IT should be obvious that alternative ways of drawing constituency boundaries can significantly influence electoral outcomes. A historical example can make the point: the 2003 redistricting (the term used in the US) in Texas, spanning the 2002 and 2004 elections, changed the composition of its delegation to the US Congress from 15 Republicans and 17 Democrats to 21 Republicans and 11 Democrats.

It is no wonder that redistricting is a hot issue in the US whose fairness has been the subject of repeated supreme court reviews. There the deliberate manipulation of boundaries to influence electoral outcomes, termed gerrymandering, is along two lines — favouring one party over another, as in the case mentioned above, or attempting to reduce the representation of racial minorities.

In this context, it is surprising to find no analysis of past practice in Pakistan, nor much interest now that we are undergoing the process after a gap of nearly 20 years. This could suggest universal agreement on the fairness of the delimitation process in the country. Even so, one should be curious to know if any biases exist in past exercises and how they have evolved under changing demographics over time.

Continuing urbanisation suggests the issue could be important with the distinct possibility that the urban population is underrepresented in the legislature. Historical parallels can be employed again to underscore the relevance. The 1919 election in Germany is considered the first truly fair one because it was the first held after scrapping constituencies that grossly overrepresented rural areas. In India, where about a third of the overall population is recorded as urban, only about 85 of the 543 constituencies for the Lok Sabha, or under 15 per cent, have a majority urban population.

Why is delimitation so contentious in the US and so ignored in Pakistan?

How can such underrepresentation of the urban electorate come about? Simply by ‘splitting’ the urban population into a number of seats most of which have rural majorities. This can be shown easily with a hypothetical example. Imagine a district with a population of four million including a city of 1m and suppose the population per electoral seat is also 1m yielding a total of four seats for the district. Constituency boundaries can be drawn such that there is a constituency with an urban population of 1m and three constituencies with rural populations of a million each. On the other hand, the urban population can be split to yield the following four constituencies (populations in millions with U and R representing urban and rural shares, respectively) (i) 0.3U, 0.7R; (ii) 0.3U, 0.7R; (iii) 0.4U, 0.6R; (iv) 0U, 1R.

The urban population is fairly represented in the first case — 25pc of the population having 25pc of the seats. With the second set, it is completely unrepresented with no seats at all. The actual situation in Pakistan is likely to be one, as in India, where the urban population is considerably underrepresented in the legislature.

Aside from the fact that urban-based political parties have much to lose from dilution of the urban vote, there are other negative consequences of such underrepresentation, if it exists.

First, the Constitution guarantees each citizen a vote of equal value and underrepresentation devalues that of the urban citizen. Second, Election Commission guidelines stipulate that constituencies be demarcated such that homogeneity of the community is ensured. Urban and rural communities are, however, very heterogeneous and one can expect a representative dependent on a rural majority to neglect the interests of the rump of urban voters in his/her constituency.

It can be inferred from this that unless cities and towns acquire a political voice commensurate with their numbers they will lack the attention they need to serve their residents nor get the resource allocations needed for national development. The latter is relevant since almost three-fourths of gross domestic output of the country now emanates from urban areas.

Over the years observers have noted the persistent dominance of ‘feudals’ in legislatures, the term used loosely to denote members of notable families repeatedly elected on the basis of dependent clienteles that are much more a feature of rural than urban demographics. Since such rural clienteles are easier to control it is natural that the beneficiaries would not want the status quo to change in their constituencies. It is, therefore, reasonable to speculate that medium and small urban centres would be split almost entirely into constituencies with rural majorities, a speculation supported by their condition. Only a rigorous study can provide the evidence for a correction like the one that marked the beginning of fair representation in Germany.

It is also of interest to consider why delimitation or redistricting is so contentious in the US and so ignored in Pakistan. It could be because there are easier alternatives available to the establishment and political parties to influence electoral outcomes in Pakistan — these include, rigging, bribing, inducing military takeovers, and outright dismissals of governments. No such measures are available to political parties in the US forcing them to rely on indirect methods like redistricting and the electoral college. It is, therefore, not a surprise that in the US the redistricting process has been retained under the political control of state legislatures while most other countries, including Pakistan, have transferred it to the jurisdiction of neutral commissions.

This last observation raises a related issue meriting attention in Pakistan. Election laws stipulate that constituencies should preferably lie within district boundaries which means that creating new districts perforce necessitates delimitation. Since creating new districts is a political prerogative in Pakistan one can speculate that it could have had underlying electoral imperatives. A retrospective study could test this hypothesis since the stated rationale of better governance advanced for the creation of new districts cannot bear the weight of objective evidence.

An analytical exercise seems warranted with the objective of ensuring that election outcomes reflect the popular will and that the preferences of voters are translated faithfully into policy outcomes. Both these are dependent on unbiased representation.

The writer was dean of the School of Humanities and Social Sciences at Lums.

Published in Dawn, January 20th, 2018