Martial races as clubs? The institutional logic of the martial race system of British India

Abstract

Military institutions can be seen as a solution to a type of principal-agent problem, in which a government principal contracts with a military agent to produce violence on their behalf. Absent extensive monitoring, low effort should be expected from the military agent. This should especially be the case for a mercenary force which cannot rely on patriotism or ideology. Yet the mercenary army of the British-ruled India consistently produced high and dedicated performance. In this paper, I argue that the otherwise curious institution of recruiting predominantly from the so-called ‘martial races’ helps to explain this puzzle. Drawing on Iannaccone’s club goods model, I argue that the martial race recruitment system represented an effective solution to the principal-agent problem by allowing the British to benefit from a repeat business arrangement with a small number of ethno-religious groups whose costly in-group prohibitions and poor outside options helped both to screen out potential shirkers and deter poor performance on the battlefield.

Keywords

Club goods institutions and rational choice

Introduction

A soldier is among other things an individual hired by a political ruler to produce violence on his or her behalf. As such, relations between rulers and their armies have been seen in terms of the principal-agent framework (Feaver, 2003; Miller et al., 2018). According to the most parsimonious version of this framework, the agent should, absent extensive monitoring on the part of the principal, exert minimal effort in their allotted tasks. In the military context, the principal can guard against this risk by relying on appeals to patriotism and ideology (Barber and Miller, 2019; Miller et al., 2018). If, however, the principal relies upon foreign soldiers, especially to maintain his rule over their country, these motives cannot be relied upon.

In this respect, the British colonial regime in India (the Raj) should have struggled to field effective military forces. The Raj was safeguarded primarily by Indian troops (Sepoys). Almost from the beginning of the East India Company’s time as a military actor in Indian politics to the end of British rule in 1947, British troops formed only a minority of the Raj’s military forces. Indian troops in British service, moreover, served as a spearhead for British imperial ambitions outside India, including service in Afghanistan, Burma (Myanmar), Africa and China, as well as the Middle East and Southeast Asia during the World Wars. Recruitment into this force, however, excluded a substantial proportion of India’s manpower, relying disproportionately on small ethno-religious groups such as the Sikhs, Gurkhas, Rajputs, Jats, Dogras and Pathans (Pashtuns) to the exclusion of most others (Streets, 2017).

However, this system, which might have been thought to produce an army of poorly motivated mercenaries who would desert at the earliest opportunity, instead produced a highly motivated army which was renowned for its acts of valour. At the battle of Saragarhi in 1897, for instance, 21 Sikh soldiers in British service defended a small outpost against several thousand insurgent Afghan tribesmen. Refusing offers to surrender, the Sikhs fought until to the death.¹

I argue that the martial race system represented an astute solution to the principal-agent problems inherent in recruiting and fielding an army, especially an army in which the agents and principals are drawn from radically different societies. Drawing on the logic of the club goods model of Iannaccone (1992), I show how the costly prohibitions inherent in the cultural and religious practices of many of the martial races helped to screen out low commitment types prior to their recruitment into the military. Further, I show how the repeat business model whereby the British serially recruited from the same small and tightly knit communities with few good alternative options to military service provided an incentive for those communities to monitor and sanction poor military performance.

In this article I first outline what the institution of martial race recruitment was and why it has so often been characterized by historians as irrational. I then recap the basic principal-agent and social dilemma dynamics which characterise the incentive problems facing a policymaker seeking to create a well-motivated army. I proceed to outline my adapted club goods theory of military recruitment before proceeding to describe how four of the group disproportionately recruited by the British - Brahmins, Sikhs, Rajputs and Gurkhas - fit into the theory. I then examine which groups were not recruited by the British and explain why. Finally, I examine two alternative explanations for the martial race system - that it was designed by the British to ‘divide and rule’ the Indian people (Farooqui, 2015) or that it was the result of Victorian racialist ideology (Streets, 2017).

This work represents a contribution not only to the economics of religion (Carvalho et al., 2019; Carvalho, 2013, 2016a, 2019, 2016b; Chaudhary and Rubin, 2016; Kuran, 2012; Iyigun, 2019; Johnson and Koyama, 2019; Platteau, 2017; Aimone et al., 2013; Aksoy and Gambetta, 2016; McBride, 2015, 2016; Iyer, 2016) but also to the growing literature on the economic organization of early modern institutions such as the British system of officer commission purchases in the army (Allen, 2011), the navy spoils system (Allen, 2011) and pirate crew organization (Leeson, 2007).

The martial race system

While the origins of the ethnic and religious groups which comprised the ‘martial races’ predate British colonization, the system of the ‘martial races’ was evident in the English East India Company (EIC)’s recruitment practices from early on in the company’s days as a military power on the Indian subcontinent. India had long had a burgeoning military labour market with skilled soldiers available for hire to the highest or most reliable payer. The EIC was divided into three ’Presidencies’ based on its three main bases in India - the Bengal, Madras and Bombay Presidencies (Streets, 2017). Each Presidency had its own army with separate recruitment practices. The Bengal army especially favoured the recruitment of high caste Brahmin (priestly caste) troops (Peers, 1995). These troops, however, provided the basis for the Indian Mutiny/War of Independence in 1857. After the rebellion was crushed by the British, the East India Company was wound up and replaced by direct British rule. The British Government reformed the Indian army and changed the basis of its recruitment from the Brahmin class to a number of ethnic and religious groups of primarily Northern Indian (including modern day Pakistani, Nepali and Afghan) origin - Sikhs, Gurkhas, Jats, Dogras, Rajputs, and Pashtuns (Streets, 2017). This created what might be termed the ‘classical’ martial race system, which dominated recruitment into the British Indian army from 1880 until World War Two (Roy, 2008).

The system was justified via a racial and gendered ideology which, while discordant to modern ears, was also contradictory and historically inaccurate even given the information available at the time. The martial race ideology sometimes stressed commonalities between the martial races composed of lighter skinned, Indo-European speaking ‘Aryan’ North Indians and the British themselves, which suggested racial superiority over darker, Dravidian language-speaking South Indians (Streets, 2017). Yet the ideology also stressed the supposed contrast between the ‘intelligent and cunning but weak and cowardly’ urban intellectual class of Indians and the ‘simple minded but brave and loyal’ soldiers of the martial races. The martial races were also seen as embodying ‘masculine’ virtues in contrast to the alleged ‘femininity’ of other Indian groups. As the British General Sir George MacMunn (1911), author of introductory textbooks on both the British Indian Army and the martial races for British officers, wrote:

“In the East, with certain exceptions, only certain clans and classes can bear arms; the others have not the physical courage necessary for the warrior …It is extraordinary that the well-born race of the upper classes in Bengal should be hopeless poltroons, while it is absurd that the great, merry, powerful Kashmiri should have not an ounce of physical courage in his constitution, but it is so …The gradual raising of the military standard by the change of times and terrain, and the enervating effect on Asiatics of a few generations of peace, have already been referred to. It has resulted in a gradual enlistment of an increasing number of men of the northern races, whom a cold winter, and a lesser period of enjoyment of peace, have for the present preserved from military deterioration.”

These racial and gender stereotypes have dominated previous scholarly attention to the martial races and have understandably created a strong impression of the irrationality of the system. Even given the knowledge available to the British at the time, it should have been clear that the alleged military deficiencies of substantial proportions of the Indian population were not grounded in fact. Indeed, the British themselves had recruited extensively from southern India and the Bengal upper classes before the mutiny (Streets, 2017). Moreover, restricting military recruitment to only small groups within the overall population would seem, on the face of it, to make little military sense, complicating manpower needs and cutting the British Indian Army off from huge pools of potentially talented and capable soldiers. As MacMunn (1911) himself recognized:

“The existence of this condition, therefore, much complicates the whole question of enlistment in India. It renders any form of levy en masse impossible, or any form of militia service.”

The question then arises as to why this seemingly wasteful practice was maintained by the British Empire in India for such a long time, breaking down only finally under the severe manpower pressures of the Second World War. One potential explanation is that the restrictions on recruitment to the martial races formed part of an internal political strategy to maintain British rule in India through a policy of ‘divide and rule’. As we shall see, while this explanation has some merit, it fails to account for many aspects of the case - why were some particular ethnic and religious groups chosen over others, for instance, and why aspects of the system were maintained by the independent Indian government after 1947.

Instead, I argue that the martial race recruitment system in fact represented an efficient solution to a set of military problems facing all armies, but especially those recruiting from a foreign country on which they have incomplete information. Below, I will outline my argument, before going on to show how it can explain the operation of the martial race recruitmemt system.

The army’s incentive problems

The principal agent problem

My argument begins with a description of the two of the main types of problems an army has to solve. The first is a principal-agent problem - the principal (the decision maker who raises the army) contracts with the agents (soldiers) to provide military services, where the agents’ actions are not perfectly observable by the principal and the outcome is a noisy function of the agents’ efforts. The principal wishes to get the maximum amount of effort from the agent (application in training, bravery in combat and so on) in order to achieve his or her goals in warfare. The agent by contrast wishes to obtain the maximum reward from the principal for the minimum effort (for example, risk of death or disability).

Holding the technology of monitoring fixed, screening can provide a means for resolving the principal-agent problem. If the principal can screen out unmotivated or unsuitable agents ex ante then his ability to extract maximum effort is enhanced, even if he cannot fully monitor the agent’s actions.

The social dilemma

At the same time, fighting in warfare is also a type of social dilemma amongst the agents. As Brennan and Tullock (1982) note, individual soldiers have strong incentives to flee the enemy, though the collectively rational decision would be to stand and fight - both soldiers would have be more likely to survive if the other stood and fought. Part of the problem of organizing an army is to find a solution to this social dilemma.

Armies have found various means of resolving this dilemma throughout history. Harsh discipline, including the death penalty for desertion or cowardice, transforms the structure of the soldier-agents’ payoffs (Chen, 2017). Ideological appeals, such as propaganda and indoctrination, can serve the same purpose (Barber and Miller, 2019). Small unit cohesion - that is, the formation of tight social bonds between soldiers - can also resolve the social dilemma, either by causing troops to attach positive utility to the well being of their comrades or alternatively by creating trust through repeated interaction (Costa and Kahn, 2003, 2009).

Another means by which the social dilemma of fighting might be resolved is through social pressure from the home front. If a soldier’s non-military social networks are both tight and motivated to sanction military underperformance, then this may also alter the soldier’s incentives. The prospect of ostracism at home significantly raises the costs of desertion or cowardice. Indeed, if the costs of overall military underperformance fall on a soldier’s home community, he will not only have incentives to exert greater effort, but also to monitor the effort of his peers, reducing the principal’s monitoring costs.

How do these two dynamics help explain the martial race system?

The club goods model

Iannacone (1992)’s club goods model helps to provide an answer. Iannacone attempted to provide an explanation for the apparently irrational prohibitions and sacrifices demanded by some religious sects. Iannacone viewed religious sects as clubs offering their members mutual aid but which required methods to prevent free-riding. Prohibitions and sacrifices serve both to screen out potential freeriders ex ante and to lower the value of their outside options relative to membership in the group. The club goods model has proven influential in the economic study of religion and has been adapted and modified in various directions since Iannocone (Aimone et al., 2013; Aksoy and Gambetta, 2016; Carvalho, 2013, 2016a, 2016b, 2019; Carvalho et al., 2017, 2019; Carvalho and Koyama, 2016; Chen et al., 2019; McBride, 2015, 2016). The club goods model has also been extended to conflict scenarios by Berman (2008), (2011), who used it to explain the success of modern day religious terrorist and insurgent groups such as the Taliban and Hezbollah.

The theory I present here is an adapted version of the club goods model in which a principal external to the club (the British Empire) effectively contracts a religious or ethnic ‘club’ to serve as its agent in war-fighting. The benefits are provided both to the individual soldier and to the club as a whole, with the club’s costly prohibitions and sacrifices co-opted by the principal to prevent shirking in battle.

Let us first examine how the principal might screen potential military recruits. The principal would wish to have a reasonable assurance that potential soldiers would be committed, disciplined, obedient and physically courageous. He would wish to know that each recruit is used to hardship and will be able to tolerate the regimentation and restrictions of military life. If the pay associated with the army is sufficiently attractive, of course, many potential recruits would have incentives to make these claims even if false.

If, however, the recruit is a member of a religious or ethnic group which requires costly sacrifices or prohibitions as the cost of membership, then this recruit can credibly signal to the principal his suitability for military service. Consequently, one potential solution to the principal agent problem is precisely to recruit the army disproportionately, even exclusively, from such groups.

Moreover, recruitment from one particular social group can help to resolve the principal agent and social dilemma problems. Suppose, for instance, that the army recruits repeatedly from a given village and pays well. The economy of the village may then come to depend on the income from military salaries. Essentially the village has become a specialist exporter of soldiers. It is then in the interests of the villagers as a whole to ensure that the soldiers perform satisfactorily in battle, lest the village’s principal income source be jeopardized. It will thus pay villagers, whether soldiers or not, to cooperate in imposing social sanctions on soldiers who do not perform in battle. If a soldier fails to fight, word of his conduct will be transmitted by his comrades back to his home village, where he will be ostracized, perhaps excluded from marriage alliances and a share of the village’s economic surplus. The opportunity costs of expulsion from the club are thus a crucial part of the story (Berman, 2000; Carvalho et al., 2017; Carvalho and Koyama, 2016). If the club itself is of high status, then the costs of exclusion from the club will be correspondingly high and the soldier will be even more constrained in his temptation to defect. Conversely, ejection from a lower status club is simply less of a deterrent. At the same time, however, the principal would not want to recruit from clubs whose average level of human capital was very high, since this would imply that the club’s members would have good options outside both the club and the military. Finally, the principal would wish to exacerbate the differences between the club and the rest of society, so as to further lower the value of club members’ outside options. The principal would therefore look to recruit from clubs which are high in status but low in human capital, with the status linked to military performance.

Most importantly, recruitment from club-like religious or ethnic groups allows the principal to tap into those sects’ established mechanisms for screening out and deterring free-riding. How might this be achieved? If military service comes to provide a substantial proportion of the club’s income, then soldiers from this sect who do not perform well in battle are consequently endangering the sect itself. By fighting for the principal, therefore, members of the sect are aiding their fellow sect members just as much as if they were helping to gather the communal harvest or providing each other with childcare. Poor performance in battle then becomes akin to the type of free riding within the group which the prohibitions and sacrifices are intended to prevent.

The clubs

Brahmins

The attraction of the British authorities towards a club goods model of recruitment in fact predates the Mutiny and the classical martial race system which followed it. Before the Mutiny, the British recruited the majority of Indian soldiers for the Bengal army from the (Hindu) Brahmin caste of northern central India, especially Bihar and Oudh. The Brahmins are the highest caste in the Indian social hierarchy and enjoy considerable social prestige, although Brahmins are not necessarily financially wealthy (Hutton, 1963). A Brahmin may, however, lose their high caste status in the event of transgressing caste prohibitions, an occurrence which would not only lead to a radical diminution of social prestige but also exclusion from marriage and kinship networks. This made the Brahmin ‘club’ highly attractive to the British initially as a source of soldiers. Peers (1995) notes that:

“Membership in these societies was consolidated through the adoption (or invention in some cases) of rituals intended to mark them out as exclusive and elevated above the rest of rural society. Such customs, in British eyes, came to define these recruits as being from the higher castes. British officers were convinced that the sense of honour that they identified as pervading these groups made for a superior recruit, one who would be impelled to obey and perform well by his own sense of self respect”.

Such was the enthusiasm of the British to reap the benefits of the club goods model of recruitment that they were willing to accommodate caste prohibitions even to the extent of allowing hiring servants for high caste Sepoys to perform prohibited tasks such as digging trenches or ringing guard bells (Peers, 1995). The British were also far more reluctant to use corporal punishment against high caste Indian soldiers than against British recruits, as this would be contrary to the former’s caste status (Peers, 1995). In spite of this, surviving statistics on disciplinary infractions from the 1830s Bengal show that British recruits accounted for the vast majority (or in some cases all) of the cases of desertion, absence without leave, mutinous conduct, mutiny, robbery, theft and striking an NCO, even though Indians formed a majority of recruits overall (Peers, 1995). So much better did British officers consider high caste Sepoy discipline to be that they often assigned Indian soldiers to watch over British soldiers’ barracks in case of indiscipline (Peers, 1995). The British also viewed the combat motivation of Brahmin soldiers favorably (Hodgson, 1850).

In the years prior to the Mutiny, however, the British had come to see some of the shortcomings of reliance on recruiting Brahmins. While the British appreciated the role that the Brahmins’ caste prohibitions played in weeding out low commitment types and thus fostering discipline and combat motivation, they also found many of the specific prohibitions, especially relating to manual labour and overseas deployment, detrimental to military efficiency (Streets, 2017). The ideal would be to recruit from groups which had similar club-like prohibitions which served a screening function without undermining troops’ performance of more mundane military duties or from serving overseas. After the British conquest of the Punjab at the end of the 1840s, the possibility of recruiting large numbers of Khalsa Sikhs offered a solution. British officers began to argue that Brahmin soldiers could be compelled to break caste prohibitions which interfered with their military duties and that if they were not prepared to do so, they could be replaced with Khalsa Sikhs and others (Hodgson, 1850). The trigger for the Mutiny can be seen then as the excessive demands which the British made on Brahmin troops with respect to breaking caste prohibitions, which they believed they could do because of the presence of an alternative source of military manpower in the form of the Khalsa Sikhs. Indeed, after the Mutiny was suppressed, the Khalsa Sikhs indeed came to replace the Brahmins as the mainstays of the British Indian Army (Streets, 2017).

Sikhs

The post-Mutiny British Indian army therefore consisted disproportionately of Sikh recruits. Sikhism is a syncretic religion which arose in early modern India under the teachings of a series of Gurus, beginning with Guru Nanak, who sought to combine elements of Hinduism and Islam (Dhavan, 2011). Many of the beliefs and prohibitions of Sikhism emerged under subsequent Gurus. In 1606, the fifth Guru, Guru Arjan, was tortured and executed by the Mughal Emperor Jahangir for refusing to convert to Islam. This event inaugurated an era of persecution for the Sikh community and gave rise to many of the characteristics which were ultimately to make Sikhs attractive soldiers for the British Raj. The sixth Guru, Guru Gobind Singh, established a Sikh militia in the late seventeenth century to protect the faithful and made martial accomplishment a key part of the faith (Dhavan, 2011). As the Mughal Empire’s authority began to decline, the Sikhs began to accrue greater military power, eventually forming a Punjabi Sikh Empire centered on Lahore. In the early nineteenth century the Sikh Empire clashed with the rising power of the British East India Company in the Anglo-Sikh Wars (Dhavan, 2011). The British were victorious but were impressed with the fighting skills and motivation of the Sikh army, which they began to incorporate in the East India Company armies (Streets, 2017). The Sikhs remained loyal to the British during the Indian Mutiny of 1857, primarily because they feared that the success of the Mutiny would lead to the reestablishment of the Mughal Empire which had persecuted their faith in its early years (Imy, 2019). Although the Sikhs therefore came to be seen as quintessentially loyal by the British, their support for the Raj was therefore more a case of preferring the lesser of two perceived evils, and many Sikhs continued to yearn for the independent Sikh Empire of the pre-colonial Punjab (Imy, 2019).

The religious militia established by Guru Gobind Singh - the Khalsa - has many of the attributes of a club in the Iannaccone sense. To join the Khalsa, Sikhs must take part in a ritual initiation known as the $p \bar{a} h \overset{͝}{u} l$ . Khalsa Sikhs are then required to wear the ‘five Ks’ - $k \bar{e} s$ (uncut hair), $k a n g h \bar{a}$ (hair pin), $k a r \bar{a}$ or iron bracelet, kirpan or small sword and kachera or undergarments. The turban, for which the Sikhs are perhaps best known, is designed to keep in place the long, uncut hair. These visible signs of religious affiliation would, of course, mark the Khalsa Sikh out for persecution if living in a hostile environment (in which historically Sikhs often did) and hence serve as a costly signal of devotion to the faith. The long hair, for instance, served to differentiate many Khalsa Sikhs from non-Khalsa members of their own families. Caste traditions in the Sikh heartland of the Punjab expected that family members should shave their head during the period of mourning for a deceased relative (Dhavan, 2011). Khalsa Sikhs who foreswore this custom were subsequently ostracized and boycotted by their own non-Khalsa family members (Dhavan, 2011). Khalsa Sikhs are also expected to observe a number of prohibitions, including those against intoxication, adultery, slanderous talk and the excessive accumulation of material goods (Dhavan, 2011).

As a result of the persecution of the religion by the Mughals, Sikhism developed a strong martial strain, symbolized by the inclusion of the kirpan amongst the five Ks and reinforced in the Sikh prayer or Ardas which recites the bravery and sacrifices of Sikhs for the faith in the past, and states that ‘victory belongs to God’ (Dhavan, 2011; Streets, 2017). As Heather Streets notes:

“According to the teachings of Guru Gobind Singh, extreme courage and/or death in the heat of battle was said to bring honour to the whole community. Conversely, cowardice in battle was considered unacceptable and unmasculine, and those who demonstrated it were likely to face the contempt of family or of entire villages.”

An individual wearing the distinctive clothing of a Khalsa Sikh was therefore sending a credible signal of their adherence to a faith which preached bravery in battle and contempt for death. In this light, it is not therefore surprising that the British authorities found it attractive to recruit disproportionately from the Sikh community. Indeed, the British would recruit only Khalsa Sikhs precisely for this reason. Otherwise identical Sikhs who did not take on the costly prohibitions of the Khalsa were rejected, as it was believed they were not credibly committed to fight. As MacMunn notes:

“A Sikh is baptised into his sect and not born into it, so that no man is a (Khalsa) Sikh till he has taken the $p \bar{a} h \overset{͝}{u} l$ , that is to say, till he has been baptized. Sikhism is an austere faith, demanding some simplicity and rigour of life from its adherents. So much is this so, that for many years there has been a tendency for young men to avoid the $p \bar{a} h \overset{͝}{u} l$ and grow up as ordinary Hindus, for whom life has few irksome restrictions. But as the value of the Sikh as the simple, faithful soldier, has lain in his adherence to the simple tenets and hardy life of his forebears, no non-baptized Sikh is admitted into a regiment of the Indian Army. So careful are the regiments in this matter, and so much are regiments the home of the old martial and simple Sikh principles, that it has been said, not without some shadow of truth, that it is the British officer who has kept Sikhism up to its old standard.”

Once in the British Indian army, every effort was made to maintain the prohibitions of the Sikh religion. As Streets writes:

“Within each Sikh regiment, daily ritual and prayer meetings based on the Guru Granth Sahib - the Sikh holy book - were intended to keep the faith of the Khalsa strong. In addition, the Granth was given a place of honour in regimental quarters, and was carried at the head of the regiment on the march.”

As the above quotes suggest, the British also attempted to harden the differences between the martial races and other Indians - for instance the difference between Sikhs and Hindus, or the differences between Rajputs and Jats on the one hand, and other peasant populations on the other, attempting to turn what had been relative fluid groups into fixed categories (Imy, 2019; Omissi, 2016). While this has traditionally been seen as another aspect of ‘divide and rule’, this phenomenon is also consistent with the club goods model as it reduces the outside options of club members, and has been observed in other, non-colonial contexts such as Orthodox Judaism (Carvalho and Koyama, 2016).

At the same time, British recruitment of Sikh troops had many of the characteristics of the repeat business model. The majority of Khalsa Sikhs were of an agricultural Punjabi background and were vulnerable to drought induced agricultural depressions. Consequently, a key motivation for Punjabi Sikh recruitment into the British Indian army was to generate cash revenue which could help to smooth over business cycles and prevent families at home from falling into debt (Streets, 2017). Streets (2017) notes that ‘for many Sikhs living in the area (the Punjab) enlisting in the army was one of the primary ways by which they could secure the financial solvency of their families. Contemporary observers noted that the ‘most important point’ for Sikhs in the army was ‘whether he can put by a sufficient sum from his pay’ and that Sikh soldiers had a habit of nearly starving themselves to save money (2017). Omissi (2016) writes that “some villages depended almost entirely on military income, most of their able bodied men having joined up”. So attractive was service in the British Indian army in terms of pay and prestige that many Indians are understood to have converted to Sikhism in order to increase their chances of being recruited (Streets, 2017).

There is strong evidence that villagers on the home front came to sanction and, if possible, monitor martial race soldiers’ military performance in order to protect a key portion of local community’s livelihoods. As one Sikh civilian wrote to a fellow villager on the front in World War One (Streets, 2017):

“I shall be very pleased to hear of your valorous deeds. You are a brave soldier. Now is the time to display your manhood. Now is the time for loyalty. You are a true Sikh. By the Guru’s order you must remember the promise of the Almighty, who said: Recognize the hero in him who fights for the faith; Though cut to pieces he will not quit his ground”

Similarly, the mother of a trooper in the Poona Horse wrote:

“I shall be well pleased with you, if you do not turn your back on the enemy, so that our neighbours may not taunt us.” (Omissi, 2016)

Streets (2017) writes that ‘Sikh soldiers were expected by their home communities to live up to the ideal of the Khalsa soldier’. She quotes a number of letters from Sikh soldiers (mostly from World War One) demonstrating awareness of this fact, for example:

“We do not wish the war to stop yet. We want it to continue. You will understand our worth when we return with our chests full of medals.” (Streets, 2017)

“We will do our best to uphold the family traditions and the reputation of our tribe.” (Streets, 2017)

“This is the time when he who desires to do so may illuminate his name and his clan by sacrificing himself.” (Streets, 2017)

“If we get the chance we will show the stuff our caste is made of.” (Omissi, 2016)

The Khalsa Sikhs therefore fit the club model very well. The prohibitions and injunctions which Khalsa Sikhs had to obey - especially the five ks and particularly the long hair, turban, bracelets and swords - served to visually distinguish them from other Indians and serve as a costly demonstration of their devotion to the faith which also prized valor in battle. This made the Khalsa Sikhs immensely attractive military recruits for the British, and in return for providing these benefits this club received extensive financial benefits and social prestige.

Rajputs

In addition to the Sikhs, the British also recruited from a number of religiously mixed ethnic/caste groups such as the Rajputs, Jats and Dogras. The Rajputs, in particular, were described as the ‘noblest and proudest in India’ by the British, being likened to a ‘military aristocracy of the feudal type’ and were especially favoured in military recruiting.

The origins of the Rajputs are not clear, but they are believed to have emerged from the warrior castes of northern India. In fact, the British considered Rajput to be synonymous with the kshatriya warrior caste which they traced back to India’s earliest times (Bingley, 1899 p.3). In medieval India Rajputs had constituted a travelling warrior caste, which sometimes fought and sometimes served invading Muslim regimes (Kolff, 1990). The Rajputs from whom the British recruited were composed of a number of clans residing in particular areas of northern India (Rajputana) (Bingley, 1899 p.27). Although Rajputs formed an elite, high caste group, by the late nineteenth century the Rajputs had lost ground economically as their traditional role as tax farmers had been circumvented by Sikh and British rulers, their villages were built on infertile high ground for defensive purposes and they considered farm labour to be demeaning (Omissi, 2016). Rajput villages consequently came to be highly dependent on employment as soldiers by the British for their economic viability. For example, Kangra district sent 12,000 Rajputs to the Great War from a total Rajput male population of 56,400 (Omissi, 2016).

While the Rajputs were subject to fewer restrictions than the Brahmins, who constituted the priestly caste, they had to obey a number of costly prohibitions which, if they were breached, could result in a loss of caste and social status. A young man from a given Rajput clan could only marry into a certain set of other Rajput clans, not including his own. The Ahban clan of Oudh, for instance, could only marry women from the Gaharwar, Chandel, Raikwar, Janwar, Gaur, Sombansi, Dhakre and Nikumbh clans, though Gaharwar men, for instance, could not marry Ahban women (Bingley, 1899 p.29-31). Marriages were arranged between families by professional matchmakers (Bingley, 1899 p.150). A higher dowry or bride payment could procure a marriage with a woman of higher status from a suitable clan (Bingley, 1899 p.150). Marriage ceremonies were also expected to be extravagantly expensive, with a British author complaining that ‘the whole savings of a lifetime are wasted over a single ceremony’ (Bingley, 1899 p.151). The Rajputs also obeyed a number of prescriptions around cooking and cleaning - eating only once a day, for instance, and bathing every time prior to a meal (Bingley, 1899 p.165). Non-Rajputs, or even Rajputs of non-allied clans (that is, clans into which a Rajput could not marry) were considered to make food or drink ritually unclean and therefore Rajputs would often cook for themselves individually (Bingley, 1899 p.164-5). Consistent with frequently observed religious prohibitions/injunctions related to shaving (Berman and Laitin, 2008), Rajput men were expected to shave their underarms (Bingley, 1899 p.167). Rajput women were often subjected to the expectation that they would sacrifice themselves on a ceremonial funeral pyre in the event of their husband’s death (sati) or of defeat in battle, to prevent their becoming victims of sexual violence at the hands of the winners (johur), though both of these customs were suppressed by the British (Bingley, 1899 p.170-171).

The youngest son of a Rajput family was traditionally expected to leave his home village to earn his way as a soldier, even before the British era. The Rajputs, as the putative descendants of the warrior caste, were like the Sikhs expected to display valour in battle. For Rajputs, remittances from the military could help his family pay higher bridal prices to secure better matches and to attain greater social prestige via more lavish wedding ceremonies.

So attractive were the rewards to Rajput soldiers in British service that many non-Rajputs attempted to falsely claim Rajput status in order to be able to join Rajput regiments (Bingley, 1899 p.177). The British thus carefully noted the best recruiting grounds for ‘true’ Rajputs. Bingley (1899) (Appendix C), for instance, noted that the Hissar and Rohtak Districts of Delhi division ‘furnish some of the best Rajputs in the service’, while Rohilkund division was ‘on the whole a bad ground, as a large number of the Rajputs are spurious’.

The British also took great care to ensure that a given recruit was a ‘true’ Rajput. AH Bingley’s Handbook on Rajputs outlines a set of questions which British recruiters should ask potential recruits to determine whether they were genuinely Rajput or not (Bingley, 1899 p.177):

1. “What clan does he belong to and what district?

2. What is his gotra (clan)?

3. What clan did his mother come from?

4. From what clans have male members of his family taken wives?

5. Into what clans have female members of his family been married?

6. What food will he eat, by whom cooked and from what vessels he will drink?”

Recruiters were also instructed to examine the recruit’s janeo, an item of Hindu religious clothing, with Rajput janeos being shorter than those of Brahmins and worn only after marriage.

As with the Sikhs, the Rajputs ‘club”s prohibitions served to make the club members attractive military recruits for the British. The reluctance of Rajputs to engage in non-military occupations such as agriculture, for instance, reduced their outside options relative to soldiering, while the tight social control which Rajputs exercised over marriage choices, for instance, gave these communities strong power over young Rajput men. Poor performance on the battlefield could lead to complete loss of caste or at least exclusion from favorable marital prospects. At the same time, Rajput communities’ economic precarity made the pay offered by the British and remitted home by recruits highly attractive.

Gurkhas

Along with the Khalsa Sikhs, the Gurkhas formed one of the major stalwarts of the British Indian army (Imy, 2019). The Gurkhas were recruited from Nepal, which was not formally part of the British Empire and culturally distinct from most of India. The term ‘Gurkha’ referred to Nepal’s ruling dynasty, but the majority of recruits into British Gurkha regiments were derived from two Nepalese ethnic groups - the Gurung and the Magar (Vansittart, 1906 p.77).

The Gurung and the Magar resembled the Rajputs in the sense that they were characterized by high prestige tied into military service, low income and low human capital (almost all recruits into Gurkha regiments were illiterate (Vansittart, 1906 p.61)). Nineteenth century Nepal was transitioning from Buddhism to Hinduism and was partially adopting Hindu customs, including the caste system (Vansittart, 1906 p.11, p.49). This meant that Gurung and Magar men could lay claim to high status which could nevertheless be rescinded by the community in the event of transgressing social norms (Vansittart, 1906 p.48). At the same time, however, the Gurung and Magar did not adopt some of the caste prohibitions characteristic of Indian Brahmins which had made military service difficult, such as the prohibition on manual labour (Vansittart, 1906 p.57).

The Gurung were divided into two major groupings - the Charjat and the Solahjat. The Charjat were considered the higher status group (for instance, a Solahjat would carry a Charjat’s load if travelling together) (Vansittart, 1906 p.80). A Charjat could, however, be outcasted and reduced to the status of a Solahjat for various offenses (Vansittart, 1906 p.80). Although ‘some excellent recruits’ could be found amongst the Solahjat, the British prefered Charjat recruits, even to the extent of noting which families produced the best recruits from within each Charjat clan (Vansittart, 1906 p.79).

The Magar were similarly divided into a number of tribes. Of these tribes, the British were especially keen to recruit from the $T h a \bar{a} p \bar{a} s$ and the $R \bar{a} n \bar{a}$ . The $R \bar{a} n \bar{a}$ were a group of families believed to be defined by the fact that they were of $T h \bar{a} p \bar{a}$ descent and descended from three consecutive generations of men who had lost their lives in battle (Vansittart, 1906 p.88). Thus a $R \bar{a} n \bar{a}$ Gurkha would be motivated to fight by a desire to uphold family prestige and a $T h \bar{a} p \bar{a}$ by the possibility of having his family elevated to $R \bar{a} n \bar{a}$ status. By contrast, the $P \bar{a} r \bar{e}$ tribe were ‘never to be enlisted’, as they were low status outcastes or descendents of outcastes (Vansittart, 1906 p.88).

Gurkha recruitment was also heavily characterized by the repeat business model. Recruitment into Gurkha regiments was the responsibility of experienced serving Gurkhas, usually NCOs (Vansittart, 1906 p.154) They were dispatched by the British to Nepal to recruit initially from their home villages and then from close neighbours. The recruiters were rewarded with cash bonuses for each recruit and favorable treatment in terms of promotion. The new recruits then normally served in the same units as their recruiters, both incentivizing the recruiter to use his local knowledge to recruit good soldiers and the recruit to perform well, since poor performance could be transmitted back to his home village (Vansittart, 1906 p.153).

The financial incentive for Nepalese villages to supply recruits for British service was clear. The majority of Gurungs and Magars were precarious subsistence hill farmers (Streets, 2017). For Nepalis, recruitment into Gurkha regiments could supplement very low incomes. This could be especially attractive given that debt peonage and slavery were both common in contemporary Nepal (though the latter could only be imposed as a punishment for certain offenses against caste (Vansittart, 1906 p.52-53)). So important was military service to the Nepalese economy that Streets (2017) writes that, until very recently, Gurkha soldiers’ remittances were Nepal’s largest source of foreign currency.

The Gurkhas therefore exhibited many of the same characteristics as the Rajputs. The Gurkha regiments were disproportionately recruited from groups in Nepalese society (such as the $T h \bar{a} p \bar{a}$ ) which displayed many of the same characteristics that Rajputs did in India - high prestige tied in to a reputation for martial valor combined with low and precarious economic prospects. A Gurkha soldier who performed poorly in battle would face social ostracism at home which would, as with the Rajputs, exclude him from favorable marriage prospects among other social benefits. Nepalese communities were incentivised to impose these penalties by the crucial importance of continued business with the British in supplying effective soldiers.

Why not other groups?

The question remains, however, as to why the British did not recruit from other groups within Indian society which also maintained sacrifices and prohibitions similar to those of the Sikhs or the Rajputs. While the extent and rigidity of the historical caste system remains a subject of debate, there is no question that many if not most groups within Indian society also maintained caste prohibitions and sacrifices similar to those of the Rajputs or Khalsa Sikhs. Why did the British not also recruit from these groups?

For some castes, the answer is that their adherence to caste prohibitions did not serve to signal their dedication to military performance but rather to other roles within society - the Brahmins, for instance, to priestly/scholarly pursuits or the Vaishas (farming caste) to agriculture and trade. It is not clear, for instance, that a Vaisha recruit would be subjected to the same kind of social pressure to perform in battle as a Rajput.

Consistent with the club goods model, the groups which the British designated as martial races tended to have lower than average human capital, as measured by literacy. For instance, the Punjab, from which the martial races disproportionately came, had a adult male literacy rate of 6.7% compared to the national average of 8.7% in 1901 (Omissi, 2016). Indians often complained that the British grouped Indians into two types of individual - the martial races who were ‘brave but stupid’ and the educated intellectuals who were ‘clever but cowardly’ (Tyagi, 2009). The British reluctance to recruit well educated intellectuals from the urban classes into the military has often been seen as a result of fears that such individuals would also be more prone to Indian nationalism. However, the British were also reluctant to recruit other groups with similarly high outside options – for instance, groups which specialized in urban trade, even if, as in the case of the Khatri group, they were (non-Khalsa) Sikhs (Imy, 2019).

Yet this does not explain why the British did not choose to recruit from other warrior castes (termed kshatriya) besides the North Indian Rajputs. The British did not, for instance, consider any South Indian groups to form part of the martial races and rarely recruited South Indians into combat roles. Yet South Indian society had traditional warrior groups who had performed military service for pre-colonial kingdoms. Why were these not incorporated into the British Indian army or considered to be ‘martial races’?

The answer lies in the ambiguous position of the kshatriya caste in South Indian society. Whereas the Rajputs are recognized as kshatriyas in North Indian society and thus accorded high social status, the status and prestige of traditional warrior groups in South Indian society is much less clear. Many parts of South India do not recognize kshatriyas as part of the caste system at all, preferring instead a simpler twofold caste distinction between the priestly Brahmins and the rest of society (Fawcett, 1901). Indeed, a British census superintendent in 1891 stated that: ‘of course, there is no such thing as a Dravidian kshatriya’ (Thurston, 1909).

Many of the traditional warrior groups of South Indian society claimed kshatriya status but these claims were treated with scepticism by other South Indians and by the British (Thurston, 1909). The reason was that many South Indian warrior groups had over time lost (or had never followed) many of the prohibitions and sacrifices which marked them off as kshatriyas and had come to be considered part of the lower or ‘polluting’ castes.

For example, the $S \bar{a} m a n t a s$ caste of Malabar in South-west India were a Dravidian people who claimed kshatriya ancestry. “Tradition,” wrote the Travancore Census Superintendent in 1901, “traces the $S \bar{a} m a n t a s$ to the prudent Kshatriyas, who cast off the holy thread, to escape detection and slaughter by Parasu $R \bar{a} m a$ . They are believed to have then fled to uninhabited forests till they forgot the $S a n d h y \bar{a} v a n d a n a$ prayers, and became in certain respects no better than $S \bar{u} d r a s$ . Thus they came, it is said, to be called Amantrakas, $S \bar{a} m a n t r a k a s$ , $S \bar{a} m a n t a s$ , or having no mantra at all”² (Thurston, 1909). The $S \bar{a} m a n t a s$ s, and other South Indian warrior groups, therefore, had either lost or never first possessed the costly sacrifices and prohibitions characteristic of the martial races (Thurston, 1909). Similarly, the Tamil-speaking $S h \bar{a} n \bar{a} n$ group claimed kshatriya status but were considered by others as a low status polluting caste (Thurston, 1909). The British anthropologist Edgar Thurston (1909) nonetheless noted that ‘ it is by no means certain that the $S h \bar{a} n \bar{a} n s$ were not at one time a warlike tribe, for we find traces of a military occupation among several toddy-drawing³ castes of the south, such as the Billavas (bowmen), $H a l \bar{e} p a i k$ (old foot soldiers), $K u m \bar{a} r a p a i k$ (junior foot). Even the Kadamba kings of Mysore are said to have been toddy-drawers.”

The reason why South Indian castes claiming kshatriya descent could not have formed part of the martial races therefore is not that they would have been more likely to oppose British rule. Indeed, the $S h \bar{a} n \bar{a} n s$ for one were widely praised by British clergy for their adoption of Christianity and of British social values (Thurston, 1909). Rather groups such as the $S h \bar{a} n \bar{a} n s$ were unattractive to the British as military recruits because their opportunity costs of military underperformance were lower than those whose warrior status was recognized by the wider society such as the Rajputs. A $S h \bar{a} n \bar{a} n$ who deserted, for instance, may very well lose caste in the same way as a Rajput would, but this would constitute a far smaller loss as the $S h \bar{a} n \bar{a} n s$ had little prestige to begin with.

A similar logic explains why the British did not recruit urban or outcaste populations. As Omissi (2016) notes: “The Indian army also took hardly any men from the towns and cities. The British saw the urban lower classes as mixed, degenerate and potentially threatening. The shifting urban population could not be classified, ordered and ‘placed’ as easily as that of rural India. “The inhabitants of Hyderabad City’, remarked one handbook, ‘are of such doubtful character that it would be dangerous to experiment in enlisting them’. Similarly, Hodgson (1850) argued that: ‘men of low and degraded caste …would never answer as soldiers. Despised through countless ages, they have never felt the sentiment of self-respect. When no disgrace can humiliate, it is folly to imagine that military glory would be sought at the peril of existence.’

The British did, in fact, attempt to broaden the basis of Indian army recruitment by identifying more ‘martial races’ in Southern India (Omissi, 2016). Specifically, the British focused on two Southern Indian groups - the Mappilas and the Coorgs (Omissi, 2016). The Coorgs, a traditional South Indian warrior caste, proved hard to recruit given their family and inheritance systems. The Mappilas were a low income Muslim group from Malabar in southwestern India who were notorious for their violent attacks on landlords, agents and Hindus more generally. Some British officials believed that this reputation for violence could be harnessed to military purposes and raised two Mappila battalions in 1900. However, the experiments was a failure as the Mappila troops were notorious for indiscipline and the two battalions were disbanded in 1907 (Omissi, 2016). Given the logic of the club goods model this is understandable - unlike the Rajputs or the Sikhs, for instance, the Mappila community did not levy social sanctions for indisciplined behaviour (as shown by their propensity for violence prior to 1900), nor would a loss of British military employment be a source of shame for a Mappila man, as the Mappilas had developed an ‘anti-landlord, anti-British’ mentality, fostered by traditional Islamic intellectuals. While this case may appear to show the importance of internal security considerations in recruitment, it is clear from Omissi’s discussion that the Mappilas were considered unsuitable for service not because they posed a political threat to British rule but rather because of their tendency to indiscipline in the form of petty crime, including what was termed addiction to ‘unnatural crimes’ (Omissi, 2016).

Alternative explanations - divide and rule?

An alternative explanation for the martial race policy is simply that it represented a strategic policy of divide and rule to maintain British control of India through the provision of selective side benefits to small groups of specialists in violence (Farooqui, 2015). There is some evidence for this point of view:- most obviously the removal of the traditional Brahmin castes from recruitment into the army and their replacement by groups such as the Sikhs and Gurkhas who had mostly remained loyal to the British during the Rebellion. However, a close examination of British decision making after the Rebellion reveals a more complex story.

After the mutiny, the British established two commissions to enquire into the causes of the Mutiny and to make recommendations for the structure of the British Indian Army - the Peel Commission and the Lawrence or Punjab Commission (Streets, 2017). The two commissions in fact came to diametrically opposed views as to what type of recruitment system was best suited to maintaining British rule. The Peel Commission criticized the club goods solution of relying on Brahmins from Bihar and Oudh and the Bengal Army’s ‘indulgence’ of their ‘superstitious’ religious and cultural prohibitions. The Peel Commission believed that this had created an ‘arrogance’ and ‘insolence’ on the part of the Sepoys which lay at the heart of the rebellion and that the best option to maintain British rule was to broaden the army’s recruitment base (Streets, 2017). By contrast, the Lawrence Commission believed that broadening the army’s recruitment base and integrating different ethnic groups and castes at the small unit level would produce a pan-Indian consciousness which would endanger British rule (Streets, 2017).

In short, then, British decision makers were divided as to whether a club goods model of recruitment would be optimal for maintaining British rule against internal challenges. The evidence, in fact, would at the time have appeared to support the idea that a broader rather than a narrower recruitment base would best serve the interests of preventing further rebellions. Indeed, as Roy (2008) argues, the Raj’s predominant recruitment strategy from the end of the Mutiny to 1880 was a ‘balanced’ one in which the army integrated different racial and religious groups even down to the company level in many cases. As one British advisor to the colonial government, CB Pritchard, later wrote: “Class regiments are more combat effective against external threats, but from the point of view of internal insurrections, it is dangerous to convert the class company and mixed regiments of the Bombay and Madras armies into class regiments recruited from upper India” (Roy, 2008).

The reason why the British believed class regiments (that is, regiments recruited from a single ethnic or religious group such as the martial races) to be superior in fighting ability to mixed regiments in part lay in the combat performance of the Madras and Bombay armies, which recruited ethnically mixed regiments. While the Madras and Bombay armies had not participated in the Mutiny, their combat performance against external enemies was judged wanting (Streets, 2017). During the Third Anglo-Burmese War of 1885, for instance, the Madras army had performed poorly. As Omissi (2016) notes, during one attack the 10th Madras Infantry were ‘very ready to lie down and extremely difficult to get up’ and on another occasion the 10th Madras Infantry were reluctant to advance even though they were not under fire’. After a difficult retreat, the same regiment became ‘utterly done up and demoralized” on the march the men had tried to hide behind their baggage animals and had ‘fired all the way at nothing from funk’. By the end of the campaign, the Officer Commanding had concluded that ‘in all regiments of the Madras Infantry there are very many men whom it is an absurdity to call soldiers’. Explanations for this phenomenon range from the Madras army’s lack of combat experience, to its inability to recruit more able British officers and the advanced age of many Madrassi sepoys and (at the time) to racialized theories of South Indian soldiers’ innate weakness (Omissi, 2016; Roy, 2008). However, many British officers attributed the Madras army’s poor performance precisely to its mixed recruitment system which prevented the formation of an esprit de corps (Roy, 2008).

Ultimately, the British decided to return to a club goods model of recruitment but simply switched the ethnic and religious groups who formed the backbone of this model - from the Bihari and Oudh Brahmins to the Sikhs, Rajputs, Dogras, Gurkhas and other martial races. Moreover, as the examination of the statements of British commanders at the time make clear, the reasons for the adoption of this more exclusionary policy were twofold. First, technological innovations such as the steamship and the telegraph gave the British more confidence in their ability to put down a new Mutiny since they could rapidly request and receive reinforcements from the rest of the Empire (Roy, 2008). The British therefore perceived a lower degree of internal threat in 1880 than in 1859. Second, the advance of the Russian Empire to the borders of Afghanistan created a renewed sense of external threat to the Raj. Indeed, the perceived greater internal risk to the security of the Raj from the martial race system was seen as a necessary evil in return for the greater effectiveness which it would bring to the British Indian Army against the Russians. In this regard, the words of Field Marshall Sir Frederick Roberts, the Commander in Chief of the British Indian Army, are especially noteworthy. As Roberts asserted in 1890 (Streets, 2017):

“I have no hesitation in stating that except Goorkhas, Dogras, Sikhs, the pick of Panjabi Muhammadans, Hindustanis of the Jat and Ranghur (Muslim Rajput) castes …and certain classes of Pathans, there are no native soldiers in our service whom we could venture with safety to place in the field with the Russians”

Crucially, Roberts made the case for basing recruitment on the martial races in spite of the internal threat they posed to British rule, not because he perceived them to be a defence against this threat (Streets, 2017):

“Should it be urged that these measures would have the effect of increasing to a dangerous extent the more warlike races in our armies, and that there would no longer exist the same counterpoise to this element which is now afforded by having in our ranks a certain proportion of the races of Hindustan and southern India, I would reply that whatever weight this argument may have, it sinks into insignificance when compared with the peril of confronting a European enemy on our frontier, without an army fit in all respects to cope with him.”

Of course, internal political considerations were not irrelevant in deciding which groups would form the basis of army recruitment. Active evidence of hostility to British rule from a group, such as the behaviour of the Brahmins in the Mutiny or the Mappilas in the early twentieth century, would exclude a group from consideration for recruitment. But ideological affinity with British rule was neither a necessary nor a sufficient condition for inclusion in the ‘martial races’ if it conflicted with combat ability.

The British also excluded from the martial races many of the groups in Indian society who should have been most attached to British rule. After the Mutiny the British considered recruiting heavily from Eurasians (that is, children of mixed British and Indian descent) and Indian Christians (Roy, 2008) but ultimately discouraged and even in some cases prohibited Indian Christians from serving in the combat branches of the army (MacMunn, 1911). Similarly, while Eurasians formed important units at some points in the history of British India - for instance the Skinner’s Horse cavalry regiment during the Mutiny - Eurasians were generally not recruited into the British Indian military and instead typically found employment in lower level clerical positions especially in the railway service (MacMunn, 1911). In the case of Indian Christians, they had either abandoned the costly prohibitions and sacrifices required of groups such as the Rajputs or the Sikhs or more often belonged to outcaste groups which did not engage in such practices even prior to conversion (MacMunn, 1911). At most, the British required from the martial races an absence of active hostility (as in the case of the Khalsa Sikhs), rather than a positive attachment, to British rule. Even this minimal requirement could be qualified - as we have seen, the British recruited extensively from among the Afghan Pashtuns, even though this group were one of the Raj’s most longstanding adversaries (Ridgway, 1910).

Thus the British came to rely on the martial race recruitment policy not because they believed it to be the most effective means to ‘divide and rule’ Indian peoples, but rather because they believed it to foster military motivation and effectiveness, even if it increased the internal security risk.

Moreover, if the martial race recruitment policy were a divide and rule tactic designed to hamper Indian moves towards independence, why would the post-independence Indian government have kept substantial proportions of the policy? Since independence, the Indian government has relied disproportionately on Sikhs and other martial race peoples (despite the threat of Punjabi Sikh separatism) in its military recruitment system. Moreover, martial race regiments were, on independence, not disbanded but simply transferred over into the Indian army. Thus, for instance, the Rajputana Rifles, the Dogra Regiment, the Jat Regiment and six Gurkha regiments were directly incorporated into the new Indian army (Tyagi, 2009). This stands in contrast, for instance, to the Republic of Ireland, which although incorporating many former British army soldiers on independence, disbanded the principal Irish regiments of the British army such as the Connaught Rangers or the Royal Dublin Fusiliers (Morrissey, 2017).

A more subtle variant of the divide and rule theory is that the British wished to base their rule on arming small, distinct minorities who differed from the broad mass of the Indian population and would therefore be implausible alternative rulers. This dynamic has been observed in many authoritarian states. While this explanation would be plausible for some of the martial races who were ethnically or religiously distinct from the majority of Indians - such as the Gurkhas or Sikhs - it is less plausible in the case of predominantly Hindu martial races such as the Rajputs. Although the Rajputs were socially distinct from most Indians, this was far from inconsistent with ruling legitimacy, as the Rajputs were historically considered to be the ruling class of northern India (Kolff, 1990).

Another alternative explanation is that the British genuinely believed the pseudo-Darwinian racial ideology behind the martial race recruitment policy. Certainly, the pronouncements of British policymakers including Roberts and MacMunn are replete with racial ideology - the superiority of the more ‘Aryan’ Northern Indians over the Dravidian South, the allegedly beneficial racial effects of a cold and harsh climate and the equation of the martial races with ‘masculinity’ as opposed to other Indian peoples and references to pseudo-sciences such as phrenology (Roy, 2008; Streets, 2017). The club goods model advanced here does not require that British policymakers fully understood club goods logic, although the practice of Sikh regiments discussed by MacMunn or the Rajputs by Bingley suggests that the British were on some level aware of it. Rather, the club goods model explains why the martial races were motivated and effective fighters. This was apparent in the military history of these peoples in combat against the British, as in the Anglo-Sikh or Anglo-Nepalese Wars, or indeed before the British arrived in the sub-continent (MacMunn, 1911). The British, realizing that these peoples enjoyed a comparative advantage in warfare, employed them in the army and retrofitted a racist ideology to explain this performance after the fact, whereas the club goods model represents a superior explanation of the same phenomenon without relying on innate racial differences.

Conclusion

In this paper I have outlined a novel theory to account for the phenomenon of the martial race recruitment policy in British India. Previous research has seen this policy either as an expression of irrational racial ideology or of a politically driven divide and rule strategy. Drawing on Iannacconne’s club goods model of religious sects, I have shown that the martial races policy was in fact a rational solution for the British to the problem of fielding a motivated and effective colonial army. The martial races were in fact ethnic and or religious clubs characterized by costly sacrifices and prohibitions which served both to screen out and deter potential shirkers. The British were able to tap into the effectiveness of these clubs by preferentially hiring their members for military service to the point where the clubs themselves came to be dependent on the material and prestige goods the Raj provided. Soldiers from the martial races could be relied upon to fight since they were both pre-screened for loyalty to their club and deterred from shirking by the prospect of expulsion from the club. The British refused to hire from other groups within Indian society which were characterized by a lack of such costly club-like prohibitions, an orientation towards non-military vocations or lower prestige. Consistent with the club goods model, and inconsistent with the divide and rule theory, this extended to barring recruitment into the military for groups within Indian society which should have been expected to be the most favorable to British rule, such as Indian Christians and Eurasians. Also consistent with the club goods model and inconsistent with divide and rule is the fact that the Republic of India continued to recruit disproportionately from the martial races even after independence from Britain. Of course, internal security was not irrelevant to the British - many groups such as the Mappilas who might otherwise have provided military manpower were excluded because of a tradition of active hostility to British rule. Active affinity with British rule, however, was neither a necessary (as the case of the Khalsa Sikhs or the Pashtuns shows), nor a sufficient (as the case of the Eurasians and Indian Christians shows) condition for consideration as a major source of manpower for the Raj.

This study has applications beyond the history of British India. Studies of authoritarian rule, for instance, stress the importance of rent seeking in recruitment to the security services and military. Yet this study suggests that this dynamic may not be the whole story. While all groups in society would like to gain rents by serving in an authoritarian power structure, some groups may, for reasons similar to those of the martial races, have a comparative advantage in the production of violence, and this may account for why they are chosen for an authoritarian military or security forces. Some of these groups may also not appear to be those with the most obvious ethnic, religious or ideological affinity towards the regime (see for instance the importance of the Chechen Kadyrovtsy in the modern Russian military). Indeed, this dynamic may also extend beyond autocracies. Many modern day democratic militaries recruit disproportionately from certain segments of the population, including the over representation of white Southerners in the US military or, historically, the over representation of aristocrats and especially Anglo-Irish aristocrats in the British officer corps.

Future research could look to extend this model to recruitment for other colonial or imperial forces, such as those of the British in other parts of the Empire, or other colonial powers such as France or the Netherlands. More broadly, future research could also examine whether exclusionary recruitment policies in other types of organization, which would appear to be damaging to principals by excluding potential talent from consideration, may have a more functional explanation.

Intuitively, one would suppose that a meritocratic policy which seeks agents from all portions of society equally would be most conducive to successful organizational performance. Indeed, substantial research argues this to be the case (Lyall, 2020). However, this study suggests that this view is mistaken in some respects since it ignores the transaction costs of identifying motivated agents and aligning their incentives with those of the army, which may be differentially distributed across different social groups. Under certain circumstances, therefore, a discriminatory policy may in fact be optimal from the point of view of the principal. A key scope condition may be transaction and monitoring costs (Allen, 2011) - discriminatory practices such as the martial race recruitment policy may be less optimal in modern contexts where the ability to identify and monitor talented agents is greater.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Charles Miller

Notes

References

Aimone

Iannaccone

Makowsky

, et al. (2013) Endogenous group formation via unproductive costs. Review of economic studies 80(4): 1215–1236.

Aksoy

Gambetta

(2016) Behind the veil: the strategic use of religious garb. European Sociological Review 32(6): 792–806.

Allen

(2011) The Institutional Revolution: Measurement and the Economic Emergence of the Modern World. University of Chicago Press.

Barber

Miller

(2019) Propaganda and combat motivation: radio broadcasts and German soldiers’ performance in World War II. World Politics 71(3): 457–502.

Berman

(2000) Sect, subsidy, and sacrifice: an economist’s view of ultra-orthodox Jews. The Quarterly Journal of Economics 115(3): 905–953.

Berman

(2011) Radical, Religious, and Violent: The New Economics of Terrorism. MIT press.

Berman

Laitin

(2008) Religion, terrorism and public goods: testing the club model. Journal of public Economics 92(10-11): 1942–1967.

Bingley

(1899) Handbook on Rajputs. Asian Educational Services.

Brennan

Tullock

(1982) An economic theory of military tactics: methodological individualism at war. Journal of Economic Behavior & Organization 3(2-3): 225–242.

10.

Carvalho

(2013) Veiling. The Quarterly Journal of Economics 128(1): 337–370.

11.

Carvalho

(2016a) Identity-based organizations. American Economic Review 106(5): 410–414.

12.

Carvalho

(2016b) Sacrifice and sorting in clubs. In: Forum for Social Economics. Taylor & Francis.

13.

Carvalho

(2019) Religious clubs: the strategic role of religious identity. In: Advances in the Economics of Religion, 25–41.

14.

Carvalho

Koyama

(2016) Jewish emancipation and schism: economic development and religious change. Journal of Comparative Economics 44(3): 562–584.

15.

Carvalho

Koyama

Sacks

(2017) Education, identity, and community: lessons from Jewish emancipation. Public Choice 171: 119–143.

16.

Carvalho

Iyer

Rubin

(2019) Advances in the Economics of Religion. Springer.

17.

Chaudhary

Rubin

(2016) Religious identity and the provision of public goods: evidence from the Indian princely states. Journal of Comparative Economics 44(3): 461–483.

18.

Chen

(2017) The Deterrent Effect of the Death Penalty? Evidence from British Commutations During World War I (Feb 26, 2017). Available at SSRN: https://ssrn.com/abstract=2816255

19.

Chen

McBride

Short

(2019) Dynamics of religious group growth and survival. Journal for the Scientific Study of Religion 58(1): 67–92.

20.

Costa

Kahn

(2003) Cowards and heroes: group loyalty in the American civil war. The Quarterly Journal of Economics 118(2): 519–548.

21.

Costa

Kahn

(2009) Heroes and Cowards: The Social Face of War. Princeton University Press.

22.

Dhavan

(2011) When Sparrows Became Hawks: The Making of the Sikh Warrior Tradition. USA: OUP, 1699–1799.

23.

Farooqui

(2015) “divide and rule’? race, military recruitment and society in late nineteenth century colonial India. Social Scientist 43(3/4): 49–59.

24.

Fawcett

(1901) Nayars of Malabar. Madras: Government Press.

25.

Feaver

(2003) Armed Servants: Agency, Oversight, and Civil-Military Relations. Harvard University Press.

26.

Hodgson

CJS

(1850) Opinions on the Indian Army. Government Press Meerut.

27.

Hutton

(1963) Caste in India: Its Nature, Function and Origins. Oxford University Press.

28.

Iannaccone

(1992) Sacrifice and stigma: reducing free-riding in cults, communes, and other collectives. Journal of political economy 100(2): 271–291.

29.

Imy

(2019) Faithful Fighters: Identity and Power in the British Indian Army. Stanford University Press.

30.

Iyer

(2016) The new economics of religion. Journal of Economic Literature 54(2): 395–441.

31.

Iyigun

(2019) War, Peace, and Prosperity in the Name of God: The Ottoman Role in Europe’s Socioeconomic Evolution. University of Chicago Press.

32.

Johnson

Koyama

(2019) Persecution & Toleration: The Long Road to Religious Freedom. Cambridge University Press.

33.

Kolff

(1990) Naukar, Rajput & Sepoy: The Ethnohistory of the Military Labour Market in Hindustan 1450-1850. Cambridge University Press.

34.

Kuran

(2012) The Long Divergence: How Islamic Law Held Back the Middle East. Princeton University Press.

35.

Leeson

(2007) An-arrgh-chy: the law and economics of pirate organization. Journal of political economy 115(6): 1049–1094.

36.

Lyall

(2020) Divided Armies: Inequality and Battlefield Performance in Modern War. Princeton University Press.

37.

MacMunn

(1911) Armies of India. A & C Black.

38.

McBride

(2015) Why churches need free-riders: religious capital formation and religious group survival. Journal of Behavioral and Experimental Economics 58: 77–87.

39.

McBride

(2016) A rational choice theory of religious authority. Rationality and Society 28(4): 410–438.

40.

Miller

Barber

Bakar

(2018) Indoctrination and coercion in agent motivation: evidence from Nazi Germany. Rationality and Society 30(2): 189–219.

41.

Morrissey

(2017) A lost heritage: the Connaught Rangers and multivocal Irishness. In: Ireland’s Heritages. Routledge, 71–87.

42.

Omissi

(2016) The Sepoy and the Raj: The Indian Army, 1860-1940. Springer.

43.

Peers

(1995) Sepoys, soldiers and the lash: race, caste and army discipline in India, 1820–50. The Journal of Imperial and Commonwealth History 23(2): 211–247.

44.

Platteau

(2017) Islam Instrumentalized. Cambridge University Press.

45.

Ridgway

(1910) Handbooks for the Indian Army: Pathans. Calcutta.

46.

Roy

(2008) Brown Warriors of the Raj: Recruitment and the Mechanics of Command in the Sepoy Army 1859-1913. Manohar.

47.

Streets

(2017) Martial Races: The Military, Race and Masculinity in British Imperial Culture, 1857–1914. Manchester University Press.

48.

Thurston

(1909) Castes and Tribes of Southern India. Madras: Government Press.

49.

Tyagi

(2009) Martial Races of Undivided India. Gyan Publishing House.

50.

Vansittart

(1906) Handbooks for the Indian Army: Gurkhas. Imperial Government of India.