When Strength Becomes Weakness: Precolonial State Development, Monopoly on Violence, and Civil War

• Published date: 01 December 2023
• DOI: http://doi.org/10.1177/00104140231169017
• Author: Casper Sakstrup
• Date: 01 December 2023
• Subject Matter: Articles

Article

Comparative Political Studies

2023, Vol. 56(14) 2131–2164

© The Author(s) 2023

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/00104140231169017

journals.sagepub.com/home/cps

When Strength Becomes

Weakness: Precolonial

State Development,

Monopoly on Violence,

and Civil War

Casper Sakstrup

1



Abstract

Does the legacy of precolonial statehood affect contemporary levels of civil

conf‌lict outside Europe? I argue that places with higher levels of precolonial

state development were more likely to end up as weak modern-day states

because precolonial state structures and authority structures established by

European colonizers came to exist in parallel. This created opportunities and

motivation for civil conf‌lict still present in many countries today. I illustrate

the argument in the cases of India, Burma, and Ethiopia and test it statistically

in a global sample covering 109 countries outside Europe. The results strongly

support the theory. Countries with higher levels of state development 3500

BCE–1500 CE have weaker state monopolies on violence and markedly

higher levels of intrastate armed conf‌lict in modern times (1946–2018). The

f‌indings remain robust across numerous alternative specif‌ications, including

using the timing of the Neolithic Revolution as an instrument.

Keywords

civil war, conf‌lict processes, state building, colonial legacies, precolonial states

1

Aalborg University, Denmark

Corresponding Author:

Casper Sakstrup, Department of Politics and Society, Aalborg University, Fibigerstræde 3,

Aalborg Ø 9220, Denmark.

Email: casa@dps.aau.dk

Introduction

Strong state structures are widely believed to reduce the risk of civil war

1

(e.g.,

Fearon & Laitin, 2003;Fjelde & De Soysa, 2009;Müller-Crepon et al., 2021).

While this is true for contemporary levels of state capacity, this article

demonstrates that strong precolonial state structures had the opposite effect

outside Europe. I present a theoretical argument detailing how precolonial

statehood

2

adversely affected state building outside Europe, causing a higher

risk of intrastate armed conf‌lict that persists in many countries today. The

catalyst of this surprising transformation from strong precolonial states to

weak modern-day states was the arrival of European colonizers. In places with

high levels of precolonial state development, it was costly for European

colonizers to defeat, dissolve, and replace indigenous state structures (Ertan

et al., 2016;Hariri, 2012). As a consequence, precolonial authority structures

persisted with varying degrees of autonomy in parallel with authority

structures set up by European colonizers (see also Hariri, 2012;Müller-

Crepon, 2020). The existence of parallel authority structures during the co-

lonial period meant that a state-wide monopoly on violence was less likely to

develop, which increased opportunities and motivation for intrastate armed

conf‌lict in a vicious circle still visible today: Opportunities for violence in-

creased since it is more feasible to organize armed rebellion in areas outside of

central state control (e.g., Fearon & Laitin, 2003;Fjelde & De Soysa, 2009;

Thies, 2010), while motivation for violence increased when central gov-

ernments sought to dissolve parallel authority structures, in particular in

connection with decolonization.

The theoretical argument is illustrated in three cases with comparatively

high levels of precolonial state development: India, Burma, and Ethiopia. All

three cases illustrate the same general mechanism, but they also demonstrate

how this general mechanism manifested itself differently in three different

contexts. Across India, precolonial state structures survived with varying

degrees of autonomy under the indirect colonial rule of the British, most

notably in the Princely States. In neighboring Burma, the British eventually

defeated and dismantled the main precolonial state, the Burmese Kingdom,

but precolonial state structures persisted in several areas, for example, in the

Highlands. In Ethiopia, the precolonial state managed to remain largely

autonomous, while a smaller part of the territory, Eritrea, was colonized by the

Italians, thus also creating parallel authority structures. In all three cases, the

presence of parallel authority structures and consequent weak state monopoly

on violence increased opportunities and motivation for intrastate armed

conf‌lict still visible in recent times, for example, in the Eritrean War of In-

dependence, in armed conf‌licts in the highlands of Burma, and in armed

conf‌licts in several former Princely States in India.

2132 Comparative Political Studies 56(14)

I test the theory statistically

3

in a global sample of 109 countries outside

Europe using the State Antiquity Index (Bockstette et al., 2002;Borcan et al.,

2018) to measure precolonial state development (3500 BCE–1500 CE), and

the UCDP/PRIO Armed Conf‌lict Dataset (Gleditsch et al., 2002) to identify

intrastate armed conf‌licts in present times (1946–2018). In accordance with

the theoretical expectations, I f‌ind that higher levels of precolonial state

development are strongly associated with a higher probability of intrastate

armed conf‌lict and a greater share of conf‌lict years 1946–2018. Across all

models, the effect sizes are striking. For example, the predicted share of

conf‌lict years is four times greater in countries with high levels of precolonial

state development than in countries with no history of precolonial state in-

stitutions above the village or tribal level. Furthermore, the f‌indings remain

robust across numerous different model specif‌ications, including using the

timing of the Neolithic Revolution as an instrument for precolonial state

development, using alternative operationalizations of precolonial state de-

velopment, adding more geographic and climatic control variables, and ex-

cluding inf‌luential observations.

To probe the theoretical mechanism, I conduct mediation analyses with

seven potential mediators including the main mediator of interest, state

monopoly on violence, measured using data from the Varieties of Democracy

Project (V-Dem) (Coppedge et al., 2019a;Pemstein et al., 2018). Due to the

strong assumptions underlying mediation analysis in general, measurement

challenges pertaining to the main mediator of interest, and high levels of

statistical uncertainty, this analysis should only be considered as a cautious

probe of the mechanism. While keeping this in mind, the descriptive results

display a clear pattern in which state monopoly on violence and population

size are the strongest mediators, while wealth, political regime type, ethnic

fractionalization, positive horizontal inequality, and negative horizontal in-

equality all contribute smaller or even negative shares of the total mediation

effect.

Research on the historical roots of contemporary civil conf‌lict has mainly

focused on the consequences of ethnic partitioning through colonial border

drawings, precolonial and colonial conf‌lict persistence, and various forms of

colonial rule and administration (Besley & Reynal-Querol, 2014;Dincecco

et al., 2019;Fearon & Laitin., 2014;Ito, 2020;Lee & Paine, 2019;

Michalopoulos & Papaioannou, 2016;Mukherjee, 2018a,2018b). However,

three studies have investigated how ethnic groups with a history of precolonial

statehood in Africa and former British colonies affect present-day risks of

ethnic conf‌lict (Paine, 2019;Ray, 2019;Wig, 2016). Specif‌ically, Wig (2016)

f‌inds that ethnic groups in Africa with a history of centralized precolonial

political organization are less likely to be involved in postcolonial ethnic

armed conf‌licts, arguing that these groups have been better suited to bargain

with the central government. Partly to the contrary, Paine (2019) and

Sakstrup 2133