern and central Punjab topped the list with 96 per cent).

This is a clear reflection of the impact of state’s policies. After all, nations are also shaped by their states, and national political cultures are influenced by the delimitation of institutional arenas. Pakistan being the framework for competition between elite groups, these groups have had to put their regional identities on the back-burner in order to project themselves as national.

This transition has taken place when the groups in question have achieved power. Sindhi nationalism is a case in point: it has declined after Pakistan Peoples Party (PPP) emerged as a vehicle for the Sindhis’ upward political mobility. Why should the Sindhis' ask for more autonomy when they could have access to power at the Centre?

The Punjabis had already “Pakistanised” themselves according to the same logic: why should they retain a separate regional identity when they could dominate the whole country? In The State of Democracy in South Asia survey, 64 per cent of the Punjabi respondents declared that they felt only or predominantly Pakistani whereas 19 per cent felt only or predominantly Punjabi. Their role in the army partly explains this commitment to nationalism.

At the same time, marginalisation does not necessarily lead to regionalism, as is evident from the case of the Mohajirs. In the same survey, Urdu speakers top the list of those who say that they feel “only” Pakistanis at 60 per cent. This identification with the nation does not reflect a sentiment of domination of the state – that has been replaced by a sentiment of victimisation since the 1970s – but the feeling that Pakistan is a creation of the Mohajirs (who cannot identify with any other place anyway).

Yet, in most cases, the smaller (and the more marginalised) an ethnic group is, the more inclined towards regionalism it is. But each group includes elite groups which have been co-opted by the national establishment (see how the numbers of Pakhtuns has increased in the army since Ayub Khan). Hence, the split verdict that we find in the survey cited above: while 33 per cent of the Pashto-speaking interviewees felt themselves to be “only national”, 22 per cent felt “only regional”.

These figures were, respectively, 18 per cent and 17 per cent in the case of the Baloch, 32 per cent and 29 per cent in the case of the Seraikis and 23 per cent and 36 per cent in the case of the Sindhis.

The resilience of ethnic identities is consistent with different political cultures. In The State of Democracy in South Asia survey, only 21 per cent of the Punjabis considered that democracy was preferable to any other regime, against 35 per cent of the Sindhis, 32 per cent of the Pakhtuns and 29 per cent of the Mohajirs. But as many Punjabis and Baloch – 34 per cent – thought (for different reasons) that democratic and military regimes made no difference whatsoever.

Interestingly, democracy did not mean the same thing for every ethnic group. For the Baloch it meant primarily “peace and security”, for the Sindhis, “justice and welfare”, and for the Mohajirs “freedom”. But for none of these groups (except 3 per cent of the Sindhis) it meant “election”. This is exactly the conclusion one draws from the evolution of voter turnout in recent years. In spite of a positive trend in 2008 (in tune with 2002 figures), it has remained 10 percentage points below what it was in 1977.

Will the coming election make a significant change in terms of turnout? Does the main challenger of this election, Imran Khan, have the potential to become a game-changer in this respect at least? Although, relying upon a smaller sample than the aforementioned 2008 survey, the pre-election survey conducted by the Herald in March 2013 indicates that, even if it will not deliver the “tsunami of change” that it promised, Pakistan Tehreek-e-Insaaf (PTI) has already started upsetting Pakistani politics as we knew it. The voting intentions of the respondents suggest that PTI is currently the only party – besides PPP– which finds significant support across the ethnic spectrum.

Is PTI all set to become Pakistan’s next truly multi-ethnic party? Yes and No. On the one hand, like PPP (but to a lesser extent), PTI has its core constituency in one particular ethnic group, namely, the Pakhtuns (with 31.40 per cent of the registered Pakhtun voters interviewed for the Herald survey declaring their intention to vote for PTI).

On the other hand, though, the PTI also seems to have made a dent into the Punjabi electorate (30.80 per cent of voting intentions), among the Seraikis (23.49 per cent of voting intentions) and to a lesser extent among the Baloch (11 per cent of voting intentions). Even the Mohajirs – the paradigmatic example of ethnic voters in Pakistan – seem to have succumbed to Khan’s charm (16.67 per cent of voting intentions), although the limited number of registered Urdu-speaking voters interviewed for this survey makes any generalisation hazardous.

Only the Sindhis seem to have remained unaffected by the Khan factor (nine per cent of voting intentions), with more than half of the Sindhi respondents remaining loyal to PPP. And if we exclude the limited sample of Hindko-speaking registered voters, PTI is the only major party which does not register more than 40 per cent voting intentions in any ethnic group, a characteristic that makes it less “ethnic” than PPP.

Could it then be that ethnicity is becoming increasingly irrelevant in Pakistani electoral politics? Once again, the answer should be nuanced. First of all, the Herald survey suggests that all political parties retain a core ethnic constituency, while two parties – PPP and Muttahida Qaumi Movement (MQM) – register 50 per cent of voting intentions or more in one specific ethnic group.

Secondly, as suggested above, ethnicity seems to remain an important ingredient in shaping political cultures in Pakistan, although this is more obvious for meta-political issues (attitudes towards the institutional set-up of the state of Pakistan at large and the distribution of power within it) than for socio-economic issues. (Pakistani voters across the ethnic spectrum seem to be primarily concerned with the same issues: poverty, corruption and the power crisis).

Thirdly, ethnicity is far from being the only variable shaping the vote of Pakistanis. The Herald survey provides interesting insights in this regard, by suggesting that the vote for certain parties is over-determined by age. Thus, Pakistan Muslim League-Nawaz (PMLN) would be the “oldest” political party in Pakistan, with 46.2 per cent of the registered respondents aged 70 and above declaring their intention to vote for it.

However, the influence of the age variable seems to be less significant at the other side of the generational spectrum and no political party (even PTI, which is surprising) registers more than 28 per cent of voting intentions among the 1835 age group (although, on the whole, PTI’s vote share seems inversely correlated to the age of its voters).

Ethnicity seems to remain an important ingredient in shaping politicalcultures in Pakistan, although this is more obvious for meta-politicalissues (attitudes towards the institutional set-up of the state ofPakistan at large and the distribution of power within it) than forsocio-economic issues.

The urban/rural divide, unsurprisingly, also continues to make its influence felt and the Herald survey suggests that while some parties have a significantly more rural base (PPP, PMLN, Awami National Party), others remain predominantly urban-based (MQM and to a lesser extent PTI).

Class positions remain equally significant. PPP registers the highest share of voting intentions among the poorest voters (with Rs 3600 to Rs 10,000 monthly family income), followed by PMLN. This gap decreases among the lower-middle class and middle-class voters, although PPP still registers the largest share of voting intentions among the Rs 15,001-Rs 30,000 income groups.

MQM, for its part, lives up to its reputation as a lower middle-class/middle-class party, with 43 per cent of its voting intentions coming from respondents with a monthly income between Rs 15,000 and Rs 45,000.

However, this phenomenon probably has less to do with the ideology of MQM than with the sociology of the electorate of ethnic/nationalist parties — the same trend can be observed for Balochistan National Party-Mengal (BNPM), for instance. This is less true for ANP, though, which seems to have a stronger support base among poor Pakhtuns. PTI, for its part, could be a major challenge for PMLN as far as the middle/upper class vote is concerned. While PMLN had the highest share of voters from the middle and upper class in 2008, PTI registers the highest share of voting intentions among all the Rs 30,000+ income groups.

The level of education seems to play a more complex role. It does affect some parties — voting intentions for PPP, for instance, are inversely correlated to the level of education, at least until the masters level, while the vast majority of PTI sympathisers seem to come from the educated classes (77 per cent of the respondents declaring their intention to vote for the party have done matriculation or above).

PMLN seems to have a more balanced ratio of highly educated/less educated voters. Interestingly, the level of education seems highly and positively correlated with the vote for nationalist/ethnic parties, such as BNPM and to a lesser extent MQM and ANP (with respectively 82 per cent, 71 per cent and 66 per cent of the respondents declaring their intention to vote for these parties having done matriculation or above).

Finally, the influence of gender remains quite limited, although some significant discrepancies seem to prevail in the case of two parties: PPP (which, proportionately at least, would find larger support among women than men) and PTI (so much for Khan’s charm: only 18.3 per cent of registered women voters declared their intention to vote for PTI, against 21.8 per cent of registered men).

It is worth mentioning that the coming elections are taking place in the context of an unprecedented recognition of regionalisms. While the federal dimension of the 1973 Constitution had been repeatedly curtailed under military – and civilian – regimes, including during the rule of General (retd) Pervez Musharraf when the local and district governments were promoted at the expense of provincial governments – the 18th amendment has given the provinces more autonomy than ever before.

Not only have the old bureaucratic names of two provinces been replaced by others — the Northern Territories becoming Gilgit Baltistan and NWFP becoming Khyber Pakhtunkhwa — but the 18th amendment has also given new prerogatives to the provinces. In the original version of the 1973 Constitution, there were two lists, the Federal Legislative List which comprised of 67 domains and the Concurrent List which comprised of 47 subjects about which the Centre could legislate.

The latter has been abolished and its contents distributed between the Center and the provinces which have gained a lot of power as a result of this distribution.

Clearly, political centralisation – an old obsession that had resulted in the 1950s from the fear of India (and its instrumentalisation by the ruling elites) – is not legitimate any more. In this context, political parties have articulated regionalist agendas during the ongoing political campaign.

PPP, for instance, has promised to create a new Seraiki province if it is brought back to power, claiming that it was prevented from doing so until now due to the opposition of “anti-Seraiki forces”. The party seems concerned with being rooted out from south Punjab, probably less for its inability to deliver on its promises to create a Seraiki province than for the failure of the federal government in addressing the concerns of the residents of the districts of south Punjab most affected by the 2010 floods (Dera Ghazi Khan, Rajanpur, Muzafargar). PMLN, for its part, has committed itself to declare the four provincial languages as national languages if voted to power.

However, it continues to oppose the creation of a new Seraiki province and its top leadership supports, instead, the restoration of Bahawalpur province, which was abolished with the implementation of the One Unit system in 1955.

This project could become a bone of contention within the party, though, as some Punjabi leaders of PMLN remain opposed to any division of Punjab. Among the smaller parties, MQM supports the creation of new provinces in Punjab and Khyber Pakhtunkhwa, claiming that this will bring greater stability to the state of Pakistan.

However, its electoral manifesto does not make any mention of this issue and advocates, instead, “a national dialogue and consensus” on the controversial issue of “full provincial autonomy” in the light of the 18th amendment. The position of PTI on these issues is even more cautious. As he can reasonably expect his party to make a dent into south Punjab, Khan has carefully avoided taking a stand on the issue of new provinces, especially on the restoration of Bahawalpur. At the same time, though, PTI entered into a seat adjustment agreement with the Nawab of Bahawalpur, in order to benefit from the vote bank of this undeterred promoter of the Bahawalpur province.

By requesting political candidates to refrain from seeking votes “on the basis of religion, sect, caste or ethnicity” and threatening violators with a three-year jail term, the Electoral Commission of Pakistan recently confirmed the persisting relevance of ethnic bonds in shaping electoral choices. Yet, while most political parties aim at cashing in on this elusive “ethnic vote”, their leadership remains aware that ethnicity is only one factor among others shaping the decisions of the Pakistani electorate.

Regrettably, the pre-election survey of the Herald did not factor in the role of another crucial variable in this regard, at least in rural areas: biradari/clan affiliations. But as the experts engaged by the Herald in another survey (see Inexpert language) emphasise, these bonds and the vote banks that they sustain remain determinant come election day. This is particularly true in Punjab, where the fate of the next government will be sealed.

All in all, the next elections could see a resurgence of national political forces to the detriment of regionalist ones (this is the case, in particular, in Khyber Pakhtunkhwa where PTI could be a serious challenge for ANP). Nevertheless, in some parts of the country, ethnic parties will probably reassert themselves (BNPM in Balochistan) or at least save the day for themselves (MQM in Karachi).

Both in Balochistan and Karachi, this persisting relevance of nationalist/ethnic forces is linked to a political context characterised by escalating armed confrontations. The 2013 elections should, therefore, confirm that, in Pakistan as elsewhere, the political relevance of ethnicity is the by-product of political conflicts and socio-economic struggles, rather than the opposite. In the same logic, even the vote for PTI in Khyber Pakhtunkhwa will partake in an ethnic colour, as war-weary Pakhtuns will be tempted to give peace a new chance, Imran-style.

This article was originally published in the May 2013 issue of the Herald under the headline 'Mosaic or muddle'. To read more subscribe to the Herald in print.