Title: The Changing Pakistan Army Officer Corps

Authors: C. Christine Fair and Shuja Nawaz

C. Christine Fair
Assistant Professor at Georgetown University

Shuja Nawaz
Director of the Atlantic Council's South Asia Center

Title: The Changing Pakistan Army Officer Corps1

Abstract: The Pakistan army elicits many concerns about terrorism, nuclear
proliferation and the coherence of the state. However, very little is actually known
about this institution. This essay mobilizes unique data to address one important
facet: the army’s geographical recruitment base. We find that the Pakistan army
has been successful at expanding the geographical recruitment base while some
groups (namely those who are native to Sindh) remain highly under-represented.
We also find that the officer corps is increasingly coming from urban areas. We
conclude with a discussion of the implications of these important shifts subject to
the limitations of our data.

1. Introduction

The Pakistan army has governed Pakistan directly through military rule for more than
half of Pakistan’s history since it became independent in 1947. The army has governed
indirectly for the remainder by manipulating and undermining civilian political parties
and other institutions and through pro-active manipulations of domestic political
developments. The army has long justified its preeminent role in the state by arguing that
it alone is the sole institution capable of protecting Pakistan against external threats posed
by India and against internal threats posed by centripetal ethnic fissures and sectarian

1 The authors thank Stephen P. Cohen, John H. (aka Jack) Gill, David O. Smith, and Ashley J.
Tellis for reading earlier versions of this work. The authors are also thankful to Lisa Myashiro, who did our
GIS analyses; Aine McCarthy, who performed data entry; and Shiv Sarin who helped create needed
district-level crosswalks. Despite the guidance and assistance of these persons, the authors are alone
responsible for any errors that persist.

7/12/2010 1
conflict. With the exception of notable and extraordinary periods in Pakistan’s history,
Pakistanis accept this role and hold the army in high esteem.2
Yet the Pakistan army discomfits many analysts and policymakers outside of
Pakistan for several inter-related reasons. First, the Pakistan army is the custodian of the
country’s nuclear weapons program with a demonstrable nuclear proliferation track
record. The notorious activities of A.Q. Khan since the 1970s underscore Pakistan’s
intent to engage in subterfuge to acquire strategic capabilities. Pakistan’s relationships
with Iran, China and North Korea have been characterized by extensive nuclear and/or
missile technology proliferation with presumed army sanction. Second, along with the
intelligence agency it oversees—the Inter-Services Intelligence Directorate (ISI)—it has
raised, nurtured and employed Islamist militants to secure its interests vis-à-vis India and
Afghanistan since 1947 with varying intensity. Third, the Pakistan army plays a pivotal—
but controversial and highly criticized—role in the U.S.-led efforts to contend with the
threats posed by al Qaeda and the Afghan Taliban.3
Fourth, in recent years, the Pakistan state has come under direct attack from
domestic militants operating under the moniker “Tehreek-e-Taliban-e-Pakistan” (a.k.a.
“The Pakistani Taliban” (TTP).4 Elements of the TTP have attacked military,
paramilitary, police and other governance targets; perpetrated a suicide bombing
campaign against government entities and citizens alike; and seized pockets of territory in

2 Following the 1971 war when the Pakistan army lost Bangladesh, antipathy towards the army was
prevalent. In 2007, Pakistani regard for the army again plummeted owing to President Musharraf’s
unpopular policies, including alignment with the U.S.-led global war on terror, the military operations
against domestic militants, and failed domestic policies with respect to energy and the economy among
other concerns. Since then, the army has recuperated its standing. See International Republic Institute,
‘IRI Index: Pakistan Public Opinion Survey July 15- Aug 7,’ October 1, 2009. Available at
http://www.iri.org/mena/pakistan/pdfs/2009_October_1_Survey_of_Pakistan_Public_Opinion_July_15-
August_7_2009.pdf.
3 See U.S. Government Accountability Office (GAO), ‘Combating Terrorism: U.S. Oversight of
Pakistan Reimbursement Claims for Coalition Support Funds,’ (Washington D.C.: GAO, 2008); John D.
Negroponte, Deputy Secretary of State, ‘Pakistan’s Fata Challenge: Securing One of the World’s Most
Dangerous Areas,’ Testimony before the Senate Foreign Relations Committee, May 20, 2008; Peter
Beaumont and Mark Townsend, ‘Pakistan troops ‘aid Taliban’: New classified US documents reveal that
mass infiltration of Frontier Corps by Afghan insurgents is helping latest offensive,’ The Observer, June
22, 2008, http://www.guardian.co.uk/world/2008/jun/22/pakistan.afghanistan; Ann Scott Tyson, ‘A Sober
Assessment of Afghanistan: Outgoing U.S. Commander Cites 50% Spike in Attacks in East,’ Washington
Post, June 15, 2008;
http://www.washingtonpost.com/wpyn/content/article/2008/06/14/AR2008061401639.html.
4 Hassan Abbas, ‘A Profile of Tehrik-i-Taliban Pakistan’, CTC Sentinel, 1/ 2 (January 2008).

7/12/2010 2
the Pashtun belt and set up parallel structures of government. These events have
precipitated further concerns, with varying degrees of validity, about the stability of the
state and its strategic assets. These concerns have been mitigated in some measure by the
army’s redoubled military efforts in 2009 to regain lost territory.5 However, the
international community is not confident that the army will succeed against the TTP
much less commit to tackling the Afghan Taliban and turning its guns against its
erstwhile proxies operating against India.
These facts motivate many apprehensions about the Pakistan army. Some
analysts and policy makers have speculated that the army may split, imperiling the
territorial integrity of the state as well as the security of Pakistan’s nuclear weapons.
Variants of this uneasiness involve the army cleaving with Islamist radicals forming, a
dangerous rump that supports Islamist terrorists and may provide nuclear weapons or
technology to terrorists. More plausibly, the Pakistan army, with its long association with
militant proxies, may be unwilling to act against militant groups comprehensively. This
concern remains even though the army has taken on those particular militant groups that
have attacked the Pakistani state. Speculation about the impact of the army’s sustained
internal security operations upon the morale and discipline of the army are cause for
further unease.6
The army is also believed to be at the bottom of significant internal discord.
Speculation is rife that the army is Punjabi dominated and seeks to protect Punjabi ethnic
equities. 7 The presumed ethnic imbalance is posited as a source of ethnic strife in
Pakistan.8

5 C. Christine Fair and Seth G. Jones, ‘Pakistan’s War Within’, Survival, 51/6 (December 2009-
January 2010) 161-188.
6 Russ Wallen, ‘Keeping Pakistan’s Nukes Extremist-Free,’ Asia Times Online, June 16, 2009.
http://www.atimes.com/atimes/South_Asia/KA16Df01.html; Shaun Gregory, ‘The Terrorist Threat to
Pakistan’s Nuclear Weapons’, CTC Sentinel, 2/7 (July 2009).
7 Mushahid Hussain, ‘Reforming the Armed Forces,’ June 10, 2001. Originally published in The
Nation. Available at tp://chowk.com/articles/5060. Selig S. Harrison, ‘Pakistan’s Ethnic Fault Lines,’ The
Washington Post, May 11, 2009. http://www.washingtonpost.com/wp-
dyn/content/article/2009/05/10/AR2009051001959.html.
8 Selig S. Harrison, ‘Drawn and Quartered?’ New York Times, February 1, 2008.
http://www.nytimes.com/2008/02/01/opinion/01harrison.html?_r=1&scp=1&sq=Selig+S.+Harrison&st=ny
t.

7/12/2010 3
 Unfortunately, there are few sources of empirical data about the Pakistan army
despite these varied concerns about the army and its role in Pakistan’s security and that of
the region. This essay, employing unique and difficult to obtain district-level officer
recruitment data from 1971 to 2005, explores an important aspect of the institution: its
geographical representativeness over time. These data and this analysis cannot answer
the most pressing questions about this institution. While these data may be the best
available, they are far from ideal. These data are district-level; not officer-level.
Therefore we can only describe changes in the districts that produce officers over time.
We cannot infer anything about the characteristics of the officers themselves. Despite
these debilitating data limitations, this analysis empirically documents important changes
in the recruitment base of the institution and lays out a number of implications that may
stem from these changes.
The next section of this essay lays out important methodological notes and
describes data limitations and analytical caveats. The third section provides a description
of the recruitment process because this essay is fundamentally about officer recruitment.
Because the current composition of the Pakistan army derives from past recruitment
practices, we discuss the historical lineaments of Pakistan army recruitment and their
enduring impacts upon current army composition in the fourth section. This section also
describes past and current army efforts to reshape its composition. As will be evident in
this section, the army’s composition is not driven by a predatory Punjabi drive to
dominate the country. The penultimate section presents the results of our data analyses
which demonstrate considerable change in the geographic representation in the Pakistan
army. This essay concludes with a discussion of the implications of the changes we have
observed, subject to the limitations of our data.

2. Data, Data Handling, and Analytical Caveats

Shuja Nawaz acquired these data in hard copy, in spreadsheet form from the Pakistan
Army’s General Headquarters in Rawalpindi as part of a historical account of the
Pakistan army.9 A critical—but untestable—assumption is that these data reflect the
army’s efforts to collect and present authentic data. This spreadsheet enumerates the

9 Shuja Nawaz, Crossed Swords (Oxford: OUP, 2008).

7/12/2010 4
annual number of officers recruited from specific districts between 1970 and 2005.10 A
fundamental but untestable assumption undergirding this study is that the data are
genuine and reflect actual army records.11 The team cannot independently verify the
authenticity of the data. However, general features of the data do accord with the general
literature on the Pakistan army. Specific data integrity problems are addressed where
appropriate herein.12
Because the Pakistan army data occasionally contained non-standard district
identification, misidentified provinces for a small number of districts and because
districts have changed over time due to reorganization and population growth, the project
team devised a series of rules to clean and reorganize the data to ensure maximum
accuracy and comparability across time and across geographical areas.13
This reorganization allowed us to merge the recruiting data with GIS (Geographic
Information System) codes for the boundaries of Pakistan’s districts. (This file detailing
GIS district boundaries is referred to as a “shape file.” This file is from Princeton
University’s data library and dates to 1995.) Using this shape file, we depicted shares of
recruitment outcomes cartographically.
Finally, because the army provided numbers of officers recruited from particular
districts per year and because the total number of officers recruited each year varied, we
converted district figures to “market shares.” (We use this terminology because
recruitment is a market and we want to know the overall share of district recruitment as
relative to the overall number of annual recruits.) The share of any given district is
defined by dividing the number of recruits for any given district by the total number of
recruits for the year. (Figures 1 through 4 show the cartographic depiction of shares by
district for 1974, 1984, 1994 and 2005.)

10 Districts are the third layer of administration in Pakistan. The most central level is the federal
government, followed by provincial and finally district governments.
11 Nawaz, Crossed Swords.
12 Two persons independently entered the data into Excel spreadsheets. The team’s quantitative
analyst converted these two spreadsheets into Stata (a statistical program) and electronically compared the
two independently entered datasets for officer recruitment data. This allowed us to identify and correct data
entry errors in the spreadsheets to render them identical to each other and to the hard copy originals. The
raw data contained many errors in spelling, districts were sometimes put under incorrect provinces, or
(rarely) the Pakistan army used a geographical area that we could not identify in any census or district
report. The raw data file only corrects spelling.
13 For more information about data handling and data cleaning, please contact the authors.

7/12/2010 5
 Because we have district-level, not officer-level data, we can only characterize the
districts’ characteristics and not those of the officers. It cannot be facilely assumed that
any given officer has the same characteristics of his (or less likely her) district because
the army engages in a selection process and because officers with a propensity for
military service may differ in fundamental ways from other residents of the district. This
is an “ecological fallacy” problem that inheres in such a study of characteristics of
geographical areas of recruitment instead of individual characteristics. While officer-
level data is ideal for understanding changes in the officer corps, in reality such data are
unlikely to ever be available to researchers inside or outside of Pakistan. Thus these data
are likely to be the best data that exist on the recruitment base of the army. Finally,
because we only have annual intake data, we can only describe changes in the yearly
intake of officers, not the composition of the entire army.

3. Overview of Army Recruitment Practices

Since this essay examines historical district-level data on officer intake into the Pakistan
Military Academy (PMA) at Kakul some description of both the PMA and the
recruitment process is appropriate.
Admission to the PMA is very competitive: there are roughly 3,000 applicants
nationwide for the roughly 320 cadet places in each regular long course in the PMA.14
Normally the PMA conducts two long courses each year. One cohort is inducted in the
spring and another in the fall. The PMA normally has five courses running
simultaneously with the regular, two-year course being the longest.15 After completing
the two-year long course the officer cadets are commissioned with the rank of second
lieutenant and are assigned to a particular arm of service and a regiment or battalion.
To qualify, candidates must be single and hold at last an intermediate degree (i.e.,
12 years of schooling) and be between 17 and 22 years of age. 16 Candidates with a
bachelor degree (aka “graduates” as they are known in Pakistan) must be between 17 and

14 Brigadier (now Major General) Khawar Hanif, Defence Attache, Pakistan Embassy, 2007.
15 The graduate course and the technical graduate courses (for Signals and Engineering services)
run one year. The integrated courses (for doctors and specialized Masters Degree holders) and the
Women’s courses run only six months. Officers commissioned from the integrated course are inducted
directly as captains in the Pakistan army.
16 Except serving soldiers of JCOs of the armed forces, who may be married.

7/12/2010 6
23 years of age. Serving armed or civilian armed forces personnel are also considered
provided that they are between 17 and 23 years of age.17 Recruits must obtain a score of
at least 50 per cent in their matriculation (10th grade) or FA (12th grade) exams. The
initial testing and screening occurs at eight regional selection and recruitment centers at:
Peshawar (NWFP); Rawalpindi, Lahore, Multan (Punjab); Hyderabad, Karachi (Sindh);
Quetta (Baluchistan), and Gilgit (Northern Areas). Eligible candidates next take an
“intelligence” exam. Candidates must score at least 50 percent in initial testing. (For
persons from Sindh and Baluchistan, candidates must achieve a score of 45 percent as
described below). Successful candidates are next subjected to physical and medical tests
and a preliminary interview at the recruitment centers.18
Candidates who are selected from this interview process proceed to the Inter-
Services/General Headquarters Selection and Review Board in Kohat or its ancillary
centers in Gujranwala, Malir, or Quetta. Here candidates undergo four days of
observation and testing to assess their intelligence, psychological profile, leadership
potential, and physical fitness. After the four-day testing program at the ISSB, successful
candidates are recommended for the PMA, according to requirements issued by the Army
General Headquarters and derived from regimental reports on shortfalls.
Officer candidates generally are selected for the PMA based upon on merit,
although the army occasionally has relaxed standards to increase recruitment among
under-represented ethnic groups, as described below. Those candidates who are selected
for continued processing following the written exam are called for a preliminary physical
examination, which is conducted at recruitment centers established throughout Pakistan.
Candidates are declared physically fit by a board of officers headed by a Brigadier
General.19

17 Pakistan Army, ‘Regular Commission in Pakistan Army,’ available at

www.joinpakarmy.gov.pk/intr-procedure-for-comission.php. There was also a (now defunct) Junior Cadet
Course (JCC) was co-located with the PMA and was designed to be a feeder school for the PMA, providing
military instruction in addition to other subjects. The JCC began in 1976 and was shut down in 1988. The
aim was to compensate for the lack of high quality candidates by preparing JCC entrants through intensive
coaching.
18 Official website of the Pakistan Army, http://www.joinpakarmy.gov.pk/intr-procedure-for-
comission.php.
19 Official website of the Pakistan Army, http://www.joinpakarmy.gov.pk/intr-procedure-for-
comission.php.

7/12/2010 7
 In effort to attract more officers from Sindh and Balochistan, the Pakistan Army
relaxed the academic and testing requirements for recruitment in the early 1990s. The
army leadership sought to increase the number of Baloch and Sindhi officers despite their
lesser qualifications rather than offer remediation to increase their qualifications. (This
may have the unfortunate affect of increasing such officers’ attrition if they are relatively
less qualified than others when they are up for promotion.) The army has episodically
decreased the testing requirements for applicants from Sindh and Baluchistan, who
needed a 33 per cent (that is just passing) in their exams. (These minimum levels were
raised later as discussed below.) During the 1980s, in effort to more quickly induct
Sindhi and Baloch recruits into the PMA and then the army, the army allowed the Sindhi
and Baloch recruits to study an abbreviated course at the now defunct Junior Cadet
Course after which they could go to the PMA. Sindhi and Baloch applicants to the
infantry enjoyed even further relaxed grading standards with respect to their FA/FSc
exams relative to other candidates.20
Moreover, Sindhi and Baloch candidates recommended by the Inter Services
Selections Board were selected into the PMA irrespective of their rank on the “merit list.”
The merit list ranks applicants on the basis of their cumulative test scores. If the PMA
required 300 persons for a given class, they would typically take the top 300 persons in
rank order. However, for Sindhis and Baloch this was relaxed. If a Baloch or Sindhi
recruit were to be ranked at 320, for example, they would be given preference to someone
else who is ranked higher on that list. Thus Baloch and Sindhi recruits were favored in
two ways: first their threshold scores for passing was lower than for others and second
they need not place as highly on the merit list due to a quasi-quota system that is in effect
for selection.21
Subsequently, the educational scoring standards were raised although the authors
could not obtain the exact year when this occurred. Now, candidates resident in
Balochistan, the Federally Administered Tribal Areas, the districts of Kohistan, Chitral,
and Dir, and Tehsil (sub-district) Balakot of Mansehra in NWFP, Tharparkar and Tehsil

20 The percentage scores of candidates in their Intermediate (12 year) examinations are placed into
grade levels A, B, C, etc. B is the standard minimum grade level, except for areas that are being favored by
an affirmative action policy. Candidates from the favored areas may be admitted with C grades.
21 Information provided to S. Nawaz from Army General Headquarters.

7/12/2010 8
Umerkot of Sindh, Rajanpur, Cholistan, Desert Derawer Fort (also in Cholistan),
Salamsar, Mojgarh, Dingarh of Punjab, and the Northern Areas can be admitted for
preliminary testing with a score of 45 per cent in their intermediate examinations in
contrast to 50 per cent required for all others.

4. Ethnicity and Pakistan Army: The Weight of History

As is well-known, the army sustains criticism about its over representation of Punjabis
and, to some extent, Pashtuns. While this ethnic distribution reflects in part the
population composition of Pakistan, it also is a legacy of colonial recruitment practices
inherited by Pakistan. These practices endure despite the army’s varied efforts since
1947 to move away from those traditions.
One of the important concepts that continue to influence army recruitment is the
problematic notion of “martial races.”22 The so-called martial races included Punjabis
(from contemporary East and West Punjab, comprised of Sikhs, Hindus and Muslims),
Pashtuns (from the Northwest Frontier Province and the tribal areas) as well as the
Gurkhas of Nepal. In contrast, the British considered South Indians and Bengalis to be
“non-martial” and sought to exclude them from military service. Because the British were
motivated by a desire to build an effective military to protect their interests throughout
the empire rather than to develop an ethnically representative institution, they encouraged
recruitment of those martial races and discouraged the induction of non-martial races.23
Apart from their beliefs about martial races, the British sought to increase
recruitment from the NWFP in particular for geo-strategic reasons. The NWFP was the
outer ring separating the British colonial sphere of influence from that of Russia.
Afghanistan, per agreement between the two imperial powers, was the formal buffer
between the two. They negotiated the northern and eastern boundaries of the Afghan state
to ensure that the empires of Russia and Britain did not share a border. For this reason,

22 See Douglas M. Peers, ‘The Martial Races and the Indian Army in the Victorian Era’ in D.
Marston and C. Sundaram Eds. A Military History of India and South Asia (Bloomington: Indiana
University Press 2008) 34-52.
23 The British remained convinced of the notion of martial races for most of its tenure in South
Asia. See Hassan Askari Rizvi, The Military and Politics in Pakistan: 1947-1997 (Lahore: Sang-e-Meel
2000); Hassan Askari Rizvi, Military, State and Society in Pakistan (London: Palgrave 2000); Peers, ‘The
Martial Races and the Indian Army in the Victorian Era’.

7/12/2010 9
Afghanistan was forced to accept the oddly shaped Wakhan corridor which provided a
narrow band of territory between the two empires. Both reasoned that the likelihood of
dispute would be minimized if they shared no border.24
As a result of these demand-side considerations, the British Indian army
predominantly drew personnel from the northern part of the empire and Nepal.25
However, the composition of the army was also shaped by several important supply-side
considerations stemming from the socio-economic standing and preferences of potential
recruits. The Punjab peasantry was facing difficult economic times when the British
entered the region. Army service afforded those peasants opportunities to augment
family income and insulate them from the vicissitudes of agricultural dependence. Many
of these recruits came from the Salt Range (Potwar, also spelled Potohar) regions of
northern Punjab (e.g. districts of Jhelum Rawalpindi, and Attock-formerly Campbellpur)
as well the areas adjoining the NWFP.26 The British encouraged peasant interest in the
armed forces by granting agricultural land as a reward for military service.27
The British resolved that by the time India and Pakistan became independent
states on 15 and 14 August 1947 respectively, they would have operational control of
their own independent armed forces and the colonial government formed a committee to
execute this complicated task.28 While British and Indian stakeholders hotly debated the
process of dividing the armed forces, in the end, religion and territory guided the division
of personnel. While most men in service could elect to join either the Indian or Pakistani
military, there were important exceptions. A Muslim from the area that went to Pakistan
could not opt for India and a non-Muslim in the territory that became India could not opt
for Pakistan. There was no restriction for Muslims in India to join Pakistan or for a non-

24 See Rubin 2002 for more on this competitive history and the role of Afghanistan and the
Northwest Frontier Province.
25 Rizvi, The Military and Politics in Pakistan.
26 Rizvi, Military, State and Society in Pakistan.
27 In addition, in 1885, the British Indian government began developing an important network of
canals to irrigate the Punjab, which made those lands even more productive and thus attractive as
allurements into army service. The British allotted these lands in reward for service for the Raj, especially
to ex-servicemen (both officers and ranks). The net impact of these policies made the army an extremely
lucrative profession for Punjabi peasants seeking to improve their socioeconomic standing. Rizvi, Military,
State and Society in Pakistan.
28 For more information, see Stephen P. Cohen, The Indian Army: Its Contribution to the
Development of a Nation (Oxford: OUP, 1990). In particular see 32-56.

7/12/2010 10
Muslim in Pakistan to opt for India. While a few non-Muslims opted to stay in Pakistan
and a comparatively more Muslims chose to stay in India, the communal violence that
marked partition compelled some to seek permission to reverse their decisions.29
By August 15, 1947 the future disposition of the various units had been assigned,
with the exception of those troops that were either abroad at the time of partition or those
that were engaged in the Punjab Boundary Force, both of which were divided later. The
ratio for division was 64:36 for India and Pakistan, excluding the Gurkha troops, which
were divided between India and Britain.30 This was roughly the communal balance of
Hindus and Muslims in undivided India.
As noted above, there were no purely Muslim regiments in India and many of the
units with large Muslim representation were in areas that were to become part of India.
While some of these Muslims could have chosen to move to Pakistan, many chose to
remain in India. In contrast, virtually no Sikh or Hindu officers within Pakistan’s territory
elected to stay in the Pakistan army. Consequently, Pakistan did not receive any
regiments in full strength. Generally, as the smaller successor state, Pakistan was slated to
receive substantially fewer personnel, stores, supplies and facilities.31 Training
institutions remained with the state in which they were situated. India retained for
example, the prized Indian Military Academy at Dehra Dun, among other important
schools. Pakistan retained seven army training centers including the Staff College
(Quetta) and the Royal Indian Service Corps (Kakul).
In short, Pakistan’s army was “weak, less organized and ill equipped. Despite
these institutional problems, the new army soon had to perform internal security duties
from the first day of Pakistan’s independence. Moreover, it soon found itself embroiled in
armed confrontation with India in Kashmir.”32
British recruitment policies bequeathed several problems to independent Pakistan.
Pakistan emerged with two wings: West Pakistan (which is today Pakistan) and

29 Rizvi, Military, State and Society in Pakistan; Stephen P. Cohen, The Pakistan Army (Berkeley:
University of California Press, 1984); Pervaiz Iqbal Cheema, The Armed Forces of Pakistan (Karachi:
Oxford University Press, 2002).
30 Rizvi, Military, State and Society in Pakistan; Cohen, The Pakistan Army; Cheema, The Armed
Forces of Pakistan.
31 Cohen, The Pakistan Army, p. 7.
32 Rizvi, Military, State and Society in Pakistan, p.56.

7/12/2010 11
predominantly Bengali East Pakistan (which is today Bangladesh). In 1947, most non-
Muslim Bengali officers and jawans (enlisted) opted for India. This meant that the east
Pakistanis formed less than 1 percent of the total strength of Pakistan’s armed forces,
which exacerbated mounting ethnic and political tensions that were emerging between
Bengali-dominant East Pakistan and West Pakistan, with its Punjab and Pashtun ethnic
domination.
While the bulk of Pakistan’s military personnel were from West Pakistan, the
west did not produce recruits equally. The Punjab and NWFP continued to produce the
majority of officers and cadres while Sindh and Baluchistan remained massively under-
represented. The districts of Kohat, Peshawar, Campbellpur (now Attock), Rawalpindi,
Jhelum and Gujrat were the main recruiting areas for the Pakistan army. In those areas,
nearly every second family had some kind of ties to the army.33
Liaquat Ali Khan was troubled by the extreme imbalance in army recruitment
across the two wings of West and East Pakistan. He appointed a committee to investigate
why East Pakistanis (Bengalis) were inordinately underrepresented in the army. The
committee was also charged with finding ways of increasing the representativeness of the
army. While the report was not made public, the army did take a number of steps.
Notably, it raised two battalions of the new East Bengal Regiment. While some of these
recruits came from the pioneer (construction) units or were Muslims who had served in
the Bihar Regiment of the united Indian army, some of the junior commissioned officers
came from the Punjab Regiment. As Bengalis became available the Punjabis were
replaced. These regiments were distinct in that they were exclusively Bengali. (Other
regiments could not claim exclusive ethnic representation.) In 1968-1969, ten more
exclusively Bengali battalions were raised and recruitment to all branches of the Pakistan
military was opened to East Pakistanis.34
In 1959, Ayub Khan reduced the physical standards for recruitment into the army
in hopes of luring East Pakistanis to join the armed forces. (The standards were not
relaxed for those in West Pakistan.) While these efforts did increase the numbers of
Bengalis across the army (as well as the air force and navy), Bengalis never achieved a

33 Rizvi, The Military and Politics in Pakistan, Rizvi, Military, State and Society in Pakistan;
Cohen, The Pakistan Army.
34 Rizvi, The Military and Politics in Pakistan; Cohen, The Pakistan Army.

7/12/2010 12
level of representation consistent with their population distribution which was nearly 50
percent of Pakistan’s population.35
While Bengali numbers slowly increased, the army resisted further expansion of
Bengali representation at least in part because many within the army leadership cadre
harbored considerable “distaste for the quality of Bengali officers and other ranks.”36
The maltreatment of Bengalis and their lack of representation within the military was a
festering problem that undermined the ultimate coherence of a united Pakistan. Those
Bengali officers and other ranks that were in service comprised the backbone of the
Bengali resistance during the civil war of 1971. Despite clear warnings that their policies
of discriminating against Bengalis, the Pakistan army remained ambivalent whether
Bengalis should be “taken into full partnership or completely eliminated.”37
While the army’s anti-Bengali preferences are well-known, the reasons for the
lack of Bengalis in the army was not entirely demand driven (e.g. imposed by the
recruitment policies of the armed forces). Pasthun tribesmen enthusiastically joined the
army as did others from West Pakistan. As the supply of willing recruits swamped the
actual demand for recruits (officers and jawans), recruiters could choose from among the
best of them. Bengalis, who were often (or least assumed to be by those in West Pakistan)
physically smaller, could not compete given the prevalent and essentialized notions of
martial prowess. Perhaps if Pakistan faced a shortage of recruits, the army and other
services would have been more interested in rethinking their facile assumptions about the
connections between ethnicity and military competence. However, it is not clear in the
minds of historians of the Pakistan army whether Bengalis ever had the same level of
interest in the military. Certainly the number of applications received from East
Pakistanis was a full order of magnitude fewer than those from West Pakistan.38
Because the army ran the country for much of Pakistan’s existence before the
1971 civil war, the exclusion of Bengalis was particularly problematic as they comprised
a majority of pre-1971 Pakistan’s population. Arguably the compulsion to have a
nationally representative army is direr when the army directly or indirectly governs the

35 Rizvi, The Military and Politics in Pakistan; Cohen, The Pakistan Army.
36 Cohen, The Pakistan Army, p. 43.
37 Cohen, The Pakistan Army, p. 43.
38 Rizvi, The Military and Politics in Pakistan.

7/12/2010 13
state. Until 1971, the lack of appropriate representation of the Bengalis was a pre-eminent
concern. However, from 1947 onwards there have been similar questions raised about the
“martial qualities” of the Baloch and Sindhis. Both Baluchistan and Sindh have at time
hosted ethno-nationalist insurgencies with varying degrees of severity. In both cases,
Pakistan (under civilian and military leadership) pursued military responses often with
excessive force.
The loss of Bangladesh, the persistent complaints about Punjabi domination, and
the under-representation of Sindhis and Baloch have occasioned concerns about further
disintegration of Pakistan.39 Motivated by such concerns as well as a need for an army
that is truly representative, the army has tried to expand the numbers of Baluch and
Sindhis with little apparent success.

4. Importance of Representation in an All-Volunteer Army?

The Pakistan army is an all volunteer force according to Article 39 (“Participation of
people in Armed Forces”) of the 1973 constitution, which says that the state “shall enable
people from all parts of Pakistan to participate in the armed forces of Pakistan.”40 Under
the Pakistan Army Act of 1952, the state can introduce compulsory recruitment in an
emergency but this has never been discussed or actively pursued.41 Like most national
armies, the Pakistan army generally aspires to be –and indeed considers itself to be—
representative of the country’s polity. Certainly, there is some debate and even evolution
as to what “representative” means. It is clear, as the below exposition makes clear, that
this did not always mean comparable ethnic distribution even if the army has in recent
decades become more sensitive to this issue for reasons described herein. As is true of
all-volunteer forces, its ability to be representative of its population is derived from
demand-side constraints (e.g. the preferences of the army) and supply-side constraints
(e.g. the willingness of people to join the army).

39 See Selig S. Harrison, 2009; 1980.

40See text of the 1973 constitution, with the various alterations, http://www.pakistanconstitution-
law.com/theconst_1973.asp (accessed December 2, 2008).
41 B. Horeman and M. Stolwijk, ‘Refusing to Bear Arms: A World Survey of Conscription and
Conscientious Objection to Military Service’, War Resisters International, London, 1998. http://www.wri-
irg.org/programmes/world_survey/country_report/en/Pakistan (accessed December 2, 2008).

7/12/2010 14
 For several reasons, the issue of representativeness of the Pakistan army is
important to the Pakistani state and institution of the army. First, from the optic of
citizenship and equal opportunity, it is important that all Pakistanis be allowed (or at least
believe that they are allowed) to participate equally in the obligations and rewards
associated with military service.42 The widespread belief that the army is an exclusive
club of Punjabis and Pashtuns underscores the extent to which the citizenry does not see
the army as a national institution even if the citizenry generally holds the army in high
esteem, especially compared to other national institutions.43
Second, and related to the first, ethnic representativeness within the armed forces
is important both in preventing ethno-national strife but also in contending with internal
conflicts once they emerge. Pakistan has experienced several ethno-nationalist
insurgencies, with varying sanguinity, in Baluchistan, Sindh and the Northwest Frontier
Province (NWFP). The Pakistan security forces require adequate personnel who are
familiar with the physical terrain and who understand the local languages, cultures, and
other features of the human terrain of the affected region. Without such assets, the
security forces cannot contend effectively with the rebels within its borders. As noted
below, it is far from clear to what degree the Pakistan army leadership has shared this
view and when. Certainly in recent years, the army has become more attuned to this
requirement.
There are important downsides to including within the army co-ethnics of rebels.
Veterans of the armed forces may return to the troubled areas and participate in the
insurrection. They could even provide training to the rebels. Units drawn from the local
area may also be unable or unwilling to act against combatant co-ethnics who may
include friends, relatives, or others affiliated through tribe or other forms of allegiance.
Former or retired service members may even sympathize with the cause espoused by the
militants and work to serve their cause clandestinely.
Third, since the army has ruled Pakistan directly for more than half of its
independent existence and indirectly for the remainder, the issue of representation for the

42 The Pakistan army is still largely a male institution. Women are mostly in the fields of health and
education. Pakistan is distinctive among countries in the Muslim world in that it produced the first Major
General. Dr. Shahida Malik became the first female Major General of the Pakistan Army on June 17, 2002.
She assumed the position of Inspector General of some 31 Pakistan Army hospitals.
43 International Republic Institute, “IRI Index: Pakistan Public Opinion Survey July 15- Aug 7.”

7/12/2010 15
Pakistan army is perhaps even more salient than it is for other all-volunteer forces subject
to robust civilian control and which are not involved in governing their states. This is
likely to be the case because the military has developed, over the last six decades, a
system of perquisites for its officers as well as enlisted personnel which affords special
access to economic opportunities, educational institutions for their children, special
facilities for government services catering to the armed forces’ personnel among other
benefits that are exclusively available to members of the armed forces. The belief that the
army is a Punjabi-Pashtun enterprise necessarily raises larger issues about social justice,
unequal access to services, and the general problem of rendering public goods as private
assets for military personnel.44
While the appropriate ethnic representation in a national army is important to the
Pakistan army, the army sustains criticism that it fails to meet this standard.45 Critics of
the army’s composition note that Punjabis and Pashtuns remain over-represented while
ethnic minorities (Baloch, Sindhis) and religious minorities remain under-represented.
(Note that since Punjabis comprise the largest group of Pakistanis, such prevalence would
persist even if the composition of the army perfectly mapped Pakistan’s population
distribution.) Demographic realities notwithstanding, the army’s composition has given
rise to a number of conspiracies about Punjab domination of the other provinces in
connivance with Pashtun co-conspirators.46 Critics welcomed the 2001 announcement by
the Pakistan Army’s Adjutant General to promulgate a “first ever national recruitment
policy” targeting interior Sindh and Balochistan.47
The army has been, at times, cognizant that its failure to develop a truly national
army undermines its domestic standing as a genuinely national institution and fosters
ethnic grievances. The army learned this bitter lesson from its failure to include Bengalis
(who were concentrated in East Pakistan) in the armed forces and other branches of
government in the early years following independence. This, along with numerous other
Bengali grievances, fostered resentment in East Pakistan and a systematic belief that

44 Ayesha Siddiqa, Military Inc.: Inside Pakistan’s Military Economy (London: Pluto Press, 2007).
45 Mushahid Hussain, ‘Reforming the Armed Forces.’
2001; Selig S. Harrison, ‘Pakistan’s Ethnic Fault Lines,’ The Washington Post, May 11, 2009.
http://www.washingtonpost.com/wp-dyn/content/article/2009/05/10/AR2009051001959.html.
46 Cohen, The Pakistan Army, 40-41.
47 Hussain, 2001.

7/12/2010 16
Bengalis would always be second class citizens in a united Pakistan. Ultimately,
Pakistan’s malfeasance towards its eastern wing precipitated a civil war, which
culminated in the emergence of an independent Bangladesh with Indian covert and later
overt intervention.
In the aftermath of the secession of Bangladesh, the Pakistan army has sought to
broaden its recruitment base from one centered at the time of independence in 1947 on
the Potohar Plateau of northern Punjab when three districts (Campbellpur (now Attock),
Rawalpindi, and Jhelum) dominated the recruitment in the British Indian Army and
independent Pakistan. Since the 1990s, the army has followed an explicit policy of
broadening its recruitment base to draw from provinces other than the Punjab.48
According to the Pakistan Army’s ten-year recruitment plan launched in 2001, by 2011
the army hopes to increased Pashtuns from 13.5 percent to 14.5 percent; Sindhis from 15
to 17 percent, and the Baloch from nearly nil to 4 percent by 2001. In addition, the army
hopes to increase the numbers of persons from Azad Kashmir and Northern Areas from
nearly zero to 9 percent of the force. Punjabis would make up the balance of 55.5
percent. Minority recruitment would be increased marginally.49
It is possible that the army may not be able to fully address these compositional
issues. Unlike neighboring India which faces various officer shortages, the Pakistan army
continues to attract a large number of candidates for officer and soldier training. This
means that in Pakistan, military recruitment is likely demand constrained for Punjabis
and Pashtuns even if it is supply constrained for other ethnic groups. If ethnicity were not
an issue for consideration, the Pakistan army receives more applicants than it needs to fill
required billets. Thus, the army need not take lesser qualified or undesirable candidates
(howsoever defined) unless doing so advances other goals such as ethnic diversity or
geographical/ethnic needs to execute a mission (e.g. linguistic skills, personnel needed
for specialized geographical knowledge, etc.). Without increasing the willingness of
under-represented groups to join the armed forces, there may be little that the army can
do to make the army truly representative of the country’s ethnic and sectarian
composition.

48 Background note provided to author by General Headquarters, Pakistan Army, 1990.

49 Khan, Sher Baz, ‘Punjab’s dominance in army being reduced: ISPR,’ The Dawn, September 14,
2007. http://www.dawn.com/2007/09/14/top13.htm.

7/12/2010 17
5. Expanding the Ethnic Base of the Officer Corps: What the
Data Say
In this section, we present results of our historical and geographical analysis of officer
recruitment patterns. In figure 1, we present total annual officer intake from 1971 to
2005. This figure depicts the explosion in officer recruitment that occurred during the
Soviet occupation of Afghanistan. This expansion became most obvious from 1979
through the early 1980s. Whereas in 1979 there were about 100 officers inducted, by
1982, more than 1,000 officers were being accepted into the Pakistan army per year.
From 1984 onward, the numbers of officers inducted per year fluctuated considerably
with a high of 1,350 in 1995 and low of 590 in 1996 and an average of about 1,022 per
year between 1984 and 2005.
As is well-known, during this period, Pakistan took advantage of U.S. concerns
about the presence of the Soviet Union to advance its own strategic concerns about India.
The enormous infusion of funds from the United States and Saudia Arabia likely
subsidized this expansion of the Pakistan army. 50 Stephen Cohen has observed “this
massive expansion in officer intake coincided with a peak in Indian military rearmament
and after 1986, the adoption of an aggressive strategy by India, which culminated in the
Brasstacks fiasco of 1987…[D]uring this period, both India and Pakistan were racing
towards becoming nuclear weapons states.”51

Figure 1. Annual Intake of Officers for All Provinces (1970-2005)

50 Stephen Cohen in his review of this document.

51 Stephen Cohen in his review of this document.

7/12/2010 18
 1600

1400

1200

1000
Total Officers

800

600

400

200

0
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004
Year

Source: In-house manipulations of army officer recruitment data.

Geographic Distribution of New Officer Recruits Across Time

We next depict cartographically district market shares of officer recruitments for 1974,
1984, 1994 and 2005. We used 1974 as our first end point for the cartographic analysis
of recruitment shares due to the 1971 war, which saw the emergence of East Pakistan as
an independent Bangladesh. We wanted to permit adequate time to lapse to permit both
recruitment and records of recruitment to re-equilibrate after the conclusion of the 1971
war. These analyses are presented in Figures 2 through 5.
Figures 2 through 5 dramatically demonstrate that the Pakistan army has been
successful in making the army more geographically representative. As depicted in Figure
2, in 1974 the center of gravity for officer recruitment largely centered on key districts in
the Punjab with a few in the NWFP. By 1984 (see Figure 3), officer recruitment had
expanded throughout the rest of Punjab, large parts of Sindh, NWFP, the Northern Areas
and selected districts in Baluchistan. Changes between 1984 and 1994 were less dramatic
but illustrate continued expansion of officer recruitment. Changes between 1984 and
1994 were less dramatic and illustrate some continuity as evidenced by comparing
Figures 3 and 4. By 2005, as shown in Figure 5, officer recruitment had expanded into
many more districts in Baluchistan and indeed into most of the districts in the country.
Those districts without any recruits were generally in Baluchistan. It must be recalled

7/12/2010 19
that each of these maps capture one snapshot for each year’s batch of officers, which does
vary from year to year. However, comparing the distribution of recruits between 1974
and 2005 demonstrates vividly that officers are now coming from most districts in the
country.

Figure 2. Share 1974

Source: In house data manipulation.

Figure 3. Share 1984

Source: In house data manipulation.

Figure 4. Share 1994

7/12/2010 20
 Source: In house data manipulation.

Figure 5. Market Share 2005

Source: In house data manipulation.

One serious limitation of our data must be kept in mind: these are district level
data not officer data. One cannot assume that a recruit from Baluchistan is in fact
ethnically Baluc. Nonetheless, a Punjabi, for example, who has lived and/or has been
raised in Baluchistan is likely to have a very different worldview than one who has spent
much or most of his life in the Punjab. Even though we cannot identify the ethnicity of
the recruit, the expanded geographical reach of the Pakistan army suggests that its various
efforts have born fruit at least with geographical variation.

7/12/2010 21
 A less granular analysis of officer recruitment is depicted in Figure 6, which
shows the market share of officer recruits for the four provinces (Baluchistan, NWFP,
Punjab, Sindh), the Federal Capital (Islamabad), Azad Kashmir, Northern Areas and
FATA. As demonstrated in Figure 6, the market share for the Punjab had declined
dramatically from a high of over 80 percent in 1971 to a low of less than 40 percent in
2001.

Figure 6. Market Share by Province (1970-2005)

90.00%

80.00%

70.00%

60.00%
Percent Yearly Recruitment

Azad Kashmir
Balochistan
50.00% Fata
Federal Capital
Northern Areas
40.00% NWFP
Punjab
Sind
30.00%

20.00%

10.00%

0.00%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004
Year

Source: In house data manipulation.

7/12/2010 22
 Figure 7. Provincial Population
70%
N .W . F. P .
60%
FATA
50%

P UN J A B
40%

30% S IN DH

20% B AL O CH IS T AN

10% IS L A M A BA D

0%
1972 1981 1998

Source: Pakistan Population Census Organization, “Population By Province/Region Since 1951.”

In 2005, Punjab accounted for slightly less than 60 percent of all new officers. As
shown in Figure 6 (provincial market share over time) and Figure 7 (provincial
population totals for 1972, 1981 and 1998), Punjab’s recent contribution to officer intake
is similar to the Punjab’s 56 percent share of Pakistan’s population (per the 1998 census).
Dramatic change can also be observed for the market share for the NWFP which roughly
doubled from 10 percent in 1971 to over 22 percent in 2005. According to the 1998
Census, NWFP only accounted for 13 percent of Pakistan’s overall population,
suggesting considerable over-representation of new officers. Recruits from FATA
remained somewhat stable and were roughly proportional to their population distribution
in Pakistan. Baluchistan, which accounts for about five percent of Pakistan’s population,
by 2005 accounted for about 5 percent or more of new officer recruits.
In contrast, Sindh remains under-represented but exhibits important progress in
expanding the number of officers drawn from Sindh. In the early years of the recruitment
time-series, there were few officers from Sindh. By 2005, Sindh accounted for nearly 10
percent of annual officer intake. However Sindh is the second most populous province,
accounting for some 22 percent of Pakistan’s population (per the most recent census in
1998). There are no census data for the Northern Areas or Azad Kashmir. However,
market share from these areas have been turbulent across the time series.

7/12/2010 23
 The data in Figure 6 demonstrate a peculiar anomaly from 1999 to 2003 in the
market share for both the Punjab and Azad Kashmir. One notices a consistent decline in
the Punjab’s share beginning in 1999 commensurate with increases in the market share
by Azad Kashmir. In 2001, the Punjab’s market share reaches a nadir (37%) while that of
Azad Kashmir spikes at the same level. From 2001 onward, Punjab’s market share
increases and re-equilibrates at nearly 60 percent while that of Azad Kashmir declines to
its previous low levels. Because these provincial figures are aggregation of district-level
data which were multiple and independently entered, it is not likely to be a data entry
error. Moreover, the trend is observable over several years. Given the timing of this
transition, it is likely that it is related to the conversion of the Northern Light Infantry
(NLI), which was until 1999 a paramilitary unit, into a regular army formation. While
this conversion of the NLI into a regular unit may explain the shift in market share, this
event may have triggered a change in the way in the army accounted for recruits. In the
absence of a coherent explanation, we can only note this peculiar finding and draw
attention to it.
A final way of viewing the provincial distribution is to examine trends across-
time in per capita recruits (the number of recruits from a province divided by the
population of the province). Because Pakistan census data is only available for 1972,
1981 and 1998, we estimated population growth in Microsoft Excel’s TREND function.
To do so, we used 1972 and 1998 as the endpoints, assuming linear expansion in
population.52 Per capita recruits were calculated by dividing the number of recruits in a
given year by the estimated population (in millions). These data are presented in Figure 8.
As the data depicted in Figure 8 demonstrate, Islamabad (the federal capital) has
the highest per-capita recruitment since the early 1980s.53 At distant second places are
the Punjab and NWFP which have similar per capita recruits. FATA, Sindh and
Balochistan all have very low per capita recruits.

52 Provincial population data is available at the website of the Pakistan Population Census
Organization, “Population by Province/Region Since 1951.” We estimated annual provincial population
using the TREND command in Microsoft Excel.
53 At the time of independence, Karachi was the capital. Ayub Khan undertook the decision to
establish a new capital beneath the Margalla Hills. The first office building was occupied in 1966.

7/12/2010 24
 Figure 8. Recruits per Population (Millions) by Province (1972-2005)

70 Balochistan
60
FAT A
50
Federal
40
Capital
30 NWFP

20 Punjab
10
Sindh
0
72

75

78

81

84

87

90

93

96

99

02

05
19

19

19

19

19

19

19

20
19

19

19

20
Source: In-house calculations based upon estimated population and army recruitment data.

Urbanization
Pakistan has experienced steady urbanization. There are two paths to urbanization in
Pakistan. One is the growth of cities in Pakistan and the other is the build-up of
population density in the rural areas (e.g. the formation of “ruralopolises”). According to
the 1951 Census, 82 percent of the 34 million Pakistanis were living in rural areas
compared to only 18 percent who resided in urban areas.54 According to the most recent
census of 1998, nearly one in three persons is living in an urban area. If one includes the
urbanized rural areas (high population density without municipal services), some 56
percent of Pakistanis live in urbanized areas.55
Urbanization, whether by forming megapolises or ruralopolises, has similar
impact by introducing television, radio, modern communications, expanded access to and
reliance upon modern forms of medication, transportation, and migration. This has

54An urban area is defined by the Pakistani Federal Statistics Bureau as an area that has a large
number of residents as well as a municipal or local government and the provision of local services.
Population densities of 400 persons per square kilometer (1,000 persons per square mile) are typically
considered urbanized even if the provision of public services has been organized on a communal basis
(Mohammad A. Qadeer, Pakistan: Social and Cultural Transformations in a Muslim Nation (London:
Routledge, 2005) 14-15, 51).
55 Qadeer, Pakistan: Social and Cultural Transformations in a Muslim Nation, 49-51.

7/12/2010 25
opened up the rural areas and has brought about important changes in Pakistan’s social
structures and the opportunities available to Pakistan’s citizens.
In the context of this study, the question arises how “urbanized” the Pakistan
army is relative to the country writ large. The overall urbanization of the country is easily
obtained from the Pakistan national census, which has been conducted thrice since the
1971 war and the loss of Bangladesh (1972, 1981 and 1998). (It should be noted that the
Pakistan census excludes the Northern Areas and fields a shorter census in the FATA and
data are not disaggregated by agency.)
Determining how urbanized the Pakistan army officer corps is a more difficult
task with our data. Ideally we would like to calculate the percentage of officers living in a
rural area before they joined the army. Our data do not permit this for two reasons. First,
we do not have data on the complete officer corps, rather annual intakes of new officers.
Second, we do not have officer-level data. Since our data are at the district-level, we can
only approximate the urbanicity of new officers (rather than the entire officer corps)
using information about the urbanicity of districts and the market share of officers from
each district. To approximate the urbanicity of officer intake, we generated an
urbanization index for the Pakistan army officer intake for 1981 and 1998, corresponding
to the 1981 and 1998 censuses respectively. We did this by taking a weighted average
(using the market share of recruitment as the weight) of the percent urbanization of each
district.56
Figure 9 depicts the urbanization index of army officer intakes for 1981 and 1998
compared to the urbanization of Pakistan per the censuses in both years. The results are
striking: in both 1981 and 1998 the army cohorts are substantially more urban than the
country writ large. It should be noted that the composition of army cohorts varies
somewhat by year and thus the urbanization index likely would vary by year if we had
census data on urbanization for all years in our series. Nonetheless these two data points
suggest that army cohorts may be more urban than the country overall.

56 We would have liked to have done so for the 1972 census as well. However, the hard copy
census data for 1972 do not provide urban and rural breakdown for each district. Overall, in 1972, about 25
percent of the population resided in urban areas. See Pakistan Census Organization, “Population by
Province/Region Since 1951.” Available at
http://www.statpak.gov.pk/depts/pco/statistics/pop_by_province/pop_by_province.html. Moreover, the
early years of the army data are also unreliable perhaps owing to post-war turbulence in data collection.

7/12/2010 26
 There may be a number of reasons for this higher urbanicity. Literacy is much
higher in urbanized area of Pakistan than rural areas.57 Thus the pool of qualified recruits
is larger in urban areas than rural areas. Given that recruits must pay their own way to
army recruitment centers, persons in urban areas have greater and less expensive access
to transportation (train, bus, car, etc.) than those in rural areas and the journeys
undertaken by them are likely to be less onerous than those undertaken by persons who
live in remote areas, far from public transportation. Finally, the army has a greater
presence near large cities with Corps Headquarters and Cantonments in burgeoning cities
like Lahore, Multan, Peshawar, Karachi, Quetta, Bahawalpur, among others. These
locations make the army more accessible to city dwellers and, because the population in
those cities likely benefit from the military-related businesses, the residents in turn may
be more positively inclined towards the military.
Additionally, retired military personnel may also concentrate in key urban areas,
creating coherent influencer communities that are positively disposed towards the
military. Analyses of U.S. military recruitment find that the attitudes of influencers
(friends, family, neighbors, teachers, etc.) towards military service have important
impacts upon a person’s decision to join the armed forces. When key influencers are
positively inclined towards the military, a person’s propensity to join the armed forces is
greater. Conversely, when influencers are adversely disposed towards military service, a
person is less inclined to join the military. This is often referred to as an “influencer
effect.” 58

Figure 9. Percent of Pakistani Population in Urban Areas vs. Percent of Officer Cohorts from Urban
Areas (1981 and 1998)

57 There are wide gaps in literacy between adults, between males and female, among provinces, and
between those in rural and urban areas. For example, according to the most recent Pakistan Social And
Living Standards Measurement Survey (PSLM) from 2006-07, total literacy of persons 10 years or older
was 72 percent in the urban areas compared to 47 percent in rural areas. For more disparities, see Pakistan
Federal Bureau of Statistics, 2007.
58 See Bruce R. Orvis, Narayan Sastry, Laurie L. McDonald. Military Recruiting Outlook: Recent
Trends in Enlistment Propensity and Conversion of Potential Enlisted Supply (Santa Monica: RAND,
1996).

7/12/2010 27
 45 41.7
38.3
40
35 32.5
28.3
30
25
20
15
10
5
0
1981 1998
All Pakistan Army

Source: In house tabulation of urbanicity index and Pakistan Population Census Organization,
“Population By Province/Region Since 1951.”

Conclusions and Implications

In summary, analyses of these data show that new officers entering the Pakistan army to a
significant degree resemble the geographic (as opposed to ethnic) population distribution
of Pakistan writ large with a few notable exceptions. By 2005, Sindh remains under-
represented despite improvement while the Punjab and NWFP are over-represented to
varying degrees, depending upon the year in question. However, Punjab is increasingly
less dominant in annual officer intake. By 2005, most of the country’s districts have
produced recruits in a marked change since the first year in our data series. Curiously,
using officer intake for 1981 and 1998, the urbanization of these cohorts appears to be
higher than the urbanization of Pakistan overall using census estimates.
It must be kept in mind that our data do not and cannot provide insights into the
overall composition of the entire army because we only have data for yearly officer
intake. We have no cohort information (e.g. how many attended the PMA long course or
other courses together or even how the officers broke out into the five courses) and we do
not have officer-level data on their career paths. As such, we have no information about
the relative rates of retaining persons from these various and often relatively new districts
for officer production. While the army may be increasing recruits from these areas
annually, the army may also be losing them to attrition. Such differential attrition—if it

7/12/2010 28
exists—could be due to individuals’ perceived or real discrimination in the army;
institutional barriers (e.g. bias in promotion); individuals’ discomfort with being posted
far away from home; individuals’ poor preparedness intellectually or physically for the
army; or discomfort with any other aspect of the army life style. Finally, we have data
about geographical origins not ethnic identity. While in many cases these will be
isomorphous, one cannot make this assumption. Thus the above analyses pertain to
geographic –not ethnic—characteristics of the officer intake. Despite these empirical
caveats, these analyses do suggest that the army has made important strides in achieving
its objective of becoming at least a more geographically—if not ethnically—
representative force.
The implications of this changing recruitment base over time may be very
important, if uncertain, at this juncture. As the army—a key provider of social
advancement and other perquisites in Pakistan—becomes more accessible, the army
could be a foundation of greater national cohesion rather than a source of ethnic tension.
However, as the army expands its recruitment base, it must also expand military
infrastructure into new areas. This is particularly true for provinces that are not well
served by roads and other transport infrastructure, such as Balochistan. (Note that the vast
majority of military assets are concentrated in the Punjab.)
Such expansion of army infrastructure is likely necessary for several reasons. In
poorly connected areas, potential recruits may be unwilling to undertake potentially
costly long trips to be considered for recruitment. Individuals in such areas may also have
limited access to education and may be unprepared for military service. This may make
them even less willing to undertake arduous journeys if they expect that they will not be
qualified. This concern remains even for those from Balochistan and Sindh if they are
unaware of military policies to relax standards for them. In areas with poor military
presence, individuals may have less interest in military service compared to those areas
with significant military presence due the above-noted “influencer effect.”59)
Pakistanis outside of the Punjab and NWFP have not always welcomed the
introduction of cantonments and other military infrastructure. Baloch ethno-nationalist
insurgents cite the expansion of military presence in Balochistan as concrete evidence of

59 Orvis et al., Military Recruiting Outlook.

7/12/2010 29
“Punjabi colonization.” This resentment persists even while some welcome the
introduction of cantonments because they are accompanied by expanded access to
utilities (e.g. gas connections, electrification, potable water, better roads, etc.); access to
better schools; employment opportunities due to base activities and activities supporting
such bases among other facilities and services. Conversely, some residents resent the fact
that it takes the introduction of a military facility to bring such amenities to a community.
The expansion of the military into Balochistan exhibits the entire of range of such
sentiments.60
The impact of expanded geographical recruitment base upon the army’s ability to
conduct internal security operations is less clear. On the one hand, training persons from
ethnic groups with longstanding grievances with the center may provide important
military training to co-ethnics who rise against the state. On the other hand, their
successful integration into the army creates opportunities for their near and extended
families which in turn may diminish ethnic grievances and further integrate these ethnic
groups into the Pakistani national project.
Expanded recruitment also distributes the cost of military action to new segments
of Pakistani society. As these family members become more familiar with military
operations, they may increasingly support the operation if they believe the operation to be
just or addressing Pakistani national interests. However, as the costs of military action
increases and as civilian and military casualties mount, support for such operations may
diminish putting increased pressure both on the army and the civilian administration
(assuming a civilian dispensation).61 This cluster of concerns is particularly timely given
that the percentage of recruits from the NWFP has increased in recent years and it is in
this area where the Pakistan army has been most engaged against Pakistan’s domestic

60See Ziad Haider, ‘Baluchis, Beijing, and Pakistan’s Gwadar Port,’ Georgetown Journal of
International Affairs (Winter/Spring 2005) 95–103; Frederic Grare, ‘Pakistan: The Resurgence of Baluch
Nationalism’ Carnegie Paper No. 65 (Washington DC: Carnegie Endowment for International Peace Jan.
2006); International Crisis Group, ‘Pakistan: The Forgotten Conflict in Balochistan,’ Asia Briefing No. 69
(Islamabad/Brussels: International Crisis Group, October 22, 2007).
61 C. Christine Fair, “Islamist Militancy in Pakistan: A View from the Provinces,”
WorldPublicOpinion.org, July 10, 2009.
http://www.worldpublicopinion.org/pipa/pdf/jul09/PakProvinces_Jul09_rpt.pdf. See also International
Republican Institute, ‘IRI Index: Pakistan Public Opinion Survey July 15-August 7,’ October 1, 2009;
http://www.iri.org/mena/pakistan/pdfs/2009_October_1_Survey_of_Pakistan_Public_Opinion_July_15-
August_7_2009.pdf.

7/12/2010 30
militants. Despite expanding public opinion polls, extant polling does not employ sample
sizes that permit scientifically valid comparisons across provinces much less within
provinces. At this juncture, it is impossible to tell how expanded recruitment from the
NWFP, increased operational tempo in the same area and public support or lack thereof
for these operations interplay.62
Perhaps the most important questions about the Pakistan army surround the
posited Islamization of the Pakistan army, anti-Americanism, and politicization. As is
well-known, Zia ul Haq sought to Islamize the Pakistan army during his tenure of army
chief in the 1980s. However, scholars of the Pakistan army have remained dubious that
there were deep impacts upon the institution.63 Similarly, the 1989 U.S. cutoff of military
education to the Pakistan military under the International Military Education and
Training (IMET) program has led to widespread fears that the Pakistan army harbors
strong anti-U.S. sentiment. (This is likely to be the case although the reasons may in fact
pre-date the 1990 cessation of military assistance and training.) And as is well known, the
army has long been subject to politicization. This politicization has been manifested in
the politicized selection of the army chief and other army leadership; the army’s role in
manipulating and undermining political institutions and the judiciary, often using the
state’s intelligences agencies; the army’s historical and pervasive involvement in
governing the country including introducing massive changes to the country’s
constitution. These questions about the social, political and indeed religious
commitments of the Pakistan army officer corps are no doubt at the core of international
concerns about this institution.
Unfortunately, this current effort cannot address these questions squarely with
these data alone for reasons noted throughout this essay. However, ongoing and future
research may provide modest—but important—insights into the army. Working with a
team of quantitative analysts, we are assembling a panel of district-level household
economic survey data from 1991 to 2005. (1991 is the earliest year for such data). These
additional data will allow us to further characterize the districts from which officers are
recruited in terms of social liberalism, educational attainment, urbanicity, among other

62 One illustrative exception is Fair, ‘Islamist Militancy in Pakistan: A View from the Provinces.
63 Cohen, The Pakistan Army and Rizvi Rizvi, The Military and Politics in Pakistan, and Rizvi,
Military, State and Society in Pakistan

7/12/2010 31
house-hold and district-level socioeconomic and demographic variables that may explain
variation in district-wise recruitment outcome.
However, even the results of this ongoing research are circumscribed by an
ecological fallacy problem. That is, the best we can do is characterize more fully the
districts from which officers come. We cannot assume that that any particular officer has
a similar set of characteristics as the district in part because officer recruitment is not
random; rather it is the product of both supply-side and demand-side considerations.
While our results narrowly advance our understanding of the Pakistan army, the
changes that we have documented in officer recruitment are very important in terms of
the social impacts of the army’s role in the nation; the operational and political
constraints upon the army; and even the army’s capabilities to deal with critical internal
security challenges that are likely to deepen rather than recede in the policy-relevant
future. As such these highly-caveated findings should be of interest to policy makers and
analysts alike and may help them refocus inquiries about this important institution, even
if this present and future analyses cannot answer these queries comprehensively.

References
Abbas, Hassan, ‘A Profile of Tehrik-i-Taliban Pakistan,’ CTC Sentinel, 1/2 (January 2008).

‘Combating Terrorism: U.S. Oversight of Pakistan Reimbursement Claims for Coalition Support Funds,’
(Washington D.C.: GAO, 2008)

Beaumont, Peter and Mark Townsend, ‘Pakistan troops ‘aid Taliban’: New classified US documents reveal
that mass infiltration of Frontier Corps by Afghan insurgents is helping latest offensive,’ The Observer,
June 22, 2008, http://www.guardian.co.uk/world/2008/jun/22/pakistan.afghanistan.

Cheema, Pervaiz Iqbal, The Armed Forces of Pakistan (Karachi: Oxford University Press, 2002).

Cohen, Stephen P. The Indian Army: Its Contribution to the Development of a Nation (Oxford: OUP, 1990)
32-56.

Cohen, Stephen P. The Pakistan Army (Berkeley: University of California Press, 1984).

Fair, C. Christine and Seth G. Jones, ‘Pakistan’s War Within’, Survival, 51/6 (December 2009-January
2010) 161-188.

Gregory, Shaun, ‘The Terrorist Threat to Pakistan’s Nuclear Weapons’, CTC Sentinel, 2/7 (July 2009).

Harrison, Selig S. ‘Pakistan’s Ethnic Fault Lines,’ The Washington Post, May 11, 2009.
http://www.washingtonpost.com/wp-dyn/content/article/2009/05/10/AR2009051001959.html.

7/12/2010 32
Harrison, Selig S. ‘Drawn and Quartered?’ New York Times, February 1, 2008.
http://www.nytimes.com/2008/02/01/opinion/01harrison.html?_r=1&scp=1&sq=Selig+S.+Harrison&st=ny
t.

Horeman, B. and M. Stolwijk, ‘Refusing to Bear Arms: A World Survey of Conscription and
Conscientious Objection to Military Service’, War Resisters International, London, 1998. http://www.wri-
irg.org/programmes/world_survey/country_report/en/Pakistan.

Hussain, Mushahid, ‘Reforming the Armed Forces,’ June 10, 2001. Originally published in The Nation.
http://chowk.com/articles/5060.

Khan, Sher Baz, “Punjab’s dominance in army being reduced: ISPR,” The Dawn, September 14, 2007.
Available at http://www.dawn.com/2007/09/14/top13.htm.

Nawaz, Shuja, Crossed Swords (Oxford: OUP, 2008).

Negroponte, John D. ‘Pakistan’s Fata Challenge: Securing One of the World’s Most Dangerous Areas,’
Testimony before the Senate Foreign Relations Committee, May 20, 2008.

Orvis, Bruce R., Narayan Sastry, Laurie L. McDonald. Military Recruiting Outlook: Recent Trends in
Enlistment Propensity and Conversion of Potential Enlisted Supply (Santa Monica: RAND, 1996).

Pakistan Population Census Organization, “Population by Province/Region Since 1951

http://www.statpak.gov.pk/depts/pco/statistics/pop_by_province/pop_by_province.html.

Pakistan Social and Living Standards Measurement Survey (PSLM) from 2006-07, Pakistan Federal
Bureau of Statistics, 2007.

Peers, Douglas M. ‘The Martial Races and the Indian Army in the Victorian Era’ in D. Marston and C.
Sundaram Eds. A Military History of India and South Asia (Bloomington: Indiana University Press 2008)
34-52.

Qadeer, Mohammad A, Pakistan: Social and Cultural Transformations in a Muslim Nation (London:
Routledge, 2005) 14-15, 49-51

‘Regular Commission in Pakistan Army,’ www.joinpakarmy.gov.pk/intr-procedure-for-comission.php.

Rizvi, Hassan Askari, The Military and Politics in Pakistan: 1947-1997 (Lahore: Sang-e-Meel 2000).

Rizvi, Hassan Askari, Military, State and Society in Pakistan (London: Palgrave 2000).

Text of the 1973, http://www.pakistanconstitution-law.com/theconst_1973.asp (accessed December 2,

2008).

Tyson, Ann Scott, ‘A Sober Assessment of Afghanistan: Outgoing U.S. Commander Cites 50% Spike in
Attacks in East,’ Washington Post, June 15, 2008;
http://www.washingtonpost.com/wpyn/content/article/2008/06/14/AR2008061401639.html.

Wallen, Russ, ‘Keeping Pakistan’s Nukes Extremist-Free,’ Asia Times Online, June 16, 2009
http://www.atimes.com/atimes/South_Asia/KA16Df01.html

7/12/2010 33
7/12/2010 34