Вы находитесь на странице: 1из 44

Colonial Proprietary Elites and Institutions:

The Persistence of De Facto Political Dominance

Ali Cheema 1 Shandana Khan Mohmand 2


Associate Professor Doctoral Student
Lahore University of Institute of Development Studies
Management Sciences University of Sussex

Manasa Patnam3
Doctoral Student
St. Johns College,
University of Cambridge

July 31 20094
Abstract

One of the central questions in political economy is whether there is path dependence in the political
dominance of historic elite families in post-colonial societies. This paper uses a unique combination
of household surveys, archival data and family genealogies to empirically estimate whether the
political dominance of historic elite groups in the colonial period predicts local political dominance,
today. This question is analyzed in the context of rural Punjab, in Pakistan, where the colonial state
established institutions that granted unequal political and economic power to proprietary families and
excluded non-proprietary groups. The results suggest that political dominance established 150 years
ago persists in spite of the abolition of the de jure political institutions of the colonial state. We find
that path dependence is the result of continuity in local village institutions. Within the sample of elite
families we find that the magnitude of current land ownership impacts political dominance.

KEYWORDS: Institutions, Political Economy

1
Opposite Sector U, DHA- Lahore Cantt., Pakistan. Tel: +92-42-5722670 (ext. 2334). Fax: +92-
42-5722591. Email: cheema@lums.edu.pk.
2
14 Church Place, Brighton, BN2 5JN, UK. Tel: +44-750-22-464-22; Fax: +44 (0)1273 621202.
Email: S.Mohmand@ids.ac.uk
3
St. Johns College Cambridge, CB2 1TP, UK. Tel: +447796466475. Email: mp519@cam.ac.uk
4
The authors would like to acknowledge research and funding support from the Development
Research Centre for the Future State at the Institute of Development Studies. We are grateful to Asim
Khwaja, Jishnu Das, Mick Moore, John Harriss, Andres Mejia Acosta, Farooq Naseer, Adeel Malik,
Haris Gazdar, Alice Amsden, Judith Tendler, Adnan Qadir Khan, Asad Sayeed, Imran Ali, Syed
Turab Hussain, and Mushtaq Khan for helpful discussions and suggestions. Asadullah Tahir, Ali
Asjad Naqvi, Salma Khalid, Meekal Sequeira, Faraz Hameed, Hana Zahir and Hassan Javid provided
excellent research assistance. Bindu Badshah provided crucial editorial support. Seminar participants
at the Institute of Development Studies, University of Sussex, Lahore University of Management
Sciences and UNU-WIDER provided many useful suggestions. We are responsible for any errors
that may remain.
JEL Classification: O12, O17

1
1 Introduction

A growing literature suggests that historical institutions and the past dominance of elite groups

impact current economic and political development 5 (Acemoglu and Robinson 2006; Acemoglu and

Robinson 2007; Engermann and Sokoloff 1997; Lipset and Rokkan 1967). Elite dominance, and its

adverse impact, is attributed to the historical legacy of local economic and political institutions

(Acemoglu et. al. 2007; Engermann and Sokoloff 1997). Elite dominance may exhibit path

dependence and persist even after the formal political institutions that originally underpinned it have

been abolished (Acemoglu and Robinson 2006).

Acemoglu and Robinson (2006) provide a political economy explanation for elite persistence in the

face of change in formal institutions. Their model makes a distinction between de facto political

power, which is based on the ability to engage in collective action and to use force, lobbying and

bribery, and de jure power that is derived from formal political institutions. The model results in path

dependence in the face of shocks and formal institutional change, because of persistent elite control

over de facto political power. According to them, the persistence of de facto political

power...emphasizes how the same elites are able to shape politics (pg. 329). In a similar vein,

Pierson (2000) argues that path dependence in political control may result from the differential

ability of groups to resolve collective action problems and the allocation of political authority to

particular actors (pg. 259).

This paper provides further support for the hypothesis that there is persistence in elite dominance.

We use contemporary household survey data, historical archival data and family genealogical

surveys from Sargodha district in West Punjab, Pakistan to test empirically whether at the family

5
For an excellent survey on the impact of history on politics see Pierson 2000.

2
level, de jure political and economic dominance in the colonial period predicts de facto political

dominance, today. In colonial Western Punjab, de jure political and economic power was unequally

distributed. The colonial state granted both new land and settled property rights in existing

landholdings to what were termed village proprietary families. Landowning families were designated

the unofficial agency (Douie 1931), and mandated to act as intermediaries6 between the formal

state and local society, with sole authorization to collect land-revenue, resolve disputes, and maintain

law and order within their respective villages. Members of village proprietary families were the only

rural dwellers enfranchised to vote and to be elected to provincial and central legislatures and local

rural boards. The colonial state thus granted a de jure political monopoly to these village proprietary

families, and created a one-to-one correspondence between the possession of village agricultural,

grazing and homestead land; familial blood ties; the authority to govern the village; political and

electoral rights and de jure and de facto political power.

After Independence, the de jure institutions of the colonial state and the laws that underpinned them

were either abolished or overridden. The constitution and new laws formally established universal

citizenship with universal adult electoral franchise; and reduced de jure barriers to entry into many

economic and political activities. Colonial institutions that had underpinned local economic power

were placed in abeyance, abolished or eroded by market forces. These included: the Land Alienation

Act (1900), which placed restrictions on the transfer of agricultural land from hereditary landowners

to non-agriculturalists; agricultural production based on landless share tenancy (muzair system) and

the village-centred barter system of regulated artisan services (seip system). The Land Alienation Act

(1900) lost importance with the adoption of universal citizenship; tenancy was replaced by self-

cultivation; and the seip system was eroded by the integration of the village market into the wider

cash economy. Although the distribution of land remains very unequal, there is no longer the same

6
We use the terms intermediaries and agents interchangeably.

3
degree of correspondence between de jure and de facto political power and between economic power

and de jure political power.

This paper examines whether local rural elite political dominance persists despite these post-

independence changes in de jure political and economic institutions. We use a novel measure of

political dominance: the extent to which a political agent organizes village votes during electoral

contests, intermediates public service delivery between villagers and the state, and/or manages

disputes within the village. Our question is how far politically-dominant agents, today, come from

the same village proprietary families that enjoyed de jure political power during the colonial period.

Members of these families are identified using colonial village inspection reports, land revenue

records, a customized genealogical field survey and verification through interviews at the community

level and with key informants.

We conclude that the probability that an agent exercises political dominance in a particular village,

today, is 9% higher if the agent has descended from the proprietary family of that village. The result

holds even after controlling for village fixed effects and important household attributes such as caste,

education, occupation, current land ownership and current wealth. However, the result does not hold

in villages that were resettled as a result of the Partition of the sub-continent in 1947, for in these

cases, colonial village institutions were uprooted. It appears that in the absence of any external shock

to colonial village institutions, elite dominance is more persistent. Furthermore, our evidence

suggests that 91% of agents exercise political dominance in their informal and unofficial capacity and

not as representatives of political parties or of the state.7

7
The importance of non-state or semi-state actors exercising political dominance and political

influence in developing country politics has been emphasized in the literature on clientelism

(Kitschelt and Wilkinson 2007; Krishna 2007, Migdal 1988).

4
We also test whether historical differences in family landed wealth account for differences in the

degree of political dominance, today. These differences are measured by the size of land grant

allotted to/ or settled on each sample households ancestors by the state. This data was obtained by

matching current households with the colonial land records using: genealogical surveys; colonial

village inspection reports; and detailed interviews with village and family elders. Estimates show that

a hundred-acre difference in the size of the colonial land grant results in a 10% higher probability of

the household member being a politically-dominant agent, today.

The paper also examines the extent to which landed wealth drives the persistence of elite dominance.

It is not easy to estimate this relationship: there is simultaneity between current land ownership and

political dominance that biases the OLS estimates. We overcome this problem by using changes in

inheritance rules introduced in the colonial period as an instrumental variable (IV) for current land

ownership. The use of changes in inheritance rules as an IV for current landownership is an

important methodological contribution of this paper, which allows us to treat land ownership as

exogenous. The initial rule of inheritance, followed by certain tribes and some villages, was

primogeniture. This was replaced at the end of colonial rule by Islamic law, involving equal division

of property among all sons. We use the fact of an agent descending from a patriarch, who was a first

born son, and died prior to the abolition of the rule of primogeniture as an IV to predict the

magnitude of current land ownership within the sample of village proprietors. We find that while the

IV works well in explaining current land ownership in villages where primogeniture was the adopted

historic rule, it is not a statistically sound predictor in villages where the Islamic Law of Inheritance

was adopted, historically. Our IV results suggest that the magnitude of current land ownership is an

important causal factor influencing political dominance within the sample of village proprietors.

5
The results suggest that where village land markets are inert and there is substantial continuity in

land ownership initial differences among families in political and economic power have a long-term

effect on political dominance. This suggests that the political dominance of the elite is not being

eroded in spite of the proliferation of market exchange and the extension of universal adult franchise.

Interestingly, the political dominance of village proprietary families persists even after controlling for

differences in current land ownership and household wealth. This suggests that there are other factors

at work apart from wealth. One possible explanation is that historical investments in political

networks restrict the space for the emergence of new political associations. Within the sample of

village proprietors, however, we find that the magnitude of current land ownership matters for

political dominance.

The paper contributes to the literature on path dependence, elite persistence and institutional change

by providing rich empirical evidence that will inform the theoretical literature in this area (Pierson

2000, Acemoglu and Robinson 2006, Acemoglu and Robinson 2007). Most of the empirical

literature on elite persistence derives from comparisons among countries or among large units like

districts (in South Asia) (Acemoglu et. al. 2001, 2002, Banerjee and Iyer 2005, Dell 2008). 8 There is

little direct micro-level evidence.9 Furthermore, the existing literature says little about channels that

underlie persistence of political dominance. The advantage of examining this question in the context

of West Punjab is that the colonial state established local institutions with unequal de jure political

and economic power and these institutions lost importance after Independence. The results tell us that

de jure political and economic dominance established 150 years ago persists within families, today.

This suggests that the initial unequal distribution of political and economic power can have long-term

effects in the political arena (Pierson 2000, Acemoglu and Robinson 2007), and does not bode well

8
For an excellent review see Acemoglu (2009).
9
On the importance of generating evidence at the micro-level see Udry and Pande (2005)

6
for the creation of democracy. Whether this elite persistence also adversely impacts local level

development outcomes is an important question for future research (Cheema et al 2009a and b).

2 Colonial Local Administrative, Political and Economic Institutions: Creation and

Dominance of Village Proprietors

Historical background

Sargodha10 district, in Pakistan, lies between the Jhelum and Chenab rivers in the western plains of

the Punjab (Rasul 2006). When it was annexed by the British in 1849, no less than 89 per cent (of

the district was) in a state of nature, largely due to the lack of irrigation and the paucity of rainfall

(Ouseley and Davies 1866, pg. 5). The colonial authorities instituted several measures to make the

land cultivable and boost agricultural production. Between 1901 and 1906 an extensive network of

canals was constructed to provide perennial irrigation to the district from the river Jhelum (Ali 1988,

pg. 8, Wace 1933). This was part of the Lower Jhelum Canal Colony (LJC) scheme created and

controlled by the state. Because there were no settled cultivating populations the state could easily re-

assign land rights. This combination of circumstances spurred an ambitious, state-led experiment

which created a big push in settled agricultural production.

To enhance its economic and political control, the colonial state created unequal administrative,

economic and political rights in favour of elite landowners or village proprietors who were invested

with the authority to exercise control over non-proprietors. This ensured that the state never

permeated down to the village level. Instead, it interfaced with the citizens through the privileged

10
The areas constituting the current district of Sargodha were part of the larger district of Shahpur

during the colonial period

7
village proprietors who became pivotal intermediaries between the state and the rural citizenry in

matters of governance and politics. These intermediaries were not meant to represent the village

residents; on the contrary their role was to control the non-proprietary body of the village. They were

given state-like functions and designated the unofficial agency (Douie 1931)11 authorized to collect

land-revenue; resolve disputes; and maintain law and order in their respective villages.

This section briefly describes the hierarchical village social structure that evolved in Punjab as a

result of this strategy. It also documents the unequal nature of economic, political and village

administrative institutions that underpinned colonial governance. Finally, it highlights the nature of

change and continuity during the post-independence period.

Social stratification of villages

The social structure of a typical Sargodha village, summarized in Table (1), was and is stratified into

three quoms lineage groups and reflects the dominance of the village proprietary body (Alavi

2001, Ahmad 1977, Rouse 1988). These main quoms are further sub-divided into sub-quoms, which

reflect patrilineal kinship groups. The primary distinctions in the social structure are between

Zamindar and Kammi quoms and between them and the Muslim Sheikhs (Rouse 1988, pp. 73-77).

Zamindar quoms, standing at the top of the village hierarchy, represent landowners and tenant

cultivators, the only sub-quoms declared eligible to own land. Within Zamindar sub-quoms there is a

further distinction between members of the village proprietary body who are at the helm of the social

11
Douie describes in detail the role assigned to this unofficial agency by colonial rural administration

and argues that this represents continuity with the Sikh period. He explains how the official

administrative machinery was supplemented by representatives of landowners in the shape of

village headmen, inamdars, and zaildars (1931; 137).

8
order, and the non-proprietary cultivating tenants12. These three main quom categories were used by

the colonial state to determine criteria and eligibility of land ownership in the Punjab, and were

included in the Land Alienation Act (1900). The colonial district gazetteers used for this study,

provide detailed mappings of sub-quoms into the three main quom groups.

[Table 1 about here]

The Kammis or artisans are placed below the Zamindar quom while the Muslim Sheikhs or traditional

landless labourers lie right at the bottom of the social order. Typically, they are the most depressed

and exploited group in the village social structure. Table (1) shows the close correspondence between

the quom-based social structure and ownership and control of village land.

Local administrative institutions

The dominance of the proprietary body ensured that village governance was no longer to be based

upon the consensual will of the village community (Chakravarty-Kaul 1996, pg. 198). Since all

rights were decided with reference to the proprietary body whether it comprised small or large

landowners, the Wajib-ul-arz (Village Administration Paper) defined the relationship of the village

proprietary body members with: (a) the government; (b) each other; (c) non-proprietary village

residents13; (d) potential entrants and (e) other outsiders to the village (Chakravarty-Kaul 1996, pg.

198). The non-proprietary residents comprised tenants (Muzair), artisans (Kammi) and landless

12
It is important to understand that members of the same quom might be village proprietors in one

village and tenants in another.


13
Non-proprietary residents or sakin-deh were not considered members of the village community,

they were only residents of the village.

9
labourers (Muslim Sheikhs). With its roots in customary and colony law, the Wajib-ul-arz and the

revenue record represented the charters of village institutions during this time.

Colonial village institutions ensured that the: right to participate in the management of the village;

the right to alienate land; and rights in agricultural, grazing and village residential land were

restricted to members of the village proprietary body (Nelson 2002, pg. 63, Chakravarty-Kaul 1996,

Mohmand et. al. 2007). Since tenants (whether occupancy or At-Will14), artisans and landless

agricultural retainers were not considered members of the village proprietary body they were denied

these rights. The office of the lambardar (village headman) reinforced the interface between the

village and the state. Selected from within the proprietary body of the village, the lambardar,

regulated access to land and homestead; collected revenue; and most importantly, intermediated

between the state administration, the courts and the villagers in matters of public service delivery,

policing, criminal and civil justice. There is no doubt that the structure of colonial governance

assigned a powerful and central mediating role to the un-official agency of the village proprietary

body.

Local economic institutions

The edifice of colonial village institutions was based on an unequal distribution of land, which was

protected by colonial laws that both prohibited the transfer of agricultural land from hereditary

agricultural tribes to non-agricultural tribes, and prevented the fragmentation of landholdings

inherent in the Islamic law of inheritance. The former was achieved through the Punjab Alienation of

Land Act, 1900, which was reinforced through the enactment of pre-emption rules that banned the

14
Occupancy tenants were granted security of tenure and hereditary tenancy rights at the time of the

settlements, whereas tenants-at-will had neither security of tenure nor hereditary occupancy rights.

10
sale of land by a hereditary proprietor to a non-proprietor and the repossession of land by a creditor

(Ali 1988, Gilmartin 2004). These expedient laws resulted in ossifying the property rights structure

in land. The latter was achieved through the selective institutionalization of the law of primogeniture,

which replaced the practice of the Islamic law of inheritance for the communities that came to inhabit

canal colony villages, and recognized it for other villages where it was an established tribal practice.

Hierarchical village-level economic institutions further tilted the balance of power away from

landless residents. The muzair system (share-tenancy) ensured that agricultural production took place

through customary share-cropping arrangements between the proprietary body and landless tenants.

Since all land was owned exclusively by proprietary families, the landless tenants had to negotiate

with members of the proprietary body for access to both agricultural and homestead land.

The institutions of begaar (corve) and vangaar, whereby a landlord could ask his tenant to work on

a specific project in return for a meal, tightened economic control over non-proprietors. In the case of

village artisans, the proprietary body actively administered the seip arrangement, similar to the

jajmani15 system, through which the artisan provided specialized services to agricultural tenants and

landowners in return for in kind payments made at the time of harvest.

Political institutions

The colonial state also ensured the dominance of the village proprietary body in electoral politics by

15
The Jajmani system was an economic arrangement between families of different castes through

which one family exclusively performed services for the patron family, such as providing agricultural

labour in return for protection and employment. The payment was normally a fixed share in the

harvest rather than cash.

11
restricting the number of elected seats for urban politicians and allocating the bulk of seats to rural

areas (Yong 2005, pp. 240-81).

Their dominance in the political arena was further insured by the colonial Voter and Candidate

Qualification Rules (1919-35 and 1937-47 Yadav 1987). For rural constituencies, these rules

restricted enfranchisement to revenue paying members of the village proprietary body. Non-

proprietors, with the exception of revenue paying tenants who barely existed in these villages, were

disenfranchised. Furthermore, artisans and private landless tenants were only permitted to vote on the

condition that they had at least completed their primary education. Since, the Candidates Rules also

stipulated stringent property requirements, electoral politics came to be dominated by the village

proprietary families during the colonial period.

Post-independence institutional change and continuity

The political and economic institutions of the colonial state were impacted by a number of changes

after independence that significantly altered the nature and basis of the relationship between village

proprietors and the state on one hand, and between these families and village-level social groups on

the other.

West (Pakistan) and East (India) Punjab witnessed the greatest inflow and outflow of people during

the Partition of India and Pakistan. The emigration brought about several changes, of which the most

significant was the complete transformation in the religious composition of Punjabs villages. High

levels of communal violence drove almost all Hindu and Sikh families out of West Punjab into India

and brought thousands of Muslim families from East Punjab into Pakistan. The out-migrating

families left behind entire villages and over 4.3 million acres (Bharadwaj et al 2006; 23) of rural,

agricultural land in West Punjab. Since there was plenty of evacuee land that required resettling,

12
migrants from East Punjab flocked to the canal colony villages to claim a share. Consequently, entire

villages were resettled by migrants who had no historical links in this area.

This migration and resettlement process has important implications for our study. Since the process

of resettlement occurred after colonial political institutions had been reformed (see below) the one-

to-one correspondence between landownership and de jure political rights no longer existed.

Furthermore, the process of migration and resettlement would, by definition, have eroded historic

structures of authority and dominance that prevailed in both the resettled village and the migrants

village of origin.

Prior to independence, the landed classes of Punjab had enjoyed a direct and close patron-client

relationship with the colonial government, which had ensured them special rights and benefits. It

took the newly created state of Pakistan twenty-three years before it could alter the de jure political

relationship between the state and the landed rural elite. Although universal adult franchise was

adopted as part of the 1956 Constitution, it was not until 1970 that Pakistans first General Election

was held on this basis. Therefore, it was only in 1970 that the preferential de jure rights granted to

members of the village proprietary body were altered in practice and the electoral space opened up to

all village citizens.

In addition to the political changes, colonial institutions that underpinned the economic power of the

village proprietary body were placed in abeyance or abolished after Independence. The Land

Alienation Act (1900) lost its importance with the adoption of universal citizenship; landless share

tenancy (muzair) was replaced by self-cultivation as mechanisation changed the face of agriculture

(Figure 1); and the barter systems of regulated artisan services (seip) and corve (begaar) were

eroded by the integration of the village market into the wider cash economy and by the grant of

homestead land rights to previously marginalized groups. The fading out of these institutions was a

13
result of both conscious government policy and market-based changes. At the same time, the

tendency towards land fragmentation increased due to the replacement of the rule of primogeniture

by Islamic law, and two rounds of limited land reforms.

[Figure 1 about here]

In spite of these changes, the inability of the state to implement effective land reforms and the

inertness of land markets has ensured that continuity in land ownership still persists at the village

level. Figure 2 indicates that members of the village proprietary body continue to dominate current

land ownership in the sample villages.

[Figure 2 about here]

Figure 3 shows that there is a strong correlation between historical land grants and current land

ownership.

[Figure 3 about here]

Summing up

Under colonial rule, the authority over land, village governance and electoral politics were all vested

in the village proprietary body. Post-independence changes altered the de jure authority of these

landed families, which affected the basis of the relationship between them, the state and the village

citizenry. The important question is whether these changes in de jure political and economic

institutions of the colonial state have weakened the political dominance exercised by historic village

14
proprietary families; or whether the continuity in landownership and the structure of the local state

continue to vest political dominance in their hands. A critical sub-question is whether we would

expect the same degree of continuity in political dominance by village proprietors in migrant and

settled colonial villages, given that the former villages saw the erosion of colonial village institutions

and historic structures of authority, while settled colonial villages did not.

3 Context and Description of Data

Sampling villages and survey methodology

The paper uses household survey data from a random sample of 12 villages from Sargodha district.

Following this, a random sample of households, stratified by village proprietary families and quoms

(lineage groups), was drawn from within each village. The stratified random household sample was

generated through the following process:

1. An initial census was conducted in each village to collect information on the sub-quom

(patrilineal kinship group) of each household; its primary source of income; land ownership

and type of housing structure.

2. Household sub-quoms were counterchecked through a parallel exercise conducted with

selected key respondents in each village. Inconsistencies were resolved through interviews

with village elders.

3. The sub-quoms were mapped to the three main quom (lineage group) categories based on the

information contained in district gazetteers and the colonial revenue record. In addition, all

households that had descended from colonial village proprietary families were identified

through a genealogical survey (discussed in detail below), which was triangulated against the

colonial inspection reports and the revenue record.

4. After the preliminary census information had been collected and the sub-quoms mapped the

15
sample for each village was randomly drawn from the population stratified by the village

proprietary families and the three main quom groups: the Zamindars or cultivators and

proprietors; Kammis or artisans; and Muslim Sheikhs or landless labourers. Household

surveys were used to collect information on political dominance exercised by households as

well as on household demographics; wealth; land ownership; education; occupation etc. The

size of the household sample obtained through the survey is 880, with 865 complete surveys

and 15 partially completed surveys. The response rate16 of the survey was 87%.

There are two interesting details about our sample, which we exploit in the following analysis. First,

one of the 12 sample villages is a migrant village, ChakMigrant, which was originally inhabited by

Sikh proprietors who migrated to India at the time of Partition. Subsequently, the land was resettled

and granted as compensation to migrants from India who had lost their agricultural lands in East

Punjab (India).

Second, our sample villages reflect two distinct types of land inheritance patterns. Ten followed the

primogeniture rule of inheritance during the colonial period. Of these, eight were horse-breeding

colony villages and two were proprietary villages. The two remaining villages followed the Islamic

law of inheritance. However, in 1940, Islamic law replaced primogeniture in the horse breeding

colony villages and at the time of independence it was instituted in the two remaining villages as

well. This variation in the law of inheritance is important for our IV estimations.

Identifying agents: measuring political dominance

16
The response rate is calculated according to the standard formula given by the American

Association for Public Opinion Research, Standard Definitions: Final Dispositions of Case Codes

and Outcome Rates for Surveys (Lenexa, KS: AAPOR, 2006).

16
A new measure of political dominance used in the study, identified that an agent exercises political

dominance if he organizes village votes during electoral contests, intermediates public service

delivery demands between villagers and the state, and/or manages village-level disputes.

Specific questions included in the survey helped to identify these agents. Each sample household

named up to three individuals whom they would approach to mediate on their behalf in public service

delivery, dispute resolution and electoral politics. Relevant information was collected on the political

background of the household and any member who had contested and held political or administrative

office. Detailed ethnographic field work and interviews were used to triangulate the data. After

identification, the agents or intermediaries exercising political dominance were also exposed to the

same surveys.

The results reveal that 90.63% of them had no direct affiliation with political parties and/or the state.

This suggests that an extremely large proportion of agents exercise political dominance in a de facto

or personal capacity and not because of any authority vested in them as state officials or leaders of

political parties.

Identifying village proprietary families

Since the focus of the paper is to examine whether members of village proprietary families continue

to dominate political power as in the past, one of the main challenges was to identify all census

households belonging to these families. Our task was facilitated by two factors. A family was eligible

for membership of the village proprietary body only if it had been granted property rights to land; a

land grant; or land lease by the colonial state. In our sample villages all members were either grantees

or lessees of state land. Second, since land grants were only awarded to those belonging to the

17
Zamindar sub-quoms, it was clear that all members of the village proprietary body necessarily

belonged to this social category. Colonial Village Inspection Reports, which provide detailed

economic and social information about each village, listed all the sub-quoms of families that had

received land grants by the British.

In ChakMigrant, the sub-quoms of those who received compensation for land lost were confirmed

from the revenue records that existed from the time of resettlement. To reiterate, only members of the

colonial village proprietary families were entitled to land grants post-independence because these

grants were given as compensation for loss of land ownership that resulted from migration at the time

of partition.

The next step was to identify families from these sub-quoms who received land grants and became

members of the village proprietary body or had received land as compensation at the time of

partition. A separate survey was conducted to gather genealogical information of each member in the

patrilineal order of the household and ascertain the size of the land grants received by the family at

that time. Only those households that belonged to sub-quoms of land grantees identified by the

colonial Village Inspection Reports or by the Resettlement Records were included in this process. As

the sample villages were established at the end of the nineteenth century by the colonial state, they

are young villages so the original grantee of land was no more than four or five generations old.

Since we also had access to the Settlement Records of these villages all the information was

triangulated and cross-checked against this historical document. This information was used to draw a

random sample of village proprietary households for the household survey.

In summary, we undertook the following process:

1. Selection of sub-quoms: Used colonial village inspection reports and resettlement records to

make a comprehensive list of all tribes/ sub-quoms who were given land grants at the time of

18
settlement.

2. Conducting the survey: Identified households from the village census who belonged to these

sub-quoms, for the genealogical survey.

3. Selection of families: Used the genealogical survey to identify which member in the

patrilineage of the household had received a land grant; recorded the amount of grant

received and gave the household village proprietary body (henceforth VPB) affiliation

(applicable through patrilineage).

4. Verification: Used the settlement record of each village to verify the name of the original

grantee of land against the one identified by the genealogical survey. Verification was done

for the random sample of households belonging to village proprietary families to whom the

household survey was administered.

4 Empirical Strategy

The primary aim of the paper is to empirically test for persistence in the political dominance of elite

groups. An agent is defined as a person exercising political dominance if s/he is selected by our

respondents to intermediate in the domains of public service delivery, dispute resolution and electoral

politics. Consistent with this definition all individuals who get named as intermediaries or agents will

be denoted.

We use the following regression specification to test for elite persistence:

where if the individual i from village v gets named as an intermediary, otherwise

0. is a vector of Quom (lineage group) dummies (we use Zamindar and Muslim

Sheikh, keeping Kammis as the base category); is a vector of individual level attributes, viz.

19
CURRENT LAND, AGE, EDUCATION, WEALTH, and occupation dummies (we use agricultural

occupation (AG. OCC.) and non agricultural occupation (NON-AG. OCC.), keeping day and wage

labourers as the base category); and is a vector of village dummies. Our variable of interest to test

for elite persistence is . A positive and significant coefficient on the variable indicates

the presence of elite persistence.

We also use an alternative specification for elite persistence, which tests for the effect of historical

land grants on agent status:

Here is our variable of interest: the amount of historical land grant received by a family. The

advantage of this specification is that it uses a continuous variable as opposed to a binary variable

that captures the elite status of an individual at the time of colonial settlement.

Our hypothesis suggests that elite persistence will be strong in villages where there is continuity in

colonial village level institutions. Therefore, we would expect elite persistence to be weak in

ChakMigrant, where colonial village level institutions were eroded because of migration and

resettlement at the time of partition in 1947. If this hypothesis is correct then we would expect to find

a negative coefficient on the interaction term, , in our two specifications.

The next two sections present the results of our first two specifications and the sign and significance

of the interaction term.

Section 5 identifies the channels through which the VPB effect works. It identifies attributes that

explain the success of some VPBs in comparison with others. A priori, we have strong reason to

believe that land ownership is an important determinant of influence within the VPB sample. Figure

(3) plots a locally weighted scatter plot of current land ownership against historic land grants for the

20
entire sample. There seems to be a clear pattern that emerges, which associates larger current land

ownership with larger historical land grants. This suggests that land markets have been inert and

ownership has tended to remain with the same families. Unequal land ownership can sustain political

dominance in two ways: the landlord can use employment-related sanctions as a mechanism to

control votes, or he can use the income associated with landed wealth to buy political influence.

Therefore, in order to identify the importance of land as a channel we need to estimate whether

ownership of land is a positive and significant determinant of being an agent within the sample of

VPBs. The estimation of this relationship is, however, not straightforward because of simultaneity

between current land ownership and being an agent, which will bias the OLS coefficient. Given this

limitation we employ an IV strategy that uses the change in the inheritance rules made at the end of

colonial rule to exogenise land.

As mentioned earlier, 10 out of our 12 villages followed the rule of primogeniture till it was replaced

by the Islamic law of inheritance at the end of the colonial period17. We label these villages PRIMO-

villages (PVs). In these PVs, there were households that benefited from primogeniture if they were

the descendents of a patriarch, who was the first-born son, and died before the law was revoked. This

meant that the landed estate was inherited by the eldest son in its entirety. We label these households

Primo. On the other hand, NON-PRIMO households within PVs were those whose patriarch died

after the abolition of the law, leaving a fragmented estate equally divided among the sons 18.

Therefore, the likelihood is that primo households in our sample will hold larger parcels of land than

non-primo households on account of a random historical event. Based on this, we use the fact of an

17
In the case of horse-breeding colony villages this law was abolished in 1940 and in the case of the

proprietary villages it was abolished around independence.


18
Our surveys find that in some cases daughters were also given an unequal share. In other cases,

they were disenfranchised in violation of the Islamic law of Inheritance.

21
agent descending from a patriarch, who was a first-born son, and died prior to the abolition of the

rule of primogeniture as an IV to predict the magnitude of current land ownership within the sample

of VPBs.

The two remaining villages, labelled Non-Primo-Villages (NPVs), historically followed the Islamic

law. In these villages we would not expect to find an effect of being a primo household on the

magnitude of current landownership relative to non-primo households because all households were

subject to the Islamic law during both the colonial and the post-independence period. This provides a

falsification test for our IV.

It has already been established that under the colonial state only those individuals who were given

land grants and were part of the VPB were able to retain and perpetuate land holdings. Since all other

individuals were disenfranchised, we administered genealogical surveys to members of only those

sub-quoms eligible for land grants. This captured all the information we needed to identify

primogeniture affected' families (variable denoted as primo). Our IV specification is as follows (the

subscript 1 and 2 denote the first and second stage respectively):

where is the amount of current land owned by individual I, is a dummy variable


19
taking the value 1 if the individuals ancestral patriarch benefited from primogeniture and

19
We do not use the vector of quom controls, because for this sub-sample all the individuals belong

to the Zamindar quom In addition, we drop labourers as the reference group in the vector of

22
contains all variables used in our vector of controls as defined previously.

5 Results

Descriptive statistics

[Table 2 about here]

Table (2) presents the descriptive statistics for the individuals in our sample and compares the means

of different variables across agents and non-agents. It indicates that within the sample of VPBs many

more households are agents than non-agents, and the difference is statistically significant at the 1%

level. Furthermore, there is a considerable difference in the quantum of historical land ownership

between agents and non-agents. However, this evidence is not conclusive and does not enable us to

conclude that elite persistence exists because several other variables present a significant difference

across agents and non-agents. Table (2) shows that agents predominantly tend to belong to the

Zamindar quom; they are better educated; the majority are agriculturalists and have significantly

more land and wealth. The next section presents multivariate regression results that test for the effect

of belonging to the VPB on being an agent while controlling for these other factors.

[Table 3 about here]

A comparison between Tables (2) and (3) suggests that the VPB effect is not statistically significant

in villages where colonial institutions have been eroded. Table (3) shows that within the sample of

VPBs in ChakMigrant there is no statistical difference between households being agents and non-

occupational controls (as there are none in this sample) and distinguish only between agricultural and

non agricultural occupations.

23
agents. Though the agents are wealthier and more educated than non-agents, their historical

differences in elite status and land ownership do not appear to matter. This provides some prima facie

evidence of the relationship between the continuity of village-level institutions and elite persistence.

Regression results

[Table 4 about here]

Tables (4) and (5) report the regression results of our first two specifications. Column 1 shows results

from the first specification that only includes the VPB variable and the VPB*ChakMigrant

interactive term with village fixed effects. The coefficient on the VPB variable is positive and

significant at the 1% level. This suggests that those who are members of the VPB are more likely to

be agents. Furthermore, the VPB coefficient remains positive and significant even after controlling

for Quom (lineage group) fixed effects (Column 2) and additionally controlling for (Column 3) age,

education, wealth (proxied by asset ownership of the household), occupation and land ownership.

Table (4) (Column 3) suggests that, when controlling for other variables, the probability of being an

agent is 9.1% higher if the individual belongs to the VPB family.

[Table 5 about here]

Table (5) reports similar results using the continuous historical land grant variable. Again, this

variable is a positive and significant determinant of being an agent. The result holds even after we put

in quom (lineage group) fixed effects and our controls. Estimates show that a hundred acre difference

in the size of the colonial land grant results in a 10% higher probability of being a politically-

dominant agent today.

It is important to emphasize that both our variables of interest, which capture elite status, remain

24
significant even when we control for current land ownership and wealth. The fact that the VPB

variable continues to hold importance, even in the presence of the land variable, indicates some form

of elite persistence that goes beyond current land ownership. Undoubtedly, landed individuals have

power, but within this category those who belong to elite families are even more powerful. One

possible explanation is that historical investments in political networks restrict the space for the

emergence of new political associations.

Tables (4) and (5) confirm that elite dominance does not persist in ChakMigrant, the migrant village

in our sample. The coefficient on the ChakMigrant interaction term is negative in both regressions

and statistically significant only in Table (4). This suggests that there is a strong association between

elite persistence and continuity in underlying village level institutions.

Channels

We now report instrumental variable (IV) regression results that help us determine whether land

ownership is an important channel through which VPBs sustain political dominance. Our main

interest is to understand whether ownership of land affects agent status, conditional on belonging to a

VPB household. The question is whether ownership of more land increases the probability of being

an agent within the sample of VPBs. We noted earlier that the primogeniture instrument was only

applicable to those villages that followed the rule of primogeniture and to only those households in

these villages, whose patriarch died before the law was revoked. Therefore, we restrict our sample

only to VPB households and Primo-Villages.

[Table 6 about here]

Table (6) shows the results. In the first stage, with current land as the dependent variable, the

25
coefficient on Primo is significant and positive; households whose patriarch was the first born before

the abolition of the primogeniture law are likely to have greater land ownership. In order to test for

falsification we replicated the same regression for villages where the primogeniture law was not

applicable (Column: Non-Primo Villages, Table (6A)). As expected, the coefficient on Primo is

insignificant for this sample; this is intuitive since the households in these villages did not practice

primogeniture and patriarchs of any birth order were eligible to receive equal inheritance. Similar to

the OLS estimate the IV estimate is positive and significant. The IV results in the second stage

indicate that an increase in land ownership of 100 acres increases the probability of being an agent by

6.9%; the effect being significant at 1%.

There is still a possibility that a selection bias exists in the estimates because the rule of

primogeniture could have resulted in the migration of the younger sons from the village, thereby

changing the distribution of attributes within the village population. This could potentially affect

agent selection, giving rise to endogeneity. To check for selection bias, we performed a simple test to

exclude the possibility that offspring born later in primogeniture practicing households migrated out

of the village. Our genealogical surveys captured information on the residency status of all members

of the household, across several generations. Using this we created a pseudo-panel for our primo

villages to test whether non-first born sons in primogeniture affected households are likely to have

left the village relative to first-born sons that benefited from primogeniture. Table (7) reports the

results. The variable of interest is the interaction term Primo*BO which measures the effect of being

an individual of birth order greater than one, resident in a primogeniture affected household. Its effect

is insignificant across both specifications (one with generation dummies and the other without),

which removes concerns about the effect of a selection bias on our estimates.

[Table 7 about here]

26
These results suggest that the magnitude of land owned is an important channel that gives VPBs

access to political dominance. It appears that the relationship between land and power is symbiotic

for our VPB sample.

6 Conclusion

The paper provides empirical support for the elite persistence hypothesis (Acemoglu and Robinson

2006, 2007, Pierson 2000) and suggests that initial differences in political and economic power at the

local level, have a long-term effect on an agents ability to sustain political dominance. We also find

that this dominance persists despite changes in the de jure political structure of the state and in

economic institutions that underpinned initial political and economic dominance.

It is also evident that village proprietary families who had received property rights in land or agrarian

land grants from the colonial state continue to exercise political dominance within our village

sample. Our results suggest: that the probability that a household exercises political dominance today

is 9% higher if it has descended from a historic village proprietary family; and a hundred acre

difference in the size of the colonial land grant results in a 10% higher probability of a descendant of

the family being politically-dominant, today. Interestingly, these historic village proprietary families

exercise political dominance in their informal or de facto capacity without de jure recognition.

90.63% of agents identified as intermediaries by the respondents have no direct affiliation with

political parties and/or the state.

It is important to emphasize again, that elite status is a significant predictor of political dominance

after controlling for land ownership and wealth. This suggests that the landed do have power, but

within this category, those that belonged to colonial elite families are even more powerful. We have

suggested that this could be the result of increasing returns to historical investments in political and

27
state networks, which in turn, could be based on the continuity of the administrative structure of

colonial local state institutions.

The paper also examines whether current landownership impacts the persistence of political

dominance within the sample of historic village proprietors. The methodological contribution of this

paper lies in estimating the relationship between current land ownership and political dominance

given the existence of simultaneity that biases OLS estimates. Our IV strategy uses changes in

colonial inheritance laws to exogenise current land ownership. The IV and OLS results suggest that

the magnitude of current land ownership is an important factor that contributes to political

dominance. Therefore, village proprietors with access to greater landed wealth continue to wield

more political dominance.

This study provides detailed and rich micro-evidence which suggests that the history of political and

economic power continues to affect political dominance even after the de jure institutions that once

supported it have been eroded. This could be attributed to the tenacity of the network effect, further

strengthened by the continuity of local state institutions and the wealth of the historically rich elite. It

is also evident that elite dominance only persists in contexts that experience continuity in colonial

village institutions, whereas the political dominance of colonial village proprietary families does not

persist in migrant villages that were resettled as a result of partition in 1947. It would be beneficial to

conduct research on the impact of elite persistence on village development outcomes and identify the

precise channels that allow the historic elite to sustain political dominance in the current situation.

28
Bibliography

Acemoglu, D, 2008. Persistence and Change in Institutions, Marshall Lecture delivered at the

University of Cambridge.

Acemoglu, D., S. Johnson and J. A. Robinson, 2001, The Colonial Origins of Comparative

Development: An Empirical Investigation in American Economic Review 91(5), pp. 1369-

1401.

Acemoglu, D., S. Johnson and J. A. Robinson, 2002, The Reversal of Fortune: Geography and

Institutions in the Making of the Modern World Income Distribution in Quarterly Journal of

Economics 117 (4), pp. 1231-1294.

Acemoglu, D. and J. A. Robinson, 2006, De Facto Political Power and Institutional Persistence.

Mimeo: MIT.

Acemoglu, D. and J. A. Robinson, 2007, Persistence of Power, Elites and Institutions in American

Economic Review 96(2), pp. 325-330.

Ahmad, S., 1977, Power and Influence, in Jamil Rashid and Hasan Gardezi (eds.) Class and Power

in a Punjabi Village, New York: Monthly Review Press, pp. 91-126.

Ali, I., 1988, The Punjab Under Imperialism, 1885-1947, Delhi: Oxford University Press.

Arthur, B. W., 1989, Competing Technologies, Increasing Returns, and Lock-In by Historical

Events, in Economic Journal 99(394), pp. 116-131.

29
Bharadwaj, P., A.I. Khwaja and A. Mian, 2006, The Big March: The Nature of Migratory Flows

during the 1947 Partition of British India. Mimeo: Kennedy School of Government.

Banerjee, A and L. Iyer, 2005, History, Institutions and Economic Performance: The Legacy of

Colonial Land Tenure Systems in India, in American Economic Review 95(4), pp. 1190-

1213.

Chakravarty-Kaul, M., 1996, Common Lands and Customary Law: Institutional Change in North

India over the Past Two Centuries, Delhi: Oxford University Press.

Cheema, A., S.K. Mohmand and S.A.A. Naqvi, 2009a, Bringing Electoral Politics to the Doorstep:

Who Gains Who Loses?, Mimeo Lahore University of Management Sciences

Cheema, A., M.F.Naseer, S.A.A.Naqvi, and B.M.Siddiqi, 2009b, Colonial Village Institutions, Path

Dependence and Public Good Provision: Do Peasant Republics Serve Better Than Feudal

Estates?, Mimeo Lahore University of Management Sciences

Dell, M, 2008, The Mining Mita: Explaining Institutional Persistence, Mimeo MIT.

Douie, J., 1931, The Settlement Manual with Notes, Sixth Edition, Lahore: Mansoor Book House.

Engerman, S. L. and K. L. Sokoloff, (1997), Factor Endowments, Institutions and Differential Paths

of Growth among New World Economies: A View from Economic Historians of the United

States, in Stephen Haber (eds.) How Latin America Fell Behind: Essays on the Economic

30
Histories of Brazil and Mexico, 1800-1914, Stanford: Stanford University Press, pp. 260-

304.

Gazetteer of the Shahpur District 1897 (Revised Edition), 1994, Lahore: Sang-e-Meel Publications.

Gilmartin, D., 2004, Migration and Modernity: the State, the Punjabi Village and the Settling of the

Canal Colonies, in Ian Talbot and Shinder Thandi (eds.), People on the Move: Punjabi

Colonial and Post Colonial Migration, Karachi: Oxford University Press.

Krishna, A., 2007, Politics in the Middle: Mediating Relationship between the Citizens and the State

in rural North India in Herbert Kitschelt and Steven I. Wilkinson (eds.) Patrons, Clients and

Policies: Patterns of Democratic Accountability and Political Competition, Cambridge

University Press: Cambridge.

Kitschelt, H. and S. I Wilkinson., 2007, Citizen-politician Linkages: An Introduction in H.

Kitschelt and S. Wilkinson (eds.), Patrons, Clients and Policies: Patterns of Democratic

Accountability and Political Competition, Cambridge: Cambridge University Press.

Lipset, S.M. and S. Rokkan, 1967, Cleavage Structures, Party Systems and Voter Alignments: An

Introduction, in Seymour Martin Lipset and Stein Rokkan (eds.) Party Systems and Voter

Alignments, New York: Free Press.

Migdal, J.S., 1988, Strong Societies and Weak States: State-Society Relations and State Capabilities

in the Third World, Princeton: Princeton University Press.

31
Nelson, M,. 2002, Land, Law and the Logic of Local Politics in the Punjab, 1849-1999,

Dissertation presented for the degree of Doctor of Philosophy, Columbia University.

Ousley, G. and W. G. Davies, 1866, Revised Settlement of the Shahpoor District. Lahore: Panjab

Printing Company.

Pande, R. and C. Udry, 2005, Institutions and Development: A View from Below, Economic

Growth Centre Discussion Paper No. 928, Yale University.

Pierson, P., 2000, Increasing Returns, Path Dependence, and the Study of Politics, in American

Political Science Review, 94(2), pp.251-67.

Rasul, M., A., 2006, The History of Sargodha. Lahore: Zephyr Publisher.

Rouse, S., J., 1983 Systemic Injustices and Inequalities: Maliki and Raiya in a Punjab Village,

in Hassan Gardezi and Jamil Rashid (eds.) Pakistan, the Roots of Dictatorship: The Political

Economy of a Praetorian State, London: Zed Books, pp. 311-325.

Rouse, S., J., 1988, Agrarian Transformation in a Punjabi Village: Structural Change and its

Consequences, Dissertation presented for the degree of Doctor of Philosophy, University of

Wisconsin-Madison.

Wace, F.B., 1933, The Punjab Colony Manual: Lower Jhelum Colony Conditions and Forms,

Lahore: Panjab Printing Company.

Wilson, J., 1896, Tribal Customs in the Shahpur District of the Punjab, Lahore: The Civil and

32
Military Gazette Press.

Yadav, K., C., 1987, Elections in Panjab 1920-1947. Delhi: Manohar Books.

Yong, T., T., 2005, The Garrison State: The Military, Government and Society in Central Punjab,

1849-1947. New Delhi: Sage Publications.

33
FiFe 1: FeFe FasiFicaFioF oF Fas: SaFodha FisFicF

90

80

70

60
% Farm Area

50
1960
1980
2000
40

30

20

10

0
Self-cultiavtion Sharecropping Leased

Souce: Auicuua ueus (uauious issues)

34
FiFe 2: FeF Fad FwFeFship AcFos QFoFs

35
FiFe 3: FisFoFicaF Fad Fa aFd FeF Fad FwFeFship

36
37
38
39
40
41
42
43

Вам также может понравиться