Military Fiscalism and the Politics of Market Reform at a Time of Civil War

Rajesh Venugopal, ** courtesy of the author and EPW…Note that this article appeared in the Economic and Political Weekly (EPW) vol 49, issue 49 (December 2011)

1.   The Problem: By the late-1990s, almost two decades of civil war in Sri Lanka had wrought a heavy economic cost: the physical destruction of economic infrastructure, lost production, foregone investment, the flight of human capital, and the diversion of vast resources to military purposes were quantified as over a full year’s worth of lost GDP (Arunatilake et al 2001).  Other studies quantified the cost of the war as a loss of between 2-3 percent of economic growth per year, implying that ceteris paribus, Sri Lanka’s 2002 GDP of $900 per capita was half of what it would have been if there had been no war (Central Bank of Sri Lanka, henceforth CBSL, 1998, quoted in World Bank 2004: 10). The north-eastern part of the island in particular suffered to a very disproportionate extent during these years, and came to have the lowest income levels, the highest poverty levels, and the worst provision of health and education in the island (Sarvananthan 2008). Indeed, due to the exclusion of the north-east from national accounts and most census statistics since 1990, much of the cost of the war is not incorporated into published GDP and other socio-economic data.

Nevertheless, despite the extent of economic destruction and missed opportunities, what is striking about Sri Lanka, particularly in comparison to other such countries, is the extent to which economic normality prevailed through the war years in most of the country.  A comparative study of the economic and social consequences of civil war in seven countries finds that Sri Lanka is unique not just in having experienced economic growth amid war, but in economic growth rates that exceeded the pre-war period (Fitzgerald et al 2001, O’Sullivan 2001).  Compared to other conflict-ridden countries where war has caused the substantial destruction of the formal economy and resulted in negative growth rates, Sri Lanka stands out as a curious exception (Stewart and Fitzgerald 2001).

Indeed, the war years, and particularly the decade of the 1990s, were paradoxically a time of strong economic growth, continuing market reforms, and substantial structural transformation of the economy.  During the first 15 years of war between 1983-98, real GDP growth averaged 4.6 percent annually while exports multiplied almost three-fold in real terms (World Bank 2001). Throughout the war period, governments from different political parties continued to implement market reforms while promoting a successful export-driven industrialisation process.  During the 1980s and 1990s, Sri Lanka developed a two billion dollar garment export industry virtually from scratch, ending a century of reliance on the tea crop (UNCTAD 2000). Considering that this growth occurred and was partially offset by war-related destruction (estimated at 2 percent-3 percent each year), it would appear that the underlying momentum of economic growth during the war period was actually quite spectacular.  Similarly, the ‘open economy’ policy of market reforms which began in 1977, and the privatization programme which began in 1989, also continued to unfold and expand through the war, proving resilient and even gaining in momentum through periods of economic, political and military crisis.

What accounts for this curious paradox of economic growth and policy reform amidst the destruction of civil war?  Compared to other countries, where the disruptions of war have caused major delays, reversals, and even the complete abandonment of the market reform process, Sri Lanka appears unique – even exceptional.  Indeed, within the positivist naïveté of mainstream development economics, where the quest for the sources of economic growth remains paramount, the Sri Lankan experience might even be seen as a triumph – as a useful example from which policy experiences of relevance to other conflict-ridden countries might be extracted on how economic growth and reform can be achieved despite war.  Instead in this essay, I take the paradox of growth/reform amidst war as the point of departure to formulate a different set of research questions on the relationship between development and conflict.   Importantly, this does not start off with the casual assumption that growth and reform took place despite war, but instead explores a variety of alternate mechanisms by which the two processes might have related to each other.

The simplest and most intuitive explanation for this paradox of economic growth amidst war in Sri Lanka during the 1980s-90s is in terms of the physical separation of the dynamics of the two processes.  In geographical terms, the locus of destruction in the north-east was clearly segregated from the locus of economic growth in the south-west.  Although the south has been quite seriously affected by the war through suicide bomb explosions, the constant stream of casualties from the front, and the militarization of daily life, (De Mel 2007) it bears little comparison to the scale, intensity and duration of human suffering and economic dislocation in the north-east.  The war compounded the pre-existing economic marginality and backwardness of this area, which had long before been neglected, remote, and impoverished; characterised by an arid, austere agricultural economy.  Indeed, the difference between the two parts of Sri Lanka has for long been so stark that a journey from the rest of ‘normal’ Sri Lanka past Anuradhapura to the north or Polonnaruwa to the east not only gave the appearance of entering an entirely different country with a different language, culture, topography and ecology, but of embarking on time-travel a few decades back to an era largely innocent of cars, telephones, paved roads or electrical appliances.

As a result of the fact that the war was sealed off within the bounds of the economically peripheral north-east at a time of rapid economic growth in the south, the north-east became not just more impoverished in absolute and relative terms, but ever more economically disconnected and irrelevant to the island’s economic development as a whole.  Consequently, the pre-existing physical, linguistic and psychological differences between the two parts was magnified by a growing economic rift that contributed to the separation and mutual incomprehensibility of the populations of the north-east and south.

There is as such a fairly straight-forward and intuitive answer to the paradox.  Sri Lanka experienced economic growth amidst war because the war was confined to the economically peripheral north-east, while the more prosperous south-west, with the tea plantations, export-processing zones and tourist resorts remained remarkably well insulated from its direct effects, and enjoyed a large measure of political and economic normality.

There is, however, much that remains unexplained and under-explained by viewing this problem in terms of the segregation of these two processes.  In reality, development and destruction were never entirely separated, but have interacted and even found sustenance in each other in a variety of ways.  For example, the economic growth and stability in the south has arguably been a factor in the continuation of the war in the north, as it has provided the government with a steady and growing fiscal base with which to fund its rapidly escalating military budget.  In addition, the comparative normality of civilian life in the south rendered the war a distant reality to the large majority of the population, who were as a result, less impressed with a sense of urgency in bringing about its end.

Moreover, the much vaunted economic prosperity of the south has in reality been very unbalanced such that large parts of the rural hinterland further away from Colombo have little to show for it, and are only marginally better off than the war-torn districts of the north-east that they border.  For the most part, it is the western province that has enjoyed the benefits of growth, expanding its share of national GDP from 40.2% to 49.4% between 1990-2000 (CBSL, various). The very skewed and unequal nature of the prosperity generated in the 1990s is also evident in terms of the negligible impact it has had on the poverty headcount.  Partly as a consequence of this imbalance, large parts of the south have not been insulated from the war at all, but have instead found themselves increasingly drawn into it and have even become dependent upon it in a variety of ways.

The ‘separate spheres’ paradigm explains the sources of Sri Lanka’s war-time growth by viewing the north and the south as disconnected spaces, and by categorising growth and conflict as analytically discrete, distinct phenomena belonging to separate, self-contained domains of policy formulation and implementation.  I contend that there are serious shortcomings with this paradigm, and suggest instead that civil war in the north and market reform driven economic growth in the south were not quite as separate as they might seem.  This essay argues that the civil war has actually been of functional significance for the promotion of economic growth through the fiscal multiplier effects of the growing military budget.

In a nutshell, this essay sets out an argument that the civil war has been of functional significance to the promotion of economic growth by mitigating the adverse social effects of the market reform agenda in the south.   In effect, the escalating military budget compensated for the contraction of the state due to market liberalisation, and thus made the reform agenda politically viable.  By the 1990s, the army had become the single largest employer in the country, offsetting the reduction in civilian state employment under the reforms, and containing the growing inequalities of this period.  I suggest that this had the largely unintended and unforeseen consequence of facilitating the market reform agenda by helping to win the passive quiescence, if not the active consent to an economic reform agenda that was broadly opposed by large segments of the population.

Before proceeding, there are several important qualifications and issues of relevance to note.  Firstly, the relationship between growth/reforms and war outlined here is a functional, not a causal relationship.  Although the ongoing civil war has provided sources of political, socio-economic, ideological support for the economic reform agenda, this relationship holds only in the sense of a static equilibrium, and it does not explain the evolution of the conflict or how the reform agenda came into being.  The spirit of the functional relationship between conflict and reform outlined here draws upon the similar way that James Ferguson’s ethnography of foreign aid in Lesotho frames the impact of development on the bureaucratisation of state power (Ferguson 1994).  Ferguson in turn, draws inspiration from a Foucauldian functionalism, which searches for instrumentality behind the veil of disorder and dysfunction to ask what agendas are served and upheld by the persistent condition of failure (Foucault 1979).  In much the same spirit, I seek to go beyond the rhetoric of failure and dysfunctionality that pervades the literature on the development implications of the conflict in Sri Lanka.  In contrast to the mainstream literature that views the persistent conflict as a source of development failure, I ask instead what functions did the war serve?  What interests does it serve to uphold?  What kinds of development has it given rise to? [1]

In framing the relationship in this manner, there is an important qualification to bear in mind: by looking for instrumentality within disorder, there is no suggestion here that disorder is the handiwork of those that benefit from it.  As James Ferguson describes:

If unintended effects of a project end up having political uses, even seeming to be ‘instruments’ of some larger political deployment, this is not any kind of conspiracy; it really does just happen to be the way things work out. (Ferguson 1994: 255-256)

Similarly, as I proceed to explain in the course of this essay, the functional role that the conflict played in promoting the reform agenda was not the result of a conscious conspiracy hatched by far-sighted manipulative elites.  Sri Lanka’s capitalist class and pro-reform elites did not dream up the ethnic conflict in order to smuggle in their class agenda under a cloud of false consciousness.  Indeed, Sri Lanka’s market reformist elites were arguably the segment of society who were least invested in the project of Sinhala nationalism, and least interested in the continuation of the civil war.[2]

Finally, in looking at the Sri Lankan civil war through the lens of the politics of market reform, there is also another motivation and underlying theme that is sought to be highlighted.  Due to the phenomenon of ‘ethnic over-determination’, and the relentless intrusion of the conflict into every sphere of life, the literature on Sri Lanka’s ethnic conflict often suffers from an analytical solipsism; i.e. takes place within the discursive categories and frames of reference that are self-generated by the conflict itself.  As Jonathan Spencer describes: ‘Since the mid-1980s, almost all attention has focused on the ethnic crisis and the escalating civil war, and somehow the issue of the distribution of power and resources within the polity have lost their academic urgency’ (Spencer 2002: 93).

Indeed, the dynamics of the conflict are implicitly presumed to take place within the confines of a universe where the only actors and identities available are the ethnicities in question themselves, who enter the analysis as fully-formed, timeless, monolithic, pervasive, inescapable, and pre-disposed to mutual confrontation.  What follows is that the writing on the ethnic conflict, (and not just the more spirited and partisan exchanges in the media), often suffers from what Zygmunt Bauman describes of the holocaust literature, where ‘discussion of guilt masquerades as the analysis of causes’ (Bauman 1989: xi).

I hold instead that it is important to de-link the study of the conflict from itself, as it were, and to situate it more within the larger field of politics, policy-making, and political contention, where it shares and jostles for space as just one among a series of other issues, where the important decisions on the conflict are made in a complicated environment where it is balanced and contingent upon numerous other pressing political agenda items and dilemmas.

2. The Politics of Market Reform: The problematique of the politics of market reform in Sri Lanka can for most purposes, be reduced to a simple question: how do democratically elected governments go about implementing unpopular economic policies that are certain to damage their electoral prospects?  Sri Lanka’s market reform agenda suffers from inherent unpopularity and lack of popular legitimacy that makes it an electoral handicap for any party that wishes to adopt it.  Yet, the prerogatives of a perpetually insolvent treasury, combined with pressure from foreign donors, and the powerful lobbying of domestic and foreign business groups has forced every government since 1977 to pursue this unpopular agenda.  How, one must ask, did they pull it off?

There is, and has been in Sri Lanka and in other countries, the perpetual anxiety that the contradictions of this process would become untenable – that either democracy or reforms would give way, leading to free-market dictatorships or dirigiste democracies.  And indeed, if one looks at the comparative experiences of the politics of market reform literature, one comes across very similar situations where this contradiction led to great stress.

Without digressing into an expansive review of the comparative literature, liberal democracies were typically preserved through market reforms only with great difficulty and ingenuity. As Rob Jenkins describes it: ‘Pushing through reform measures requires a broad range of underhanded tactics’ (Jenkins 1999: 206). The empirical literature describes how successful reforms were implemented through the establishment of authoritarianism or political hegemony – either in terms of political dictatorships and the sheer absence of democratic mechanisms, or more often through the domination of extant democratic mechanisms by a pro-reform party or coalition, and the insulation of the reform agenda from the pressures of electoral politics (See for example Herbst (1993), Nelson (1990), Haggard and Kaufman (1992), Przeworski (1991).  In other cases, reforms involved the containment of sources of opposition through coercion, cooption or corruption, but more often through legislative, political or administrative measures, such as restrictions on trade unions.  Yet another set of strategies involved the promotion of alternative forms of legitimacy and other sources of political division to counter the centrality of the reforms agenda, and to neutralise the legitimacy-deficit that they created.

Sri Lanka’s experience with market reforms have two important specificities to bear in mind.  Firstly, the economic reform agenda has, since its introduction in 1977, proceeded apace in parallel with the prosecution of a worsening ethnic conflict, and its escalation into a civil war in 1983.  Sri Lanka’s switch from state to market in 1977 occurred significantly ahead of the wave of market reform that swept Latin America and Africa in the 1980s, and more than a decade before India.  Market reforms and the ethnic conflict have, since 1977, been the two dominant issues on the Sri Lankan political agenda, but apart from a spirited, but short-lived burst of writing in the late-1980s that asserted a causal link between them, they are for the most part not related analytically in the study of either one. (Gunasinghe (1984), Moore (1990), Dunham and Jayasuriya (2000).

Secondly, the market reform agenda in Sri Lanka has had to contend with the fact that it is widely viewed as illegitimate within the dominant framework of political morality that regulates state-society relations.  The development of electoral politics in Sri Lanka from the 1930s-70s occurred parallel to and in close connection with the construction of a massive social democratic-welfare state.  Political parties competed and came to power on populist electoral platforms to expand the role of the state; and in doing so, helped enshrine certain moral notions of the legitimate and necessary role of the state.  The provision of public services such as free education and health, the provision of public employment, the protection of peasant agriculture and rural life against the pressures of internal capitalist expansion and international market pressures, the alleviation of poverty and social inequalities, together with the promotion of the Sinhala language and patronage of the Buddhist religion had over this period come to comprise the elements of the moral economy within which the very legitimacy and stability not just of any single government, but of the state itself and the political system was hinged.

This universe of political morality is in essence constitutive of Sinhala nationalism, which can be understood for these purposes as the ideology of an ethnicised social democratic state in which the state is seen as having deep moral obligations towards society in general, and towards the material and spiritual needs of poorer, rural Sinhala-Buddhists in particular.  As a result, the politics of market reform in Sri Lanka is not just about the implementation of an unpopular agenda, but about one that runs counter to deeply ingrained notions of what the legitimate and just role of the state should be.  How, it must be asked, did the government, the state, and the institutions of electoral politics survive such a direct assault upon the system of public legitimacy which had sustained it thus far?

I describe elsewhere how the prerogative of winning consent for the market reform agenda contributed indirectly towards the catalysis of the civil war (Venugopal 2010).  In seeking to compensate for the sources of their electoral failures in the past, the United National Party (UNP) governments of J.R. Jayawardene and Ranasinghe Premadasa in the post-1977 period sought to pre-empt the opposition by seizing the mantle of both Sinhala nationalism and social justice.  As a result, the prosecution of the reform agenda effectively created the conditions that shrivelled the political possibilities for compromise on the ethnic conflict.  The underlying issue that this pointed towards was that market reform and the political accommodation of minority rights were both controversial and to some extent mutually complementary projects of state reform.

Many of the underlying problems of the ethnic conflict were caused in terms of the way in which the social democratic state had been constructed under the development enterprise in the 1940s-70s in response to the pressures of Sinhala nationalism, demographic change, and the prerogatives of protecting and preserving the Sinhalese peasantry. As a consequence, I described how the politics of market reform revolved around two mechanisms – ideological and material.  In ideological terms, the market reforming UNP affected an exaggerated performance of Sinhala-Buddhist authenticity.  In material terms, they compensated for the withdrawal of the state and the marketisation of some spheres of activity by the aid-funded expansion of the state in other spheres, including massive rural development projects.

This essay extends that analysis of the 1980s through to the Kumaratunga years of the late-1990s to describe how the state unwittingly expanded its presence in one sphere (military expenditure) while simultaneously withdrawing significantly in other areas through privatization.  I suggest that this had the unwitting effect of diffusing and moderating the social impact of the reforms, particularly in providing a vast source of state employment opportunities to young Sinhalese men of poor, rural backgrounds.

3. Military Fiscalism: The term ‘military fiscalism’ features widely in the historical literature on state formation in early modern Europe, relating the growth of formal state institutions and domestic capital accumulation to the compulsion of funding increasingly expensive militaries. As Charles Tilly describes, ‘war made the state, and the state made war’ (Tilly 1975: 42).  Military fiscalism also features in the history of state formation in South Asia, for example in the work of Chris Bayly (1988), Burton Stein (1985), David Washbrook (1988), and Douglas Peers (2007).  As with the European experience, they also describe how military imperatives forged the innovation of centralised bureaucracies dedicated to systematised revenue collection.  In this essay, I use ‘military fiscalism’ to describe a different sequence of relationships, but one that nevertheless relates the relationship between military spending, the polity, and the economy.

In the early 1970s, a controversial study by the economist Emile Benoit made the unusual and counter-intuitive finding that there was a positive, causal correlation between military expenditures and economic growth in developing countries.[3]  Military spending, Benoit concluded, had a Keynesian fiscal effect on aggregate demand, generating positive multiplier effects.  It created beneficial externalities for the civilian economy by: feeding, clothing, and housing a number of people who would otherwise have to be fed, housed, and clothed by the civilian economy-and sometimes doing so, especially in LDCs, in ways that involve sharply raising their nutritional and other consumption standards and expectations.

Benoit’s findings provoked a series of critical rejoinders in the coming years, in which the empirical validity of the results and its theoretical foundations were subject to vigorous challenge (Ball 1983, Faini et al 1984, Deger 1986).  At about the same time, Paul Sweezy and Paul Baran’s classic 1966 work on monopoly capital also argued for a positive correlation between military spending and economic growth, albeit this argument applied to advanced capitalist countries. One of the key links posited by Baran and Sweezy is that millions of jobs are generated by military expenditure, and that this in turn, absorbs the reserve army of labour: Some six or seven million workers, more than 9 per cent of the labor force, are now dependent for jobs on the arms budget. If military spending were reduced once again to pre-Second World War proportions, the nation’s economy would return to a state of profound depression, characterized by unemployment rates of 15 per cent and up, such as prevailed during the 1930’s. (Baran and Sweezy 1966: 531).

This essay deploys a modified version of military fiscalism developed in the work of Benoit and Baran and Sweezy to the context of the politics of market reform in Sri Lanka.  I argue that the high levels of military expenditure associated with a prolonged civil war had an important positive impact on socio-political stability.  At a time of growing inequalities and social tensions generated during the market reform period, military employment created the stabilising socio-political conditions within which economic growth was preserved and a controversial programme of privatisation and market reforms could be implemented.

Before proceeding to examine exactly what this signified, it is important to recapitulate the historical role of the public sector and welfare spending in its period of expansion in Sri Lanka.  The opening of free secondary education in 1944 and the expansion and vernacularization of the social democratic state in the 1950s and 1960s had in this period, made it possible for tens of thousands of relatively poor families to escape rural adversity and join the urban middle class.  It was a time of unprecedented upward social mobility, although it is probably true that the new opportunities for social advancement and inter-generational class mobility were disproportionately exploited by relatively wealthier segments of rural society.  What occurred in those decades was more of a process of transition from the middle to upper ranks of the peasantry to the lower-middle ranks of the bureaucracy, or what I have described as ‘kulaks to clerks’ (Venugopal 2008).

As the state grew rapidly in size, and became more Sinhala-dominated from the mid-1950s onwards, the composition of white collar segments of the government service down to and including clerks and typists were gradually transformed from being dominated by Jaffna Tamils, Sinhalese Christians, and Eurasians of urban, middle class, English-educated backgrounds, to vernacular-educated Sinhala Buddhists of rural backgrounds, in many cases, literally straight from the villages.  This process occurred through the 1950s – 1980s, providing a significant portion of the rural population with a viable and realistic opportunity for upward mobility.

This dynamic had, however, begun to change quite significantly by the 1990s.  Under the impact of the privatisation agenda which began in 1989, and also because of the extended atrophy of the state, the possibilities of escaping rural poverty through education and state employment became far more meagre than they were for the previous two generations (see for example, Hettige 2000). This was also accompanied by a broad shift in this period from the public to the private sector as the source of available employment opportunities.  The 1990s were a decade when the public sector share of total employment dropped from 21.5% to 13.4 percent, most of which was picked up by the expanding private sector (Sri Lanka Labour Force Survey Annual Report, various)

The diminished opportunities for upward mobility that accompanied the shift from public to private employment are reflected in a number of different indicators, such as income inequality.  As Osmani describes, ‘there is clear evidence to suggest that post-reform growth has been of an exceedingly unequalizing kind’ (Osmani 1994: 294).  Figure 1 shows how the income shares of the top 10% versus the bottom 50% declined steadily under welfarism from the 1950s to the mid-1970s.  But following the introduction of the reforms in the late-1970s, inequality levels increased very sharply, effectively undoing in one decade what had been achieved under welfarism over the previous three decades.[4]

Subsequent studies have shown that inequality continued to grow in the 1990s, fuelled by high rates of economic growth that were matched by very low rates in poverty reduction.  The World Bank’s 2004 development policy review of Sri Lanka describes the following situation:

Of particular concern is the fact that poverty reduction has been slow while income inequality has risen in recent years. At 22.7 percent, the national poverty headcount ratio remains high for a country with US$ 900 per capita GDP. Furthermore, the rate of decline in this ratio has been modest despite sustained per capita annual GDP growth of over 3 percent per year over the last two decades. (World Bank 2004: i)

FIGURE 1: PERCENTAGE OF TOTAL INCOME BY RANKED SPENDING UNIT

Data drawn from World Bank (1995), table 1.1. (Percentage of total one month pre-tax income received by each decile of ranked spending units.)

Similarly, the Government of Sri Lanka’s Poverty Reduction Strategy Paper (PRSP) from 2002 describes: … a modest growth rate has been accompanied by little or no income redistribution.  In other words, the benefits of economic growth have not automatically trickled down to the poor. (Government of Sri Lanka 2003, Part 2: 28)

Gunatilaka and Chotikapanich (2006) use household data to explain that there was a steady increase in inequality (through the Gini coefficient) in Sri Lanka between 1980-2002 – but with a particularly sharp increase during the 1990-2002 period.  Narayan and Yoshida (2005: table 6) similarly find that in these years, the mean real per capita consumption of the top quintile of the population increased by 50.4%, while that of the bottom quintile increased by just 2.2%. Per capita income increased in this period by a total of 45%, but it had a marginal impact on the poverty headcount, which went down just 3.4 points from 26.1% to 22.7%.  This too, was overwhelmingly concentrated in urban districts such that the poverty headcount was either the same or had increased in 9 of 17 districts during the 1990s (excluding the north-east).[5]

The increase in economic inequality manifest in these data can be disaggregated further into regional and sectoral dynamics.  For example, a very disproportionate share of the growth in this period was regionally concentrated in the Western region around greater Colombo, which between 1990-2000, increased its share of national GDP from 40.2% to 49.4% (CBSL various).  There was also a significant sectoral imbalance in the growth, which came largely from the industrial and service sectors.  In contrast, there is evidence of an unusually rapid decline in the agricultural economy, both in relative and absolute terms.  In just the decade from 1990-2001, the percentage of the workforce employed in agriculture had declined from 47% to 33%, while the contribution of agriculture to national GDP dropped from 26% to 20% (CBSL various).  This extent of decline in the agricultural economy reflected a return to the burgeoning crisis of the peasant sector that was first observed in the 1930s, and which the vigorus state-led developmentalist and welfarist policies of the 1940s-80s had partially reversed.

With the global decline in agricultural commodity prices over the 1980s and 1990s, peasant agriculture and particularly paddy farming had, according to the cost of cultivation statistics, been producing negative returns, even on a pure cash basis (Government of Sri Lanka Department of Agriculture, 1990-2000)   That is, many farmers, especially paddy farmers, were barely recouping their cash costs of production in terms of seed, fertilizer, insecticide, and hired labour, even without incorporating the imputed value and opportunity cost of unpaid family labour.

The economic growth of the 1980s and 1990s created a significant expansion in private sector employment, particularly in areas such as construction, garment factories, and tourist resorts – and there was a historic reduction in the unemployment rate during the 1990s.  But upon further scrutiny, the actual kinds of private sector employment created fall mostly into three distinct categories: (i) overwhelmingly female employment in garment factories; (ii) the expansion of informal sector employment, particularly in small, service-sector enterprises that employed casual labour, i.e. the lowest paid, lowest status, most manually demanding jobs with the least security; (iii) white collar sales and clerical jobs in the corporate sector – which are typically dominated by people from urban, English-speaking, middle and lower middle class origins.

What stands out about these categories of private sector employment is they offer very little for educated rural male Sinhalese youth; a segment of society that has historically depended on free public education and state employment for social mobility.  This is also the segment of society that has formed the social base of two episodes of armed insurgency against the state in the 1970s and 1980s.

FIGURE 2. SRI LANKA, NUMERICAL STRENGTH OF SECURITY FORCES, 1982-2002

Source: IISS, The Military Balance (various).

It is within these circumstances of greater economic inequality, regional concentration, an expanding rural-urban gap, and an almost precipitous decline of peasant agriculture that military employment came to occupy an important role in the Sinhalese village.  Over the course of the 1980s and 1990s, Sri Lanka’s military grew ten-fold, from 15,000 in 1982 (the year before the war started) to 150,000 by the time of the cease-fire of 2002.  Including the police and paramilitaries, the size of the security sector increased from a total of 30,000 in 1982 to 250,000 by 2002.

TABLE 1. MILITARY AS A   PERCENTAGE OF TOTAL GOVERNMENT SALARIES AND WAGES
1979-82	1983-86	1987-90	1991-94	1995-98	1999-01
3%	4%	14%	32%	39%	41%
Source: Calculations   based on data from CBSL Annual Report (various).

Given that public sector employment has shrunk as a result of privatisation, the security sector had by 1997 come to comprise one in five of all government jobs, with the Sri Lankan army becoming the country’s largest employer.  In fiscal terms, the salaries of the armed forces by the late-1990s amounted to over 40% of the government’s total wage-bill.

Despite the burgeoning role of military employment in Sri Lanka, and the widespread anecdotal evidence of its increasing significance in the Sinhalese rural economy there is, (perhaps deliberately for security reasons), virtually no data published on the extent of military employment, and consequently very little analytical or policy discussion of its repercussions.  For example, the Census of Public Sector Employees excludes the military altogether from their data; the Quarterly Labour Force data does not separate out military employment as a category of its own.  Data on military employment is however collected at quite a detailed level in several household survey data sets, and I make use of a large, nationally representative household survey data set to outline some demographic and social characteristics of military recruits.

 4. Sri Lanka Integrated Survey (SLIS) 1999/2000: The Sri Lanka Integrated Survey (SLIS) is a household survey data-set commissioned by the World Bank between October 1999 – third quarter of 2000 and executed by the Sri Lanka Business Development Centre.  In all, a total of 7,500 households were surveyed in 500 distinct communities selected across the country, with the exception of LTTE controlled ‘uncleared areas’ in the North East – largely in the districts of Killinochi, Mullaitivu, Mannar and Vavuniya.  There is some potential for the results extracted below to be biased due to these omissions, but this is in reality not entirely relevant because these are overwhelmingly Tamil areas, and this essay is largely interested in employment patterns and alternatives among young Sinhalese.

TABLE 2 ETHNIC COMPOSITION OF   OCCUPATIONAL CATEGORIES FOR MALES 18-30
							Total
	Public	Military	Priv Sal	Casual	Business	Farming	Employed	Count*
Sinhalese	63.2%	97.1%	61.2%	60.5%	63.9%	75.5%	65.5%	1,796
SL Tamil	20.3%	1.4%	23.0%	25.6%	13.0%	21.3%	21.6%	593
Ind Tamil	0.5%	0.0%	2.8%	3.0%	1.4%	1.6%	2.2%	59
Muslim	15.6%	1.4%	12.4%	10.1%	18.5%	2.1%	9.9%	270
Others	0.5%	0.0%	0.7%	0.8%	3.2%	0.0%	0.9%	26
Total	100%	100%	100%	100%	100%	100%	100%	2,744
Count*	212	210	575	1,020	216	440	2,744
Source: SLIS 1999/2000

The SLIS data demonstrates that the structure of employment varies significantly by age, ethnicity, region and education.  Indeed, the data reveals rather unsurprisingly that military employment is concentrated amongst young Sinhala-Buddhist males aged 18-30.  Furthermore, it is very disproportionately drawn from the overwhelmingly rural areas of the outer Sinhalese periphery distant from the capital Colombo.  That is, military employment is heaviest in areas such as Polonnaruwa, Anuradhapura, Kurunegala, Trincomalee and Ampara[RV1] .  In terms of educational background, military employment is concentrated heavily on those who have completed secondary school, with 10-11 years of education, a category that accounts for almost half the total 18-25 age group. There is also data to suggest that military recruitment is also higher in the more predominantly agricultural districts, as there is also a close correlation between districts with a high degree of military employment among the male, 18-30 age group, and those where the older male generation aged 45-65 include a high proportion of farmers.

On a national basis, the military only accounted for around 3% of all employment, but this figure increases substantially for younger Sinhala-Buddhist males.  For this group, military employment varied from a low of 7% of employment in the Western province (even here it was concentrated in the more rural Gampaha district and was negligible in urban Colombo) to a high of 23% in the Eastern province, where most of the Sinhalese population are second or third generation settlers in irrigation-based colonies such as Gal Oya (Ampara) or Kantale (Trincomalee).  In terms of education, military employment is concentrated among those that have completed secondary school, but have no further education – a large category that comprises almost 50% of the total 18-25 age group.  Among young Sinhalese males of this group, which has high rates of unemployment, the military accounts for 17% of all employment, compared to only 5% in the civilian public sector.

Indeed, the contrast between employment in the military versus the civilian public sector is very stark for this demographic sub-category.  In absolute terms, civilian public sector jobs far outnumber the military: there were by the late 1990s, about 900,000 civilians in the public sector, and about 250,000 in the security forces.  But the data suggests that there is a very strong age barrier to civilian public sector jobs; for the 18-25 age group, the extent of civilian public service employment is quite negligible at between 4%-6%. The real alternatives to a life in the army are in casual labour, or for those living nearer the urban areas, private salaried jobs.

FIGURE 3. SOURCES OF CASH EMPLOYMENT FOR SINHALA BUDDHIST MEN, 18-30 WITH 10-11 YEARS EDUCATION, (Ampara, Trincomalee, Polonnaruwa, Anuradhapura and Moneragala, 114 obs.)

Source: SLIS 1999/2000.

Thus, within the multiple dimensions of segmentation in Sri Lanka’s labour market, the military has come to dominate employment in a particular demographic category defined by age (18-30), gender (male), ethnicity (Sinhalese), region (the rural periphery) and education (secondary school).  For this segment of the population (taking the districts of Ampara, Trincomalee, Polonnaruwa, Anuradhapura, and Moneragala), military employment is a necessity due on the one hand to the ‘push factor’ from shrinking agricultural livelihoods, and on the other to the absence of civilian alternatives in the private or public sector.  Some 31% of the young men in this demographic category are unemployed, a figure that would rise to almost 43% if not for the abundant presence of military employment due to the war.

The only alternative to unemployment, or the armed forces is the life of a casual labourer, a fate which most would seek to avoid to the extent possible, as it almost certainly implies a life of continued poverty.  As Alex Argenti-Pillen describes in her research among families of Sinhalese soldiers, ‘Most young soldiers in the national armed forces come from extremely poor backgrounds.  Their parents and siblings survive by means of casual labour in the tea plantations or rice paddies’ (Argenti-Pillen 2003: 2).Indeed, salary levels for military employment[6] are a magnitude of more than double what casual labourers would get, and almost double what salaried private sector jobs pay.  The only category of employment with earnings that approximate the military are civilian government jobs – which account for a very small fraction of the available employment for this group.  It is also important to bear in mind that private sector employment is concentrated in urban areas and the south-west, but it is quite absent in the rural periphery.

TABLE 3: AVERAGE SALARIES FOR   SINHALESE MALES, 18-26 (SL Rupees/month)
	Public	Military	Private	Casual
Mean	4,656	6,761	4,275	2,998
Median	4,000	6,550	3,800	2,750
Count	38	118	183	313
Source: SLIS 1999/2000

Finally, and as a corollary to this, the data shows that overall household incomes for military families (which of course includes all other sources of income) are significantly higher than those of non-military families. This is corroborated by qualitative studies which suggest firstly that families with military personnel are drawn from very impoverished backgrounds, and secondly that military employment has enabled these households to lift themselves out of poverty.  For example, one study describes how: ‘a soldier in the family can bring recognition, power and economic security to people who have no social status’:

‘Poor’ is how they generally describe their condition prior to their enlistment in the army.  Most have not possessed a permanent shelter prior to joining the military, but have lived in mud houses with cadjan or tin roofs. (Liyanage 2004: 28)

Indeed, 72% of the military households surveyed in this study had experienced a substantial improvement in their living standards, having either built or bought permanent houses.  Unlike others in the village, military households had television sets, refrigerators, proper furnishings, access to clean water and electricity, and ate better food.  One young farmer described the mobility effects of the military very clearly:

Small scale agriculture is an income source which hardly allows us to manage with the bare necessities. This source of income can never foster a significant change in living patterns, or assist us to achieve our long-term objectives. … Significantly, families which have members enlisted in the military have achieved very prominent economic development. (Liyanage 2004: 29)

These findings resonate closely with those of other qualitative surveys, such as Michele Gamburd’s study of military employment in Naeaegama in the south-western coastal strip, and the Asian Development Bank ‘Voices of the Poor’ report (Asian Development Bank 2001) from Moneragala and Hambantota – both of which find that the main reason cited for joining the army were the lack of other job opportunities and poverty (Gamburd 2004).

5. Conclusions: The statistical data on employment demonstrate the extent to which military employment had by the late-1990s, come to occupy a very particular niche in Sri Lanka’s labour market in terms of education, ethnicity and regional background.  The results in themselves are not entirely remarkable, for they are in conformity with and corroborate widely held perceptions and a considerable body of anecdotal evidence on the social characteristics of the Sri Lankan military forces.  They help to substantiate the argument that the civil war had perversely become an important source of livelihood diversification, asset accumulation and poverty alleviation for the rural Sinhalese population. Under circumstances of a secular long-term decline in small-holder farming, increasing rural-urban disparities, and the diminishing role of civilian state employment as a viable route for upward social mobility, military employment had, by the 1990s, come to occupy an important position in the economy of the Sinhalese village.  With 250,000 formal sector jobs, the security forces were the country’s single largest employer.  More importantly, they were the single most important source of employment among poor rural Sinhala-Buddhist youth, a group that has historically been at the centre of radical political activity.  By signing up a sizeable proportion of this group at the age of 18, 19 or 20, for a twelve year contract in what they perceive as a dangerous, yet lucrative job, the military  also played a unique and largely unrecognised role in redirecting this traditional social constituency of rebellion and political unrest in the direction of a hierarchical, nationalistic, conservative form of socialisation.

As discussed earlier, the instrumentality and functional benefits of military employment for the market reform agenda does not equate to a conscious conspiracy designed by self-interested political or economic elites.  As James Ferguson describes:

Whatever interests may be at work, and whatever they may think they are doing, they can only operate through a complex set of social and cultural structures so deeply embedded and so ill-perceived that the outcome may only be a baroque and unrecognisable transformation of the original intention. (Ferguson 1994: 17)

In discussing the possible reasons why military expenditure might benefit economic growth, Emile Benoit speculated that militarisation had important positive spillover effects:

Military training may be peculiarly effective in promoting certain modernizing attitudes and ways of life: in part because it has available the instruments of compulsion; in part because the value which justifies its activity national security-exercises a strong influence on most individual consciences and appears to justify imposing difficult and often painful adjustments on the individual. This process is facilitated by the fact that the military conscript is isolated from his family and community and living in an artificial community where he is under continuous government control. (Benoit 1978)

Indeed, I would argue that the ‘painful adjustment’ he refers to describes not just  the transformation of the individual military recruit, but is an apt metaphor for what is happening to Sri Lankan society and the economy as a whole, which is undergoing an equally painful adjustment to the market under the shadow of civil war (De Mel 2007).

Military recruitment, which escalated significantly in the 1990s, helped to cushion the impact of the market reform programme, in much the same way that the massive aid-funded rural development schemes did in the early 1980s; that is, by a compensatory logic through which the dismantling of the social democratic state, state employment, state patronage on the one hand, particularly through the vigorous privatisation programme of the Premadasa and Kumaratunga years, was balanced, offset, minimised and mitigated, through an expansion of the state, state employment, state patronage on the other.  As figure 4 shows, the steady growth in the numbers of the security forces over 1982-2002 has substantially compensated for the reduction in civilian public sector employment over these years.

FIGURE 4: CIVILIAN AND MILITARY COMPONENTS OF PUBLIC SECTOR EMPLOYMENT, 1982-2002

Source: CBSL Annual Report (various), and IISS Military Balance (various).

Again maybe it would be good to have a footnote here or before explaining that these are all based on your calculations

There are of course, important limitations to this analysis, as noted earlier.  The absence, and indeed, the impossibility of obtaining and testing valid counter-factuals to this proposition means that the functional relationship between war and market reforms outlined here remains tentative.  Yet I will conclude by asserting that it bears scrutiny, and has explanatory power.  For example, the idea of calibrating an analysis of the conflict against the politics of market reform can be fruitful in explaining the failure of the 2002 peace process (Venugopal 2009).  Similarly, there are important grounds to cast doubt on the idea that the parallel processes of development/ destruction, war/reform are not distinct spheres of analysis and policy formulation, but are in reality closely linked.

**  Rajesh Venugopal teaches at the London School of Economics, UK
REFERENCES

Argenti-Pillen, Alex. 2003. Masking Terror: How Women Contain Violence in Southern Sri Lanka. Philadelphia: University of Pennsylvania Press.

Asian Development Bank. 2001. Perceptions of the Poor: Poverty Consultations in Four Districts in Sri Lanka: Final Report. Asian Development Bank.

Arunatilake, N., S. K. Jayasuriya, and S. Kelegama. 2001. The Economic Cost of the War in Sri Lanka. World Development. 29: 1483-1500.

Ball, N. 1983. Defense and Development: A Critique of the Benoit Study. Economic Development and Cultural Change. 31(3): 507-524.

Baran, Paul. and Sweezy, P. 1966. Monopoly Capital, New York, Monthly Review Press.

Bauman, Zygmunt. 1989. Modernity and the Holocaust. London. Polity.

Bayly, C.A. 1988. Indian Society and the Making of the British Empire, Cambridge: Cambridge University Press.

Benoit, Emile. 1973. Defense and Economic Growth in Developing Countries. Lexington, MA: Lexington Books.

Benoit, Emile. 1978. Growth and Defense in Developing Countries. Economic Development and Cultural Change, 26 (2): 271–287

Central Bank of Sri Lanka. Annual Report (various years).

De Mel, N. 2007. Militarizing Sri Lanka : Popular Culture, Memory and Narrative in the Armed Conflict. London: Sage.

Deger, S. 1986. Economic Development and Defense Expenditure. Economic Development and Cultural Change, 35(1): 179-196.

Dunham, D., and S. Jayasuriya. 2000. Equity, Growth and Insurrection: Liberalization and the Welfare Debate in Contemporary Sri Lanka. Oxford Development Studies, 28.

Faini, R., P.Annez & L.Taylor. 1984. Defense Spending, Economic Structure, and Growth: Evidence among Countries and over Time.  Economic Development and Cultural Change, 32(3): 487-498.

Ferguson, James. 1994. The Anti-politics Machine: “Development,” Depoliticization, and Bureaucratic Power in Lesotho. Minneapolis: Minnesota University Press.

Fitzgerald, E.V.K., F. Stewart, and M. Wang. 2001. An Overview of the Case Studies  In Stewart, F., and V. Fitzgerald (eds.) War and Underdevelopment, Oxford: Oxford University Press.

Foucault, M. 1979. Discipline and Punish : the Birth of the Prison. Harmondsworth: Penguin.

Gamburd, M. 2004. The Economics of Enlisting: A Village View of Armed Service In Winslow, D. and M.Woost (eds.), Economy, Culture, and Civil War in Sri Lanka. Indiana University Press.

Government of Sri Lanka. 2003. Regaining Sri Lanka: Vision and Strategy for Accelerated Development. Colombo.

Government of Sri Lanka Department of Agriculture. Cost of Cultivation of Agricultural Crops (various years). Socio-Economic Planning Centre, Peradeniya.

Gunasinghe, N. 1984. Open Economy and Its Impact on Ethnic Relations in Sri Lanka In Committee for Rational Development (ed.), Sri Lanka, the Ethnic Conflict : Myths, Realities & Perspectives. New Delhi: Navrang.

Gunatilaka, R., and Chotikapanich, D. 2006. Inequality Trends and Determinants in Sri Lanka 1980-2002: A Shapley Approach to Decomposition. Business and Economics Working Paper No. 6/06, Melbourne: Monash University.

Haggard, S. and R. Kaufman.1992. The Politics of Economic Adjustment. Princeton, Princeton University Press.

Herbst, J. 1993. The Politics of Reform in Ghana. Berkeley: University of California Press.

Hettige, S. 2000. Economic Liberalisation, Qualifications and Livelihoods in Sri Lanka. Assessment in Education, 7(3): 325-334.

IISS (International Institute of Strategic Studies), The Military Balance, 1984-2004

Jenkins, R. 1999. Democratic Politics and Economic Reform in India. Cambridge: Cambridge University Press.

Keen, D. 2006. Endless War? Hidden Functions of the War on Terror. London: Pluto.

Keen, D. 2008. Complex Emergencies. London: Polity

Liyanage, S. 2004. ‘Coping with Vulnerability among the Families of Soldiers in a Context of Demobilization: Perspectives on Post-Conflict Peace Building’, In Asirwatham, R. and P. Thalaysingham (eds.), Poverty and Social Conflict in Sri Lanka: Integrating Conflict Sensitivity into Poverty Analysis. Colombo: Centre for Poverty Analysis and IMCAP.

Moore, M. 1990. Economic Liberalization Versus Political Pluralism in Sri Lanka. Modern Asian Studies, 24.

Narayan, A. and N.Yoshida .2005. Poverty in Sri Lanka: The Impact of Growth with Rising Inequality. Report No. SASPR-8: World Bank.

Nelson, J. 1990. Economic Crisis and Political Choice, Princeton: Princeton University Press.

O’Sullivan, M (2001): Sri Lanka: Civil Strife, Civil Society, and the State 1983-1995 in F. Stewart and E.V.K. FitzGerald (ed.), War and Underdevelopment, Oxford University Press.

Osmani, S.R. 1994. ‘Economic Reform and Social Welfare: The Case of Nutrition in Sri Lanka’. American Economic Review, 84(2): 291-296.

Peers, D. 2007. Gunpowder Empires and the Garrison State: Modernity, Hybridity, and the Political Economy of Colonial India, circa 1750– 1860, Comparative Studies of South Asia, Africa and the Middle East 27(2): 245-258.

Przeworski, A. 1991. Democracy and the Market: Political and Economic Reforms in Eastern Europe and Latin America. Cambridge, Cambridge University Press.

Pye, L. 1962. Armies in the Process of Political Modernization In Johnson, J.J. (ed.) The role of military in underdeveloped countries. Princeton, N.J., Princeton University Press.

Sarvananthan, M. 2008. Economy of the Conflict Region of Sri Lanka: From Embargo to Repression. East West Center Policy Studies 44.

Spencer, J. 2002. The Vanishing Elite: the Political and Cultural Work of Nationalist Revolution in Sri Lanka, In Shore, C. and S. Nugent (eds.) Elite Cultures: Anthropological Perspectives. London: Routledge.

Sri Lanka Labour Force Survey Annual Report (2000-2010)

Stein, B. 1985. State Formation and Economy Reconsidered. Modern Asian Studies, 19(3): 387-413.

Stewart, F., and E. V. K. Fitzgerald. 2001. War and Underdevelopment. Oxford: Oxford University Press.

Tilly, C. 1975. Reflections on the History of European State-Making,’ In Tilly, C., (ed.), The Formation of National States in Western Europe. Princeton University Press.

UNCTAD. 2000. Handbook of Statistics CD-ROM.

Venugopal, R. 2008. Cosmopolitan Capitalism and Sectarian Socialism: Conflict, Development and the Liberal Peace in Sri Lanka. D.Phil thesis, University of Oxford.

Venugopal, R. 2009. The Making of Sri Lanka’s Post-Conflict Economic Package and the Failure of the 2001-2004 Peace Process. In Newman, E., R.Paris and O.Richmond, (eds.), New Perspectives on Liberal Peacebuilding. Tokyo: UNU Press

Venugopal, R. 2010. The Politics of Market Reform at a Time of Ethnic Conflict: Sri Lanka in the Jayewardene years. in K.Stokke and J.Uyangoda (eds.) Liberal Peace in Question: Politics of State and Market Reforms in Sri Lanka. Anthem.

Washbrook, D. 1988. Progress and Problems: South Asian Economic and Social History c.1720-1860. Modern Asian Studies, 22(1): 57-96,

Wijeweera, A, and M.J. Webb. 2009. Military Spending and Economic Growth in Sri Lanka: A Time Series Analysis.  20(6): 499-508 please check if this is cited in text

World Bank. 1995. Sri Lanka Poverty Assessment. World Bank Report no.13431-CE.

World Bank. 2001. World Development Indicators CD ROM.

World Bank. 2004. Sri Lanka: Development Policy Review, World Bank Report No 29396-LK.