## Abstract

This paper examines causes and implications of Sri Lanka's political regime shift in February 2015, which has been widely hailed a victory for democracy, against the backdrop of the country's political and economic development in the post-independence era. The regime shift vividly demonstrates that voters, given the chance, turn against leaders they perceive to be corrupt, nepotistic, or needlessly divisive, even if they deliver handsome growth figures; there are limits to gaining political legitimacy in a multi-ethnic state simply by creating cleavages between majority and minority communities. For the first time, the minority communities in Sri Lanka appear to have felt themselves part of, and potentially an important influence on, the national political scene. It is, however, difficult at this stage to predict whether the regime change would usher in an era of ethnic harmony and robust economic growth. One hope is that war-weariness and discontent with the previous regime seem to have led to a greater willingness to accommodate diverse perspectives and demands within the political system.

## 1.  Introduction

Sri Lanka is one of the most heavily researched of the developing countries. It has repeatedly gained attention among scholars as a laboratory for studying issues central to the debates on sociopolitical and economic transformation in countries that gained independence from colonialism. In the 1960s and 1970s Sri Lanka attracted attention as an illustrative case in the debate on the growth-equity trade-off and the untoward consequences of prolonged adherence to a state-led import substitution development strategy. From the late 1970s it became an important case study for the analysis of the impact of economic liberalization and structural adjustment. Following the eruption of the ethnic conflict in the early 1980s there was a new focus on Sri Lanka as a test case for studying various facets of the interplay of government policies and social harmony in a multi-ethnic nation. It provided a fertile ground for studies of the art of unconventional warfare, of internationalized conflict resolution and peacemaking in protracted ethnic conflicts, and finally of how military means can be effectively used to defeat armed separatist movements. With the stunning change in the political regime at the 8 January 2015 presidential election, Sri Lanka now provides the international research community with another opportunity to undertake illustrative case studies of a range of issues relating to political transitions and regime change, post-conflict economic management and governance in a multi-ethnic country.

The outcome of the presidential election that toppled the incumbent, Mr. Mahinda Rajapaksa, and elected a new president, Mr. Maithripapa Sirisena, came as a surprise to most observers and analysts. The Rajapaksa government had crushed the Liberation Tigers of Tamil Eelam (LTTE, the “Tamil Tigers”) in 2009, ending a quarter-century-old civil war and won the election in 2010 with an 18 percent margin over his nearest rival, ex-army commander Sarath Fonseka, who had fallen out with the regime in the aftermath of the military victory. When Mr. Rajapaksa suddenly declared his intention to go for an early election in November 2014, two years ahead of schedule, his victory seemed a foregone conclusion: The regime appeared so strongly entrenched that the only issue seemed to be whether the size of the majority would be reduced.

The purpose of this paper is to study this dramatic shift in the political regime in Sri Lanka against the backdrop of longer term sociopolitical development in the post-independence era, placing emphasis on the determinants of the regime shift, and its implications for development prospects of the country and the resolution of the ethnic conflict. Section 2 begins with an overview of the historical background to the conflict and the political economy context that shaped policymaking in the post–civil war period. Section 3 discusses political developments and economic performance during the Rajapaksa era with particular attention to the sources of public discontent, which underpinned its election defeat. Section 4 deals with the regime shift and underlying causes drawing on the discussion in Section 2. The paper ends with some concluding remarks on the medium-term prospects for the Sri Lankan economy.

## 2.  Political and policy history

At independence in 1948, Sri Lanka's (Ceylon until 1972) economic prospects seemed highly promising. The colonial inheritance included a well-developed infrastructure, an efficient administrative mechanism and a thriving primary export sector with potential for substantial expansion. In terms of per capita income, literacy, and health care, Sri Lanka was ahead of most other countries in the region. It had a vibrant parliamentary democracy that had evolved through a step-by-step transition from a colonial system through a succession of constitutional reforms since the turn of the century. The population had enjoyed universal suffrage since 1931; only three years after universal suffrage was implemented in Great Britain. In these and other ways, Ceylon seemed blessed compared with its neighbors (Athukorala and Jayasuriya 1994; Wriggins 2011).

But the mobilization of this promising development potential for building the new independent Sri Lanka had to be done under the challenges of a plural polity that comprised a number of distinct ethnic groups with historically rooted cultural and linguistic differences. The major ethnic group, Sinhalese (70 percent of the population in 1946), trace their ancestry to a group of North Indian Aryansettlers believed to have arrived in the country around 500 BC. The second largest ethnic group, the Sri Lankan Tamils (11.0 percent), claim to have lived in the country for at least two millennia.1 The Indian Tamils (11.7 percent) are (mostly) descendants of migrant workers brought from South India by the British in the late 19th and early 20th centuries to work as indentured laborers on the tea and rubber plantations. The Muslims (Moors) (6.4 percent) trace their ancestry to Arab traders involved in the millennia-old seaborne trade between the Middle East and South and Southeast Asia. A smaller group, the Ceylon Burghers, are descendants of various European settlers (Athukorala and Jayasuriya 1994, Table 3.1).

These different ethnic groups, with some overlap, belong to different religious and linguistic groups. The Sinhalese are majority Buddhist and speak Sinhala. The majority of Tamils (both Sri Lankan and Indian) are Hindu and speak Tamil. Significant minorities in both Sinhalese and Tamil communities are Christians (predominantly Roman Catholic). The Muslims are the only ethnic group in Sri Lanka that are identified completely by their religion, but they speak Sinhala and/or Tamil, depending on where they live. These communities are geographically concentrated in different parts of the country.2 The Sinhalese live mainly in the southern, western and north central provinces. The Sri Lankan Tamils are concentrated in the northern province leading up to the Jaffna Peninsula and in the eastern province, though a significant number live in Colombo. Given its remote location in relation to the capital city (Colombo) and proximity to South India (Tamil Nadu), there remains distinctive sense of Tamil identity in Jaffna, which is not limited only to the language (Tamil) and religion (Hindu). Most Indian Tamils live in the plantations of the interior hill country. For the most part, they remain distinct as a community from the Sri Lankan Tamils. The majority of Muslims live in the coastal regions of the eastern province along with Sri Lankan Tamils and in small pockets in many urban centres elsewhere.

During the colonial era, a multi-ethnic, middle-class elite emerged with the expansion of the “modern” commercial and industrial sectors and the colonial administrative apparatus. Many Sri Lankan Tamils migrated from the traditional northern and eastern regions to Colombo to work in the state services and the commercial and financial sectors. English, the language of the colonial rulers, became the common language among elites of all ethnic groups, and proficiency in English has remained an important asset in both social and economic spheres. By and large, however, the deep divide among the major ethnic groups in terms of language, religion, and culture endured.

At the beginning of the 20th century the westernized Sri Lankan elite formed the Ceylon National Congress (CNC) taking a cue from the nationalist movement in India. Initially, ethnic divisions were secondary and the Sri Lankan Tamil elite played a pivotal role in the CNC, collaborating closely with Sinhalese and other ethnic elites in the political discourse with the colonial administration (De Silva 1986). This ethnic unity, however, eroded from the 1920s as the colonial administration embarked on a gradual process of devolution of power to the locals, and the basis for representation in the legislative council shifted from a communal to a territorial one.

The Sinhalese elite, recognizing the power of their majority position, began to exploit ethnic divisions by appealing to the hitherto dormant Sinhala nationalism. Nationalism also helped the elites in fending off the ideological challenge posed by the radical left, which became increasingly influential within the working class and the lower income groups from the late 1920s. Thus ethno-nationalism started to play a major role in shaping the formation and evolution of political organizations and parties. In this context, the Sri Lanka Tamils, with distinct regional concentration, a well-developed cultural identity, and a strong position within the colonial administration, began to be concerned about the likely architecture of the post-colonial independent state with growing fears that British rule would be replaced by Sinhalese rule. These concerns culminated in the quest for a “50–50” formula for representation in the legislature by G.G. Ponnambalam, the leader of the most prominent Tamil political party at the time, the Tamil Congress. To many Sinhalese, this became “a symbol of the unreasonable ambition of that energetic minority” (Wriggins 2011, 5). The Muslim elite, in contrast, opted to work with the Sinhalese elite (Shastri 1997).

At independence in 1948 Sri Lanka adopted a constitution based on the British Westminster parliamentary model. The Sri Lankan constitution was not based on a broad-based process of negotiation between the various ethnic groups. Rather, it came out of discussions between colonial administrators, the Sinhalese-elite dominated Board of Ministers, and the Sinhalese elite represented by Don Steven Senanayake. The issues of particular concern in a multi-ethnic state, such as citizenship, franchise, and individual and group rights, were not discussed or agreed to by representatives of the country's largest ethnic groups. Unlike in the Indian constitution, the Sri Lankan constitution did not include an explicit bill of individual rights or anything resembling effective formal protection for minorities other than creating a number of urban multi-member constituencies in multi-ethnic localities. Every important matter of government policy was to be defined and implemented by the national government in Colombo. Municipal, town and village councils were responsible for only a limited range of matters of strictly local concern. In the process of designing the constitution Tamil demands and proposals were treated as acts that would delay or obstruct “Ceylon's march to freedom” (De Silva 1986, 146–7). Consequently, the new constitution was not approved by 5 of the 10 members of the legislative council elected by the Tamil community.3 Thus the widely-held view that the transfer of power in 1948 was peaceful and orderly is deceptive.

The conservative United National Party (UNP), led by Don Steven Senanayake, emerged as the largest party at the first general election held in 1947, but it won only 40 percent of the seats. The UNP formed a government with Senanayake as Prime Minister and Solomon West Ridgway Dias Bandaranaike as his deputy, in a coalition with the conservative Tamil and Muslim groups. The Trotskyite Lanka Sama Samaja Party, which had a strong multi-ethnic power base, emerged as the main opposition in the new parliament. The new government moved swiftly to pass legislations to deprive Indian Tamils of citizenship and voting rights to contain the power of the left. The disfranchising of Indian Tamils enhanced the electoral weight of the Sinhalese population in the central highlands of the country, creating an additional inducement for political parties to raise issues that would attract the votes of the Sinhalese electorate (Jennings 1954).4

In September 1951 Bandaranaike formed a new political party, the Sri Lanka Freedom Party (SLFP), because of his “dismay over his prospects within the ruling party” (Manor 1989, 192). Over the next four years the SLFP developed a strong support base in rural Sinhalese constituencies on a promise to make Sinhala the official language, promote Buddhism and Sinhalese culture, and reform the system of education to benefit the Sinhalese rural middle- and lower-middle classes. The 1956 election was the first where ethnic polarization based on the official language issue played a dominant role. The UNP suffered a landslide defeat at the hands of a three-party coalition (Mahajana Eksath Peramuna [MEP]) headed by the SLFP.

The first action taken by the new MEP government was to pass the Official Language Act of 1956, a law that declared Sinhala the country's official language. An agreement was subsequently negotiated between Premier Minister Bandaranaike and S. J. V. Chelvanayagam, the leader of the SLFP, in July 1957 (the Bandaranayike-Chelvanayagam Pact) to give regional autonomy for Tamil areas (in the northern and eastern provinces) within the existing framework of unitary state and to recognize Tamil as the language of government administration in these areas. Protests and demonstration by Sinhalese extremists within the ruling party and by the UNP (led by J. R. Jayawardene), however, forced Bandaranaike to abrogate the pact. Tamil protests against efforts to extend the use of Sinhala as the official language in their areas escalated into an even more horrific clash. In 1958, a Tamil language act was approved by the parliament to provide for “reasonable” use of Tamil in northern and eastern provinces and in the national government, but it was never implemented.5

The change in government in 1956 was a political watershed in Sri Lankan history in which relations between the Sinhalese and Tamil communities took a sharp turn for the worse. From then on, Sri Lankan politics entered into a pattern of competitive bidding by the major Sinhalese-dominated parties for the support of the Sinhalese community. Within a few years even the main left-wing parties abandoned their multi-ethnic posture. During the next three decades, no serious attempt was made to address the language and regional autonomy issues with a view to achieving ethnic harmony. In 1965 the government of Dudley Senanayake (the UNP-led coalition with SLFP as one of the coalition partner) negotiated a devolution agreement (District Council Bill) with SLFP broadly along the lines of the defunct Bandaranaike-Chelvanayagam agreement. It was opposed by the SLFP-led opposition, in an almost exact replay of the 1956 event with roles reversed (De Silva 1986, 193).

The United Front government (a SLFP-led coalition), which came to power at the 1970 election, promulgated a new constitution in 1972. It accorded Buddhism the “foremost place” in the country's religious hierarchy and affirmed the role of Sinhala as the official language. The limited preferential representation given to the Tamils in the Western province under multi-member constituencies was abolished under the electoral rules of the new unicameral legislature.

The United Front government also introduced a new system of university admission involving standardization of marks across language groups and by regions, in place of the earlier system based solely on academic achievements tested through an open competitive examination (De Silva 1978). This de facto ethnic-based quota system worked to the relative disadvantage of the Sri Lankan Tamils, who had for years enjoyed a position of prominence in the science-based faculties because of their higher rate of literacy in English and better facilities for science education in the schools of the Jaffna district. Under the new system, the percentage of university places in the science-based disciplines (medicine, engineering, and natural sciences) held by Tamils fell from 35.3 percent in 1970 to 20.9 percent in 1974 and 19 percent in 1975; there was for the first time a substantial decline in even their absolute number despite a continuous increase in the total intake of students in these courses.

The Tamil discontent, resulting from policy failures in devolution and the discriminatory university entry system, was aggravated by rising youth unemployment in the context of a near stagnant economy (Wriggins 2011; Athukorala and Jayasuriya 2013). Starting in the late 1950s successive governments vigorously pursued a state-led import-substitution strategy as the basic tenet of the country's development strategy. By the mid 1970s, Sri Lanka had become one of the most inward-oriented economies of the world, with an economy dominated by extensive state ownership. The government had become the “employer of last resort” and the public sector outpaced the private sector in employment creation. As the economy began to experience massive unemployment resulting from a combination of slow economic growth and rapid population growth, employment in the public sector became increasingly subject to patronage politics and access to political power. Tamil youth were further disadvantaged in gaining public sector employment by the imposition of proficiency requirements in Sinhalese. If a development strategy that permitted greater scope for employment creation outside the state sector had been adopted, alternative jobs might have been more widely available to soften the growing resentment among unemployed Tamil (and also Sinhalese) youth.6

These developments provided the breeding ground for separatist ideologies among the Tamil youth who became alienated from mainstream politics. Several youth-based movements committed to a struggle for a separate state emerged in Jaffna. In response to the growing radicalization among the youth, the traditional Tamil political parties came together in 1976 to form the Tamil United Liberation Front (TULF), which committed itself to establishing a separate Tamil state. In turn, the demand for an independent state that claimed nearly one-third of the land area exacerbated Sinhalese hostility toward the Tamil community.

At the parliamentary election held in 1977, the UNP led by J. R. Jayawardene swept into power with a landslide victory thanks to the widespread voter disenchantment created by the dismal performance of the economy under the previous dirigiste economic policy regime. The TULF, which swept the northern and eastern provinces on a platform that called for the establishment of an independent Tamil state, became the main opposition party.

The new government launched a series of fundamental liberalization policy reforms that marked a decisive break from the decades of import substitution industrialization policies. In the aftermath of the liberalization reforms, the country seemed poised to embark on a trajectory of rapid growth that would enable it to emulate the dynamic East Asian economies. The economy failed to maintain the initial growth dynamism, however, because other components of the policy package undermined the initial stimulus to the country's international competitiveness. The opening up of the economy was accompanied by huge and often wasteful foreign-aid-supported public investment program. The inflationary side effects eroded the initial salutary effect of trade and exchange rate reforms on the competitiveness of tradable sectors, and constrained the growth of exports and efficient import-competing industries. The government also failed to put adequate resources into upgrading essential infrastructure and communications to complement trade and investment liberalization as the emphasis was on politically appealing mega-infrastructure projects and on attaining self-sufficiency in rice (the main staple) (Athukorala and Jayasuriya 1994; Snodgrass 1999; Wriggins 2011).

The designers of the liberalization policy package also failed to take into account the possible adverse side effects of its immediate distributional effects on the social fabric of a multi-ethnic, socially stratified society highly sensitive to equity concerns. The liberalization reform package was implemented in a non-uniform and discriminatory manner; it favored different regions, sectors, and particular segments of the population. The abolition of import quotas and lowering of tariffs on imports of several commodities (such as onions, red chillies, and some fruits and vegetables) cultivated mainly in the Jaffna peninsula had a disproportionate adverse impact on Tamil farmers. The growth in trade, export-oriented industrial production, tourism, urban development, and construction activity mainly benefited the Sinhalese-dominated southwest regions of the island. But even there the benefits were not shared equally: The real income of many poorer people, particularly in urban areas, declined because of drastic cuts to food subsidies and high inflation (an outcome of the macroeconomic policy slippage noted earlier) whereas Tamils engaged in private business and commerce in these areas benefited from the reforms (Tambiah 1997). As discussed subsequently, this situation was exploited by various groups as an excuse to unleash anti-Tamil violence.

The government used its two-thirds parliamentary majority to promulgate a new constitution, a presidential-parliamentary system patterned after Charles de Gaulle's French constitution of 1958. The new system established a directly elected executive president with wide powers, who would appoint the Prime Minister and Cabinet from members of the parliament, and preside over their workings, without being personally responsible to the parliament. A unicameral legislature was to be elected on the basis of proportional representation, with individual votes cast for party lists in multimember constituencies. The expectation was that the new constitution, by making it extremely difficult for any single party to get a two-thirds majority in the parliament, would provide political stability conducive to economic transformation under market-oriented policy reforms.

But any expectation that Jayawardene would use his victory to undertake constitutional reforms and satisfy moderate Tamil demands for greater devolution of power to the Northern and Eastern provinces proved to be misplaced. The TULF leadership sought some concessions from the government, pointing to the danger of violent action by extremist groups, but the government, with ethnic chauvinism rampant in its own ranks, refused to budge. This accelerated the radicalization of Tamil youth and their embrace of separatism. Among several militant separatist Tamil organizations, the LTTE) led by the charismatic and ruthless Velupillai Prabhakaran, eventually became dominant, using brutal violence to eliminate challengers. It developed a highly disciplined and well-organized military force with support networks within the Tamil diaspora in countries such as India, Canada, Australia, the UK, France, the United States, and Singapore (Wayland 2004).

The first signs of violent activity by Tamil militant groups—attacks on government officials and security personnel—began to surface in the northern province in the mid 1970s. In response to the rising hostility among the Tamils, the government implemented several institutional measures, including recognition of Tamil as a “national language” along with Sinhala, and permitting it to be used along with Sinhala for administrative, judicial, and educational purposes in the northern and eastern provinces. A belated effort was made to provide for a modest measure of regional devolution of power by passing legislation in 1981 to establish District Development Councils. The traumatic events of July 1983 (“Black July”), however, swept away the District Development Council initiative (Roberts 1986).

Black July marked a historic break in the relations between the Sinhalese majority and the Tamil minority. An attack on security forces by an armed group of the LTTE in Jaffna on 23 July led to a violent anti-Tamil program and mob attacks with widespread looting, arson, and violence in the southern parts of the country, including the capital, Colombo, that resulted in hundreds of deaths, and the start of the mass exodus of Tamil refugees from Sri Lanka. The manner in which the riots were handled by the government deepened Tamil resentment against the Sinhala-dominated government, and increased sympathy and support for separatism (Biziouras 2014).

Immediately after Black July, the government passed a constitutional amendment requiring members of parliament to take an oath explicitly disavowing support for separatism. The moderate Tamil members of parliament, who had been elected in 1977 from the north and east on a platform that called for the establishment of a separate Tamil state, could not take such an oath and found themselves expelled from the parliament, leaving no legitimate representation of the Tamils. The cycle of violence and retaliation escalated and the LTTE established a de facto separate state in parts of the north and east of the country. LTTE attacks on government establishments and the Sinhalese civilians rapidly spread not only throughout the northern and eastern provinces but also to the Sinhalese-dominated areas.

As the conflict escalated, a large number of Tamil refugees fled across the water to Tamil Nadu, an ethnic Tamil state in South India. Various militant groups started using Tamil Nadu as a base for conducting the armed conflict in Sri Lanka (De Silva 1986; Bandarage 2009). These events strained the relationship between Sri Lanka and India. In 1987 the Sri Lankan government signed an accord with India. Under the terms of the accord, India sent an Indian peacekeeping force to Sri Lanka to monitor a cease-fire in the civil war and to disarm the LTTE militants to prepare the ground for a settlement within the context of a unitary state. In return, Sri Lanka agreed to resolve the Tamil problem through regional decentralization of power. The Indian peacekeeping force, however, failed to disarm the LTTE and was consequently compelled to fight the LTTE. Because of divisions within its own ranks, the Sri Lankan government failed to devolve adequate power under the weakly instituted provincial councils and failed to win over any significant segment of Tamil opinion. President Premadasa, who assumed power in 1989 after two terms of the Jayawardene presidency, demanded that the Indian peacekeeping force leave and it completed its withdrawal in 1990.

At the 1994 presidential election, Chandrika Kumaratunga (the daughter of two previous prime ministers, S. W. R. D. Bandaranaike and Sirimavo Bandaranaike) came to power as the leader of the People's Alliance on a ticket seeking peace with the LTTE.7 The new government offered a cessation of hostilities and unconditional talks to the LTTE, which led to a brief ceasefire and hopes of a negotiated settlement. Nevertheless, the proposed devolution of power to the provincial councils fell well short of the aims of the LTTE, which stalled negotiations and then broke the ceasefire in April 1995, reigniting the armed conflict. After a botched assassination attempt by the LTTE on the president, the government launched a concerted offensive against the LTTE and in December 1995 the government forces recaptured the city of Jaffna, which had served as the de facto capital of the LTTE mini-state. This forced the LTTE to rely primarily on guerrilla warfare from then on, operating primarily in the jungle areas of the northern province and parts of the eastern province. A new attempt at a negotiated peace led to a Norwegian-brokered cease-fire agreement in February 2002, which remained largely intact until late 2006 (though its terms were frequently violated in practice).

The dynamics of the conflict began to change following the 2005 presidential election. Mahinda Rajapaksa of the People's Alliance secured a narrow victory over his closest rival Ranil Wickramasinghe (UNP) (53.0 percent and 48.4 percent of the vote, respectively). Though the new government initially continued to adhere to the peace agreement, and the UNP, on its part, backed the government, after two rounds of failed peace talks and intensified military action by the LTTE, the government began to prepare for an all-out military offensive.

In 2007, President Rajapaksa strengthened his political position for the war effort by engineering the defection of 19 members from the opposition UNP to his governing coalition and by forging alliances with several groups of Sinhala Buddhist nationalists and the left-wing nationalist Janatha Vimukthi Peramuna (JVP). The government, with support and assistance from the Western powers, India, Pakistan, and, particularly, China, sharply increased the military budget and escalated the military campaign to annihilate the LTTE. In 2009 June the LTTE was defeated as a fighting force and its leadership was decimated (De Silva 2012).8

## 3.  Post-conflict politics and economics

After the country returned to a state of normalcy at the end of the 30-year-old civil war in May 2009, President Rajapaksa consolidated power by calling fresh presidential and parliamentary elections in 2010 and winning both decisively.9 Soon after the parliamentary election victory, he used the two-thirds majority in the parliament to enact the 18th amendment to the constitution that vastly extended the executive powers of the presidency. This amendment abolished the requirement for the president to stand down after two 6-year terms, and authorized him to appoint judges of the Supreme Court, the Attorney General, the Auditor General, the Ombudsman, and the members of important bodies including the Judicial Service Commission, the National Police Commission, the Human Rights Commission and the Election Commission, and the Commission to Investigate Allegations of Bribery and Corruption (IBA 2013).

The president used the new powers and a combination of inducements and intimidation to further divide and weaken the opposition (Goodhand 2012; Jayasuriya 2013). By 2012 a tiny family group comprising the president, his brothers, and his elder son, exercised almost total control over the financial, security, and judicial powers of the state. The bureaucracy was expanded to provide an independent powerbase for the regime.10 In 2013 the Chief Justice was sacked as she delivered rulings against some government decisions, and a supporter of the regime was appointed in her place. Paramilitary agents, allegedly sponsored by the state, continued to intimidate political opponents and critics: “fear of the white van without the licence plate became a common theme in day-to-day conversations” (Holt 2011, 715).11 On the Reporters san Frontieres World Press Freedom Index, Sri Lankan's rank dropped from 51st among 139 countries in 2002 (the initial year of the index) to 165th out of 198 countries in 2014.12 As we have observed elsewhere, “The situation in Sri Lanka (was) increasingly exhibiting striking parallels with the Marcos regime in the Philippines and the final phase of the Suharto regime in Indonesia” (Athukorala and Jayasuriya 2013, 24).

There were expectations in the aftermath of the military victory that the government would move towards the long-delayed devolution of power to provincial councils. But this was not to be. The government appointed a Lessons Learned and Reconciliation Commission to come up with recommendations for political reconciliation, but then swept the commission report under the carpet. Instead, the government continued to maintain a strong military presence in the Jaffna Peninsula. Land taken over by the military allegedly for security reasons was reportedly used to run military-operated hotels and other commercial enterprises (Transparency International 2014).

Despite its authoritarian nature, the regime continued to enjoy substantial support from Muslim and other non-Tamil minority communities in the first few years of the post-conflict period; authoritarian actions were tolerated as the price that had to be paid for the successful prosecution of the war. But Muslim support for the government eroded following the campaign from 2012 onward by the Bodu Bala Sena (BBS), an extreme Sinhala Buddhist organization closely associated with Gotabhaya Rajapaksa, the Defense Secretary and a brother of the President, to entrench Sinhala Buddhist supremacy in the country. In 2013 BBS called for an end to halal certification, and declared that it will destroy a 10th-century mosque allegedly built by destroying a Buddhist temple. The BBS's anti-Muslim campaign culminated in violence in a southern Sri Lankan town in June 2014 (Bastian 2014). Though the government officially dissociated itself from the anti-Muslim campaign of the BBS, it did nothing to prosecute it or otherwise censure its actions. The BBS campaign to entrench Sinhala Buddhist supremacy also led Christians to seriously question their future in the country under the Rajapaksa regime.

### 3.1  Policy shifts

The new development strategy of the Rajapaksa regime marked a significant departure from the market-oriented policy stance maintained for over three decades from the late 1970s and harked back to the populist economic policies of the 1960s and 1970s (Athukorala 2012). It emphasized the role of the state in “guiding the markets” with a view to redressing alleged untoward effects of economic globalization. As has so often happened in the history of developing countries (Whitehead 1990), this kind of economic ideology was ideally suited to the needs of the regime that sought a developmental rationale for expanding centralized state power.

The trade regime was the first victim of this policy reversal. Import substitution in both manufacturing and agriculture was put back on the policy agenda. New export taxes were introduced on tea and rubber exported in raw and semi-processed form to promote further domestic processing of these products. Arbitrary increases in duties, driven by both revenue-raising and protectionist objectives, resulted in higher levels and inter-industry variations in the effective rates of protection.

The foreign investment approval procedures also became more interventionist and opaque following the promulgation of the Strategic Development Projects Act in 2008, empowering the minister in charge of the Board of Investment (BOI) to grant exemptions from all taxes for a period of up to 25 years to “strategic development projects.” A Revival of Underperforming Enterprises and Underutilized Assets Act was passed in November 2011, empowering the government to acquire and manage 37 “underperforming” or “underutilized” private enterprises. A number of new state-owned enterprises were set up, including a second national airline with a name adapted from the president's first name (Mihin Air).

On the macroeconomic front, there emerged a fundamental contradiction between exchange rate policy and fiscal policy and monetary policies. The Central Bank maintained a stable nominal exchange rate of the rupee vis-à-vis the U.S. dollar by drawing on foreign reserves and foreign borrowing, in a context where fiscal and monetary policy excesses continued to fuel domestic inflation. The persistent stability of the nominal exchange rate, coupled with higher domestic inflation compared to that of the trading partner countries, resulted in an appreciation of the real exchange rate by about 20 percent during 2005–13 compared with the previous five years, eroding the competitiveness of export-oriented and import-competing production in the economy (Table 1).

Table 1.
Sri Lanka: Selected macroeconomic indicators, 2004–13
Indicators2004200520062007200820092010201120122013
GNP per capita at current price, US$1,062 1,241 1,421 1,617 2,014 2,057 2,400 2,836 2,862 3,191 GNP per capita at constant (2005) price, US$ 1,182 1,241 1,323 1,400 1,471 1,505 1,611 1,727 1,818 1,920
Real GDP (2002 price) growth, % 5.4 6.2 7.7 6.8 3.5 8.3 6.4 7.3
Unemployment rate, % 8.3 7.7 6.5 5.4 5.8 4.9 4.2 4.0 4.4
Inflation (CPI) rate, % 7.6 11.6 10 15.8 22.6 3.5 6.2 6.7 7.6 6.9
Budget deficit, % of GDP −7.5 −7 −7 −6.9 −7 −9.9 −8 −6.9 −6.5 −5.9
Public debt outstanding, % of GDP 102.3 90.6 87.9 85 81.4 86.2 81.9 78.5 79.2 78.2
Of which foreign debta 47.6 39 37.6 37.1 32.9 36.4 36.1 35.6 36.5 34.1
Exchange rate, Rp/US$101.2 100.5 104 110.6 108.3 114.9 113.1 110.5 127.6 129.1 Real effective exchange rate (2004 = 100)b 100 92.8 90.6 90.5 77 75.1 73.4 71.9 77.4 73.8 Trade balance, % of GDP −10.9 −10.3 −11.9 −11.3 −14.7 −7.4 −9.7 −16.4 −15.8 −9.4 Exports (FOB) 27.9 26.0 24.3 23.6 19.9 16.8 17.4 17.8 16.4 15.5 Imports (CIF) 38.7 36.3 36.3 34.9 34.6 24.3 27.1 34.2 32.3 26.9 Current account balance, % of GDP −3.1 −2.7 −5.3 −4.3 −9.5 −0.5 −2.9 −7.8 −6.6 −3.9 Total external debt to GDP ratio, % 54.9 46.5 49.4 51.0 43.7 49.7 50.1 55.4 62.5 59.2 Debt service ratioc, % 11.6 7.9 12.7 13.1 18 22.4 16.7 12.7 19.7 25.3 Gross foreign-exchange reserves, US$ millionc 1,834 2,508 2,526 3,062 1,594 4,897 6,410 5,758 6,677 7,041
In months of imports 2.8 3.4 3.3 3.7 1.4 3.9 3.5 3.2 3.6 3.3
As a percentage of short-term debt 88.1 107.6 96.1 84.8 38.7 91.5 106.7 69.2 66.3 78.0
Net foreign-exchange reserves, US$milliond — — — — — 4150 5072 4011 4162 4597 In months of imports — — — — — 3.3 2.8 2.3 2.2 2.2 As a percentage of short-term debt — — — — — 77.5 84.4 48.2 41.3 50.9 Foreign direct investmente, US$ million 217 234 480 603 752 404 478 956 941 919
Indicators2004200520062007200820092010201120122013
GNP per capita at current price, US$1,062 1,241 1,421 1,617 2,014 2,057 2,400 2,836 2,862 3,191 GNP per capita at constant (2005) price, US$ 1,182 1,241 1,323 1,400 1,471 1,505 1,611 1,727 1,818 1,920
Real GDP (2002 price) growth, % 5.4 6.2 7.7 6.8 3.5 8.3 6.4 7.3
Unemployment rate, % 8.3 7.7 6.5 5.4 5.8 4.9 4.2 4.0 4.4
Inflation (CPI) rate, % 7.6 11.6 10 15.8 22.6 3.5 6.2 6.7 7.6 6.9
Budget deficit, % of GDP −7.5 −7 −7 −6.9 −7 −9.9 −8 −6.9 −6.5 −5.9
Public debt outstanding, % of GDP 102.3 90.6 87.9 85 81.4 86.2 81.9 78.5 79.2 78.2
Of which foreign debta 47.6 39 37.6 37.1 32.9 36.4 36.1 35.6 36.5 34.1
Exchange rate, Rp/US$101.2 100.5 104 110.6 108.3 114.9 113.1 110.5 127.6 129.1 Real effective exchange rate (2004 = 100)b 100 92.8 90.6 90.5 77 75.1 73.4 71.9 77.4 73.8 Trade balance, % of GDP −10.9 −10.3 −11.9 −11.3 −14.7 −7.4 −9.7 −16.4 −15.8 −9.4 Exports (FOB) 27.9 26.0 24.3 23.6 19.9 16.8 17.4 17.8 16.4 15.5 Imports (CIF) 38.7 36.3 36.3 34.9 34.6 24.3 27.1 34.2 32.3 26.9 Current account balance, % of GDP −3.1 −2.7 −5.3 −4.3 −9.5 −0.5 −2.9 −7.8 −6.6 −3.9 Total external debt to GDP ratio, % 54.9 46.5 49.4 51.0 43.7 49.7 50.1 55.4 62.5 59.2 Debt service ratioc, % 11.6 7.9 12.7 13.1 18 22.4 16.7 12.7 19.7 25.3 Gross foreign-exchange reserves, US$ millionc 1,834 2,508 2,526 3,062 1,594 4,897 6,410 5,758 6,677 7,041
In months of imports 2.8 3.4 3.3 3.7 1.4 3.9 3.5 3.2 3.6 3.3
As a percentage of short-term debt 88.1 107.6 96.1 84.8 38.7 91.5 106.7 69.2 66.3 78.0
Net foreign-exchange reserves, US$milliond — — — — — 4150 5072 4011 4162 4597 In months of imports — — — — — 3.3 2.8 2.3 2.2 2.2 As a percentage of short-term debt — — — — — 77.5 84.4 48.2 41.3 50.9 Foreign direct investmente, US$ million 217 234 480 603 752 404 478 956 941 919

Source: Data on net foreign-exchange reserves are from IMF (2013 and 2014); all other data are compiled from Central Bank of Sri Lanka, Annual Report (various years).

Notes:a. The figures reported here do not include purchase of treasury bills by foreign investors, which increased from US$63 million (0.1 percent of GDP) to US$ 92 million (1.6 percent of GDP) in 2012.

b. Original Central bank index inverted: an increase (decree) implies depreciation (appreciation).

c. Net of Asian Clearance Union debit balances.

d. Excluding foreign exchange swap with domestic financial institutions and short-term contractual repayment obligations of the Central bank.

e. Net of foreign-currency loans received by the enterprises approved by the Board of Investment.

— Data not available.

Rapid infrastructure development became the key policy priority of the Rajapaksa regime. A large-scale reconstruction effort with substantial public sector involvement was clearly needed after a quarter century of destruction, neglect, and decay of essential infrastructure. Nevertheless, the prioritization and economic efficiency of many government infrastructure projects, such as a modern port and other facilities (built with Chinese assistance) in the heartland of the electoral support base of the President are questionable (Law & Society Trust 2012; Sarvananthan 2015). Also, the faster than programmed, lumpy disbursements for a couple of large foreign financed infrastructure projects and for their counterpart funds (IMF 2011) became a major factor in the widening of the fiscal deficit and erosion of macroeconomic stability.

The massive construction projects also became the focal point of widespread public concern and complaints about financial excesses and rampant corruption. According to a study conducted by a senior professor of Transport & Logistic Management at a major Sri Lankan university (Kumarage 2014), the cost of highway construction (measured by cost per km) increased two to three times over a period of five to six years, a rate of increase that far exceeds the rate of construction cost increases even in developed countries such as Australia.13 In a project by project analysis, Kumarage (2014) finds that costs in projects awarded by competitive bidding exceed costs in projects offered without such bidding to be higher by about 130 percent. The value of contracts awarded without competitive bidding reached a peak of US$2,540 million in 2014.14 Kumarage's findings are consistent with those recorded in the emerging literature on the costs of road construction and associated corruption (Tanzi and Davoodi 1998; Kenny 2009; Collier, Kirchberger, and Soderbom 2013). Collier, Kirchberger, and Soderbom (2013), in particular, find that road construction costs are higher in countries with higher levels of corruption. ### 3.2 Economic performance The Sri Lankan economy grew at an average annual rate of 7.0 percent from 2010–14. This turned out to be the five-year period of fastest growth in the country's post-independent history (Table 1). Per capita income increased from US$ 1,062 in 2004 to US$3,191 in 2013. The rate of inflation came down from a historical high of 22.6 percent in 2008 to an average annual rate of 6 percent during the following five years. The unemployment rate fell from 8.3 percent in 2004 to 4.4 percent in 2013. Between 2006 and 2012, the poverty head count ratio declined from 15.2 percent to 6.7 percent, accompanied by a reduction in the poverty gap from 3.1 percent to 1.7 percent. The Gini coefficient also declined from 0.40 to 0.36 between these two years, indicating that rapid growth was accompanied by an improvement in income distribution (CBSL 2013). Nevertheless, these impressive headline economic figures need to be treated with caution for a number of reasons. Sri Lanka's system of national accounts has not been revised/updated since the early 1970s and “[t]he national accounts suffer from insufficient data sources and underdeveloped statistical techniques” (IMF 2014, Statistical Appendix). Given the unavailability of detailed data needed to measure both output and intermediate inputs, some of the gross value-added figures are estimated indirectly using fixed ratios obtained from outdated studies or based on ad hoc assumptions. Of course, although the resulting biases can go either way, such a virtual “non-system” naturally leaves room for “creative” accounting. This is particularly important because of concerns raised about political influence on the generation of sensitive data following the transfer of the compilation of national accounts and the consumer price index (CPI) from the Central Bank to the Department of Census and Statistics (DCS), which comes directly under the President.15 The new CPI compiled by the DCS since 2007 has excluded alcohol (a commodity that accounts for a substantial share of household expenditure, particularly in working class households) from the commodity basket, and some important items (such as the cost of transport and housing) are estimated using regulated (controlled) rather than market prices. Quite apart from understating the cost of living, these limitations of the CPI are bound to overstate the rate of growth in the economy because CPI and its sub-indices are used for estimating real value-added in a number of sectors (in particular, many subcategories in the services sector) (Shourie 1974). Data on poverty and income distribution, which are usually published at the national level, are naturally subject to aggregation biases. For instance, in a disaggregated analysis of unpublished household survey data, Sarvananthan (2015) shows that poverty levels in some districts and sub-regions in the eastern and northern provinces still remain stubbornly high, notwithstanding massive government infrastructure investment in these areas. Relating to income inequality, there is a large difference between the latest available figure of the Gini ratio (for 2012) as reported in the Central Bank report (0.36) and the one reported by the Department of Census and Statistics for the same year (0.48) (SLDCS 2015). Even if we take the official data at face value, there are several qualifications that must be made to this rosy picture when we analyze the overall growth experiences from a long-term sustainability perspective. First, the main drivers of growth have been the non-tradable sectors (construction, transport, utilities, trade and other services), driven largely by the major public sector infrastructure development projects. Over 70 percent of the total increment in real GDP between 2004 and 2013 originated in these sectors. The manufacturing sector grew only at a modest rate, resulting in a decline in its share in GDP from 18.5 percent during 2000–04 to 16.5 percent during 2005–13. Within manufacturing, the largest contributor to growth was the food, beverages, and tobacco product sector where production is predominantly domestic market–oriented. Sectors such as non-metallic mineral products, rubber and plastics, and miscellaneous manufacturing (where export production is concentrated), have recorded much slower growth.16 In sum, the sectoral profile of economic performance in recent years is consistent with the erosion of the competitiveness of traded goods production (real exchange rate appreciation) noted in the previous section. Second, the three-fold increase in per capita income in current US$ terms between 2004 and 2013 partly reflects domestic inflation and the artificial stability of the exchange rate of the Sri Lankan rupee against the dollar. When the data are expressed in real (2005) prices to allow for these factors, per capita income in 2013 (US$1,920) was only 62 percent higher than that in 2004 (US$ 1,182) (Table 1).

Third, the decline in the unemployment rate was partly due to an increase in public sector recruitments and a surge in overseas employment of Sri Lankans. The share of public sector employment in the total labor force increased from 10.4 percent in 2005 to 14 percent in 2012. During 2002–12, annually, on average, a quarter of a million Sri Lankans left for overseas employment, with the number increasing every year. The total stock of Sri Lankan overseas contract migrant workers would have reached two million by 2011, amounting to over 23 percent of the total labor force (Arunatilake, Jayawardena, and Weerakoon 2011).

Fourth, although the official aggregate figures show a notable increase in total foreign direct investment (FDI) inflows during the past three years, data at the sector/industry level reveal that the increase has come largely from projects in the construction and services sectors. During 2010–13, manufacturing accounted for only 31 percent of total realized FDI. The bulk of these flows were to domestic market–oriented industries (mostly food and beverages), with garments being the only export-oriented industry to attract some FDI. There is evidence that a large number of export-oriented foreign firms have closed down their operations in Sri Lanka. A comparison of the firm-level records of the BOI shows that 465 firms that were in operation in 2002 had disappeared from the BOI list in 2009. This number is too large to be interpreted solely as a recording error. Of these firms, the majority are firms with foreign capital participation (joint venture or fully foreign-owned). In contrast, the majority of newly established firms (over 80 percent) are fully locally owned. Investors from India, who have set up production bases to benefit from tariff concessions under the Sri Lanka–India free trade agreement, now dominate the list of firms whereas many firms from Korea, Hong Kong, and from a number of developed countries have left Sri Lanka (Athukorala 2012).

Fifth, the external payments position of the country has deteriorated over the past three years. In 2013, total imports were double the size of export earnings. There has been a massive contraction in exports of goods and services as a share of GDP, from an average level of 25.6 percent during 2004–09 to 16.8 percent during 2010–13. In 2013, export earnings covered only 57 percent of total outlay on imports. Although weak global demand in the aftermath of the 2008–09 global financial crisis and the recent withdrawal of “GSP Plus” tariff concessions by the EU would have played a role, a comparative analysis of Sri Lanka export performance suggests that the problem is mostly home grown (Rajapatirana 2013). Viewed against the experiences of the 1980s and 1990s, the continuous appreciation of the real exchange rate and Sri Lanka's failure to attract export-oriented foreign investors (and also to retain those who had set up production bases in the country) appear to be the main factors behind the export slowdown. Largely because of the sluggish export performance, the current account deficit widened from 0.5 percent of GDP in 2009 to 7.8 percent in 2010. It has come down since then thanks to the rapid increase in inward remittances by Sri Lankan overseas migrant workers and slower import growth. But, at 3.9 percent of GDP, it is still well above the average level in the emerging market economies (1.8 percent of GDP). 17

Total outstanding external debt almost doubled, from US$20.9 billion (49.7 percent of GDP) in 2009 to US$ 39.7 billion (59.2 percent of GDP) in 2013 (Table 1). More importantly, this increase was underpinned by a palpable shift in the composition of external debt from concessional loans from bilateral donors and international developmental agencies, to borrowings on commercial terms (IMF 2014). Within commercial borrowing, private sector debt—in the form of foreign bank borrowings and international debt securities issued by the state-owned banks under explicit or implicit government guarantees—has increased rapidly. There has also been a rapid growth of short-term debt (foreign capital flows to government securities and banking sector external liabilities), from US$5.3 billion in 2008 to US$ 9.0 billion in 2013. As a result of the overall increase in debt and the shift of its composition from concessional debt to loans on commercial terms, the debt service ratio increased from an average level of 12.5 percent during 2004–08 to 25.3 percent in 2013. The debt service ratio is bound to increase more rapidly in years to come when the accumulated long-term debt begins to mature.

By the end of 2013, total gross foreign-exchange reserves (US$7 billion) were adequate to cover 4.6 months of imports. According to the IMF estimates, however, net foreign exchange reserves18 were only US$ 4.6 billion, which was sufficient to cover the country's import bill for approximately 2.5 months—below the traditional rule of thumb level for reserve adequacy (3 months). However, this import-based reserve adequacy measure, which originated in the days of the Bretton Woods system,19 is no longer an appropriate yardstick for measuring reserve adequacy because Sri Lanka is now significantly integrated into global capital markets through foreign borrowings and its short-term debt exposure has increased significantly in recent years. An important lesson learned from the string of financial crises that engulfed emerging market economies in the 1990s was that the prudent level of reserves needs to be determined in relation to the volume of short-term foreign-currency liabilities (Fischer 2004; Eichengreen 2006; Lee and Park 2009). In terms of this criterion, Sri Lanka's ability to defend the rupee in the event of an external shock that trigger short-term capital outflow has rapidly eroded in recent years. The ratio of net foreign-exchange reserves to total short-term debt20 sharply declined from 2010, reaching 50.9 percent in 2013 (Table 1).

Finally, the other side of the coin of the worsening current account deficit and massive foreign debt accumulation is the widening budget deficit (Table 1). From about the late 1990s until 2008, the budget deficit hovered around 7 percent of GDP, with military expenditure accounting for the lion's share of deficit financing, and reached a historical high of 9.9 percent of GDP at the final stage of the conflict in 2009. Notwithstanding a mild decline during the past three years, the budget deficit still remains well above the internationally considered safety range of 3 percent to 5 percent. Even this mild decline in the officially reported deficit figures needs to be treated with caution because from about 2012 the government has been shifting budgetary transfers to the loss-making public enterprises “off budget,” by forcing these enterprises to borrow on their own from domestic banks under government guarantees.21

## 4.  Regime shift

As discussed in the previous section, by 2014 there were clear signs that the debt-fuelled growth dynamism was not sustainable. Also, despite the glowing headline economic numbers, early signs of popular dissatisfaction with the economic performance, complaints about the increasing cost of living among the general public, and concerns about the sustainability of debt-fuelled growth among the politically vigilant middle class, had begun to emerge. Authoritarianism, nepotism, and corruption of the regime turned out to be part of day-to-day conversations. A clear warning signal of waning electoral support came from the provincial council elections held in the Uva Province, where the government share of the votes slumped from 72 percent to 51 percent. In this context, in November 2014 President Rajapaksa called a snap election for 8 January 2015, two full years ahead of his second 6-year term, anticipating further erosion of his electoral support as economic conditions worsened. He expected to win on the back of his popularity as the hero who ended the civil war and started the post-civil war “economic boom.”22

Rajapaksa's decision to go to a snap election was also premised on the calculation that he would face a weak and splintered opposition, with Ranil Wickremasinghe, long-time leader of the UNP and a many-times unsuccessful presidential candidate, would be his main challenger. This, however, proved to be a major miscalculation. In a series of events that moved at a dizzying pace, the regime found itself facing a united, well-organized, generously financed opposition that launched a disciplined and focused campaign. Mr. Maithripala Sirisena, a Senior Minister and General Secretary of the president's own party (SLFP) in the ruling coalition announced his intention to challenge the President as the common opposition candidate.23 The Sirisena campaign progressively broadened its support base spanning a wide array of disparate political and social groups including the conservative UNP; the left-wing nationalist JVP; the socialist Nava Sama Samaja Party; the Sinhala-Buddhist Jathika Hela Urumaya; the main Tamil party, Tamil National Alliance; the main Muslim party, the Muslim Congress; and a number of “civil society” organizations. The opposition campaign was concentrated on pledges for sweeping changes within 100 days (“100-Day Program”) including pruning the power of executive presidency; forming an all-party (national) government; restoring the independence of institutions such as the police, judiciary, and public services; eliminating corruption; and reducing the cost of living. The promises also included some “sweeteners,” including public sector wage increases.

A month before calling the election, the government introduced a big-spending budget. The state machinery and resources were used in the election campaign in violation of the election law (Bastian 2014; Weerakoon 2015). The state-run media gave virtually exclusive coverage to Rajapaksa's campaign. Samurdi [Prosperity], a welfare payment (cash grant) system for low-income households, was used to woo voters. The Samurdi recipients and newly appointed public-sector employees were dragooned to the election campaign. The military was used to distribute and exhibit government propaganda. The government's election campaign was marred by violence and malpractices of various kinds against the opposition supporters.

But, contrary to expectations, voting proceeded in an unexpectedly calm and orderly fashion on election day. This may have been because of the pressing need for the government to display that the election was free and fair, but the mammoth election rallies (although orchestrated abusing incumbency advantage) and his astrologer's prediction probably also made the president confident that he would easily win without tampering with the voting process. The election commissioner took an unexpectedly hard-line stance on regime shenanigans. As the election turned out to be a close contest during the last few days of the campaign, there was also a “natural” compulsion for officials involved in conducting the election (the returning officers, the police on duty, ballot-box handlers) to follow the procedures correctly (Weerakoon 2015).

The voter turnout was 81.5 percent, by far the highest of the seven elections held under the presidential system since 1982. Rajapaksa's share of the total votes fell from 57.9 percent in 2010 to 47.6 percent in 2015. Sirisena won with a margin of 3.7 percentage points (51.3 percent of the total votes) (Table 2). The anti-Rajapaksa swing was stronger in the northern and eastern provinces and in the Nuwara Eliya district, where ethnic minorities are heavily concentrated (Tables 2 and 3). But there was also a massive swing against Rajapaksa in the predominantly Sinhala-Buddhist hinterland, even in the southern province, the heartland of his support base.

Table 2.
Sri Lanka: Results of the presidential elections, 2010 and 2015
2010 election2015 election
Western province 2,981 78.9 58.6 40.2 3,391 83.2 47.9 51.4
Colombo district 1,161 77.1 52.9 45.9 1,296 82.7 43.4 55.9
Central province 1,362 77.9 52.6 44.8 1,592 82.2 42.9 55.6
Southern province 1,397 79.8 65.1 33.3 1,579 83.6 58.2 40.9
Northern province 284 31.1 25.7 65.0 521 68.5 20.9 75.8
Eastern province 684 69.6 40.1 56.8 810 75.2 26.5 72.0
Northwestern province 1,268 76.3 61.9 36.6 1,445 80.4 52.2 47.0
North central province 673 78.9 65.9 32.5 781 83.4 49.5 49.5
Uva province 676 78.2 58.6 39.3 788 83.3 53.5 45.0
Sabaragamuwa province 1,072 80.1 62.9 35.3 1,217 84.3 54.0 44.8
Sri Lanka 10,397 76.8 57.9 40.1 12,123 81.8 47.6 51.3
2010 election2015 election
Western province 2,981 78.9 58.6 40.2 3,391 83.2 47.9 51.4
Colombo district 1,161 77.1 52.9 45.9 1,296 82.7 43.4 55.9
Central province 1,362 77.9 52.6 44.8 1,592 82.2 42.9 55.6
Southern province 1,397 79.8 65.1 33.3 1,579 83.6 58.2 40.9
Northern province 284 31.1 25.7 65.0 521 68.5 20.9 75.8
Eastern province 684 69.6 40.1 56.8 810 75.2 26.5 72.0
Northwestern province 1,268 76.3 61.9 36.6 1,445 80.4 52.2 47.0
North central province 673 78.9 65.9 32.5 781 83.4 49.5 49.5
Uva province 676 78.2 58.6 39.3 788 83.3 53.5 45.0
Sabaragamuwa province 1,072 80.1 62.9 35.3 1,217 84.3 54.0 44.8
Sri Lanka 10,397 76.8 57.9 40.1 12,123 81.8 47.6 51.3

Source: Compiled for data extracted from Department of Census and Statistics (DCS), Statistical Yearbook 2010 (for data on the 2005 election), and the DCS Web site (www.statistics.gov.lk).

Notes:a. Percentage of valid votes in total registered voters.

MR = Mahinda Rajapaksa; SF = Sarath Fonseka; MS = Maithripala Sirisena.

Table 3.
Sri Lanka: Population, urbanization, and ethnic composition, 2012
Ethnic composition (%)
ProvincePopulation (thousands)Urban population1 (%)SinhaleseSri Lankan TamilIndian TamilMuslimOther
Western province 5,854 38.8 84.1 5.8 1.0 7.9 1.2
Colombo district 2,324 77.5 76.5 10.1 1.0 10.8 1.5
Central province 2,572 10.5 65.9 4.9 18.9 9.9 0.3
Sourthern province 2,477 10.6 95.0 1.0 0.7 2.9 0.4
Northern province 1,062 16.5 2.0 92.9 0.8 2.9 1.4
Eastern province 1,555 25.0 23.2 39.3 0.3 36.9 0.4
Northwestern province 2,380 4.1 85.8 2.8 0.2 11.1 0.2
North Central province 1,267 4.3 90.8 0.9 0.1 8.0 0.2
Uva province 1,266 5.8 80.8 2.4 12.2 4.3 0.2
Sabaragamuwa province 1,929 6.0 86.3 3.7 5.4 4.3 0.3
Sri Lanka 20,359 18.2 74.9 11.1 4.1 9.3 0.5
Ethnic composition (%)
ProvincePopulation (thousands)Urban population1 (%)SinhaleseSri Lankan TamilIndian TamilMuslimOther
Western province 5,854 38.8 84.1 5.8 1.0 7.9 1.2
Colombo district 2,324 77.5 76.5 10.1 1.0 10.8 1.5
Central province 2,572 10.5 65.9 4.9 18.9 9.9 0.3
Sourthern province 2,477 10.6 95.0 1.0 0.7 2.9 0.4
Northern province 1,062 16.5 2.0 92.9 0.8 2.9 1.4
Eastern province 1,555 25.0 23.2 39.3 0.3 36.9 0.4
Northwestern province 2,380 4.1 85.8 2.8 0.2 11.1 0.2
North Central province 1,267 4.3 90.8 0.9 0.1 8.0 0.2
Uva province 1,266 5.8 80.8 2.4 12.2 4.3 0.2
Sabaragamuwa province 1,929 6.0 86.3 3.7 5.4 4.3 0.3
Sri Lanka 20,359 18.2 74.9 11.1 4.1 9.3 0.5

Source: Sri Lanka Department of Census and Statistics, Population and Housing Census 2012, http://www.statistics.gov.lk/PopHouSat/CPH2011/Pages/Activities/Reports/CPH_2012_5Per_Rpt.pdf.

Following the election there was a peaceful transition of power. However, whether this was an “honorable stepping down” by President Rajapaksa or the outcome of the system's checks and balance remains a debatable issue. According to some media commentaries, he explored the possibility of declaring the election results null and void, but failed because the attorney general, the solicitor general, the inspector general of police, and the military commanders were not supportive (The Economist2015).

### 4.1  What caused the regime change?

How does Sri Lanka's seemingly impossible regime change fit within the existing body of knowledge on regime changes? The available theoretical and empirical literature on this subject has specifically focused on transition from authoritarian to democratic (or hybrid) rule and/or, in a few cases, reversion from democracy to dictatorship.24 The downfall of the Rajapaksa regime does not fit neatly within this framework. In spite of the significant dictatorial tendencies developed over time, it was not a pure authoritarian regime (dictatorship). Rather it was an “illiberal democracy,” which still depended on electoral support for survival. This was, in fact, the reason why president Rajapaksa went for a snap election before the people's memories of the civil war victory faded and the debt-fuelled economic boom dissipated. The regime shift, therefore, involved two separate, but closely related, episodes: Sirisena's decision to contest the incumbent president in the election and the expression at the ballot box of the voters’ discontent that had been simmering beneath the surface. Sirisena's defection from the government that triggered the regime change was a clear case of an “elite rupture,” a well-calculated reaction to the increased concentration of power in the presidency, which alienated the president (and his close associates) within the ruling party itself (Langston 2006).

There has been considerable speculation, and direct allegations by the ex-president and others in his camp,25 that western countries and India played a role in orchestrating the elite rupture. It is certainly the case that Sri Lanka, because of its strategic location, is of significant interest to the major powers in the balance of power game in the Indian Ocean (Scott 2008; Kaplan 2009). As a reaction to the deterioration of relations with western powers and India on the issue of allegation of war crimes during the final phase of the civil war, the Rajapaksa regime had embraced China as Sri Lanka's “first friend.” Whatever role foreign powers may have played, however, our analysis of the regime consolidation in the previous section suggests that the elite rupture would not have happened without the disenchantment within the government itself caused by one-family authoritarian tendencies.

We estimated the following two simple regression equations to examine change in voter behavior between the 2010 and 2015 presidential elections:
1
2
where MRM the difference between percentage share of Mahinda Rajapaksa's votes in total votes polled and that of the major contestant (Mahinda Rajapaksa Margin); ETH represents the share of all ethnic minorities (SLT + INT + MSL) in the total population in the district (percent); SLT the share of Sri Lankan Tamils in the total population is in the district (percent); INT the share of Indian Tamils in the total population is in the district (percent); MSL the share of Muslims in the total population is in the district (percent); URB is the share of urban population in the total population in the district (percent); TNV represents the change in voter turnout in the election in the given year over the previous election (percent); and D15 is binary dummy variable, which takes value 1 for 2015 and 0 for 2010. The standard constant term is denoted by α, δ captures the joint impact of unobserved explanatory variables (unobserved effects); μ is the disturbance term that is assumed to satisfy the usual regression model conditions; i = 1, 2 , … , 22 is the district; and t = 2010, 2015 denotes the two elections.

In equation (1), ETH captures the swing votes of the three ethnic memories lumped together (non-Sinhalese) between the two candidates. The three ethnic groups (SLT, INT, and MSL) are included separately in equation (2) to test the possible difference in their voting patterns. The other two explanatory variables are common to both equations. URB is expected to capture the combined impact of voters’ discontent caused by the increase in the cost of living and of their concern about the deterioration of the quality of governance. People living in urban areas are more affected by increases in cost of living compared with their rural counterparts. The middle class, which is presumably more concerned about corruption and other malpractices and also better aware of the fragility of the economic boom, is heavily concentrated in urban areas. Unfortunately, the available data do not permit us to specify suitable variables to capture the two effects separately. TNV could capture the possible impact on the election outcome of the “last-minute” turnout of voters, who until then remained “discouraged” because of the perception of impossibility of regime change and activated as the opportunity for a change became clearer. The intercept dummy variable, D15, and its interaction terms with the other explanatory variables (slope dummies) are included to test whether the voting patterns in 2015 are different from the “average” pattern for the two elections. This approach is equivalent to estimating two separate regressions for the two elections, but has the added advantage of providing a direct test of the statistical significance of the differences between the estimated coefficients (Dougherty 2007).

We estimated the two equations using a two-dimensional (year and district) panel data set.26 The pooled ordinary least-squares and the random effects estimators were used as alternative estimation methods.27 In terms of the Breush-Pagan test for the presence of unobserved heterogeneity across the districts of the election results, we were not able to reject the null hypothesis of no difference between the two estimators of both equation (p-value = 1.00). We therefore opted for pooled ordinary least-squares on efficiency grounds.

The results are reported in Table 4. In terms of the adjusted-R2, both equations explain over 80 percent of the variation in inter-district differences in the voting patterns. The F-test for the joint significances of D15 and its interaction terms with the three explanatory variables (reported in the last row of Table 4) overwhelmingly support the hypothesis that the voting patterns in the 2015 elections are significantly different from those of the 2010 election.

Table 4.
Determinants of Mahindra Rajapaksa's margin (MRM): Presidential elections 2010 and 2015
Explanatory variableEquation (1)Equation (2)
Constant 9.13*** 9.65***
(1.16) (1.34)
ETH: The three minor ethnic groups, % −0.13***
(0.02)
SLT: Sri Lankan Tamils, %  −0.13***
(0.01)
INDT: Indian Tamils, %  −0.21***
(0.04)
MSL: Muslims, %  −0.17***
(0.073)
URB: Urbanization, % −0.01 −0.11**
(0.02) (0.05)
TNV: Change in turnout from previous election, % 0.01 0.02
(0.04) (0.0.3)
D15: 2015 election dummy 8.55** 14.68***
(3.45) (5.85)
D15*ETH −0.56***
(0.12)
D15*SLT  −0.23
(0.16)
D15*INT  −0.53***
(0.16)
D15*MSL  −1.02**
(0.42)
D15*URB −0.24* −0.23**
(0.13) (0.10)
D15*TNV −0.06 −1.09**
(0.34) (0.53)
Number of districts 22 22
Number of observations 44 44
Adjusted R2 0.83 0.85
Breusch-Pagan LM test for (p-value)a 1.00 1.00
Test for joint significance of D15 and its interaction 0.00 0.00
terms with other variables (p-value)b
Explanatory variableEquation (1)Equation (2)
Constant 9.13*** 9.65***
(1.16) (1.34)
ETH: The three minor ethnic groups, % −0.13***
(0.02)
SLT: Sri Lankan Tamils, %  −0.13***
(0.01)
INDT: Indian Tamils, %  −0.21***
(0.04)
MSL: Muslims, %  −0.17***
(0.073)
URB: Urbanization, % −0.01 −0.11**
(0.02) (0.05)
TNV: Change in turnout from previous election, % 0.01 0.02
(0.04) (0.0.3)
D15: 2015 election dummy 8.55** 14.68***
(3.45) (5.85)
D15*ETH −0.56***
(0.12)
D15*SLT  −0.23
(0.16)
D15*INT  −0.53***
(0.16)
D15*MSL  −1.02**
(0.42)
D15*URB −0.24* −0.23**
(0.13) (0.10)
D15*TNV −0.06 −1.09**
(0.34) (0.53)
Number of districts 22 22
Number of observations 44 44
Adjusted R2 0.83 0.85
Breusch-Pagan LM test for (p-value)a 1.00 1.00
Test for joint significance of D15 and its interaction 0.00 0.00
terms with other variables (p-value)b

Source: Estimated using the relevant data series given inTables 2,and3.

Note: The dependent variable (MRM) is the difference between percentage share of Mahinda Rajapaksa's votes in total votes polled and that of the major contestant. Heteroscedasticity-collected standard errors (clustered by districts) are reported in parenthesis, with the statistical significance of the regression coefficient denoted as ***statistically significant at the 1 percent level, **statistically significant at the 5 percent level, and *statistically significant at the 10 percent level.

a. A test for the choice of pool OLS estimator over random effects estimator.

b. A test of the statistical difference of the regression coefficients between 2010 and 2015 elections.

The results confirm the important role of ethnicity in determining the swing against Rajapaksa in the 2015 election. In the 2010 election, a percentage point change in the combined share of the three minor ethnic groups in total population among the districts contributed to a 0.13 percentage point reduction is Rajapaksa's margin (equation (1)). In the 2015 election, the anti-Rajapaksa swing in the ethnic vote was as high as 0.69 percentage points. When the ethnic population is disaggregated into the three ethnic groups (SLT, INDT, MSL) the coefficients of the three variables do not overlap with each other, suggesting that the voting patterns among the three groups are significantly different. For this reason, we consider Equation 2 as our preferred equation in the following discussion.

Interestingly, the coefficient of the dummy interaction variable for the Sri Lankan Tamils (D15*SLT) is not statistically significant. This suggests that there was no significant difference in the anti-Rajapaksa swing in the Sri Lankan Tamil votes in the two elections. The anti-Rajapaksa swing in the 2015 election came solely from the other two ethnic minorities, in particular from the Muslim voters. When controlling for the other variables, there was an almost one-to-one correspondence in 2015 between the degree of swing against Rajapaksa and inter-district differences in the share of Muslim population (–1.19 = –0.17 – 1.02, with a standard error of 0.43), up from 0.17 in 2010. One percentage point change in the share of the Indian Tamil population accounted for a 0.74 percentage point reduction in Rajapaksa's margin, compared with 0.21 in 2010.

Urbanization emerges as a much more important factor in 2015 compared with 2010. In 2015 a one percentage point change in the degree of urbanization across districts is associated with 0.23 percentage point of the swing against Rajapaksa, compared with 0.11 percent in 2010. As discussed earlier, this pro-opposition shift in the urban votes can be reasonably attributed to a combination of the relatively higher cost of living in urban areas and, perhaps more importantly, the fact that a higher share of the population in urban areas are presumably more concerned about the authoritarian tendencies of the regime.

The latter inference relating to the voting patterns of the urban population is consistent with the available evidence that shows sharp changes in the voting preferences of postal voters between the two elections (Gunasekera 2015). In 2010 Rajapaksa won the postal votes by a higher margin of general votes. In the 2015, he was defeated in postal votes as well; interestingly, his opponent's winning margin turned out to be larger in postal votes compared with that of general votes. Postal voters are almost entirely police and military personnel and government employees engaged in election duties and essential services, who presumably represent significant sections among the more politically sophisticated voters. Ironically, in the regime's view, they owed their allegiance to the president: As discussed, over-staffing government offices through new recruitment and increasing public service salaries were an integral part of the government's preparation for the election. Though exact figures are not available, the absolute majority of postal voters are Sinhalese (Gunasekara 2015).

In the 2015 election, unlike in 2010, the coefficient of the change in voter turnout variable (1.09) is statistically significant at the 5 percent level with a negative sign. This suggests that there was an almost one-to-one relationship between reduction in Rajapaksa's margin and the increase in voter turnout across the districts, notwithstanding the funding he lavished on the election campaign. Finally, the intercept dummy variable (D15) carries a coefficient of 14.5, which is statistically highly significant. This suggests that if it were not for the swing against him of ethnic minorities and urban voters, and the increase in turnout that overwhelmingly went against him, Rajapaksa's margin over his opponent would have increased by 14.5 percentage points at the 2015 election over the previous election.

## 5.  Prospects and challenges

The “elite rupture”—Maithreepala Sirisena's defection from the ruling party to contest the incumbent president as the common candidate—was only the proximate cause of Sri Lanka's stunning regime shift in January 2015. It provided the voters with an unanticipated opportunity to express their discontent with the Rajapaksa regime, which had been simmering beneath the surface, at the ballot box. The overwhelming support received by Sirisena from the Tamil and Muslim communities, despite his own past as a senior member of the Rajapaksa government and the presence of Sinhala-chauvinist forces in his ranks, played an important role in the regime shift. Most importantly, however, the regime change reflected an underlying assertiveness on the part of the wider electorate, encompassing all ethnic groups, of its democratic rights. For the first time in Sri Lanka's post-independence history, the minority communities appear to have felt themselves part of, and potentially an important influence on, the national political scene. But sustaining their faith and ensuring ethnic harmony, although achieving an economic transition from the debt-driven bubble economy to sustainable growth, is bound to be a formidable task for the new regime. The country is now run by a motley coalition, by far the most diverse political alliance in Sri Lankan history, which does not have a clear majority in the parliament. Ideological differences make it difficult for agreement on policy, though war-weariness and the discontent with the Rajapaksa regime may have led to a greater willingness to accommodate diverse perspectives and demands within the political system, which may also help the reconciliation process.

The new government has taken some early steps to restore institutions and the rule of law. Restrictions on the media have been removed, militarization of civilian spaces has decreased, and the judiciary seems to have become more assertive and independent. A start was also made to shift power from the presidency to the prime minister and strengthening parliamentary oversight through an amendment to the constitution (the 19th Amendment) (Fernando 2015). Although it falls far short of the election promise to abolish the presidential system, this amendment does provide for transferring many of the executive powers to the cabinet, which is directly responsible to the parliament. The term of the parliament and the president is shortened from 6 years to 5 years, the president cannot dissolve the parliament within 4.5 years, and the president is eligible to contest only for a second term.

The hardest problem faced by the new regime is to find a solution to the ethnic conflict that would enable the military to be withdrawn from the Tamil majority areas and restore civil rule with an acceptable degree of administrative autonomy. It has already begun to release some land acquired by the military in the northern province to the original owners. The anti-minority hysteria has also receded thanks to initiatives to restore law and order and putting an end to the militarization of civilian spaces. However, whether the regime shift will provide a window of opportunity for working towards national reconciliation is yet to be seen.

On the economic front, the challenges facing the new regime are quite daunting. The biggest challenges arise from the legacy of the debt-financed economic boom, driven to a large extent by Chinese-funded public sector investments. Though this boom was intrinsically unsustainable, it was also the basis of a sense of growing prosperity. The unfortunate reality is that the economic performance of the new government will be judged by the electorate with the overall growth performance during the Rajapaksa period as a benchmark. The policy dilemma of the new regime is, therefore, how to redirect policy changes to restore the international competitiveness of the economy, and to contain debt dependency while maintaining the living standards of the population and an adequate growth momentum to sustain employment levels. A lower growth rate and a slide in living standards will produce an electoral backlash that may well reverse the political transition or generate high political volatility and instability.

Here it is important to emphasize the fundamental change that has occurred in the nature of the Sri Lankan economy during the Rajapaksa regime. The large and persistent current account deficits and the shift to commercial debt financing have exposed the economy to volatile capital movements in a way that is unprecedented. In an attempt to stabilize its own direct foreign borrowings under pressure from the IMF, the government relaxed previous controls, allowing domestic banks, state-owned enterprises, and private corporations to engage in foreign borrowings, and then borrowed from these “domestic” sources. As a result, the state controlled banks have substantially increased their foreign currency exposures, which carry an implicit government guarantee and in effect create a contingent public sector liability. This makes the Sri Lankan economy quite similar in important ways to the Southeast Asian economies such as Thailand, Malaysia, and Indonesia that were hit by the 1997–98 Asian financial crisis (Athukorala and Warr 2002). It remains much more vulnerable to sudden adverse shocks from the global economy than in the past, particularly in the context of increased fragility in the emerging markets and continuing poor conditions in key markets, including the EU.

The new government's foreign policy initiatives to restoring relationships with the Western powers also pose problems, particularly in the short run, for managing the transition from a bubble economy to a sustainable growth path. The political pressures to reduce the economic dependency on China—seen as the driver of the political and strategic tilt toward China—is likely to lower the flow of future Chinese financial assistance and investment, even if current projects and commitments continue. Although many public sector investment projects funded by the Chinese are widely considered to be inefficient “white elephants”, it is not clear that compensating capital and financial flows will flow from the West or India to maintain current levels of investment and economic activity.

Even under an optimistic scenario of availability of external financing from alternative sources, the debt-driven growth dynamism can be sustained only as long as foreigners keep lending to Sri Lanka and increasing their total Sri Lankan exposure. Eventually sustaining robust growth requires macroeconomic adjustments to restore the international competitiveness of the economy through depreciation of the real exchange rate. Realistically, this will require a substantial nominal depreciation of the rupee. Under current economic conditions, however, relying on nominal exchange rate depreciation alone for achieving this economic adjustment could be a recipe for economic disaster. Given the massive build-up of foreign currency–denominated government debt, exchange rate depreciation naturally worsens budgetary woes. And given the increased exposure of the economy to global capital markets (as discussed) a large abrupt change in the exchange rate could also shatter investor confidence, triggering capital outflows. Therefore, what is required is a comprehensive policy package encompassing greater exchange rate flexibility and fiscal consolidation (which requires both rationalization of expenditure and widening the revenue base) to achieve a durable reduction in public debt, and complementary measures, including trade and investment policy reforms, to improve the overall investment climate in the country.

The current government, a coalition of disparate groups with divergent ideologies, is unlikely to effectively implement major policy changes. It is difficult at this stage to predict the composition and economic policy stance of the regime that will emerge from the upcoming parliamentary election, but it will certainly need strong economic and political skills to successfully achieve what the Sri Lankan electorate hopes for and deserves. The bottom line is that governing multi-ethnic states require special political ingenuity and care: As Sri Lankan's post-independence political and economic experiences have vividly demonstrated, maintaining ethnic harmony and sustaining growth momentum of the economy are intricately interrelated.

## References

Aneez
,
Shihar
, and
Ranga
Sirilal
.
2014
.
Sri Lanka Statistics Office to Sack Official for Data Manipulation Claim
. http://in.reuters.com/article/2014/01/22/srilanka-economy-data-idINL3N0KW3EL20140122.
Accessed June 2014
.
Arunatilake
,
Nisha
,
P.
Jayawardena
, and
Dushni
Weerakoon
.
2011
.
Sri Lanka
. In:
Migration, Remittances and Development in South Asia
, edited by
Saman
Kelegama
, pp.
112
140
.
New Delhi
:
Sage
.
Athukorala
,
Prema-chandra
.
2012
.
Sri Lanka's Trade Policy: Reverting to Dirigisme
?
The World Economy
35
(
12
):
1662
1686
.
Athukorala
,
Prema-chandra
, and
Sisira
Jayasuriya
.
1994
.
Macroeconomic Policies, Crises and Growth in Sri Lanka, 1969–90
.
Washington DC
:
World Bank
.
Athukorala
,
Prema-chandra
, and
Sisira
Jayasuriya
.
2013
.
Economic Policy Shifts in Sri Lanka: The Post-conflict Development Challenge
.
Asian Economic Papers
12
(
2
):
1
28
.
Athukorala
,
Prema-chandra
, and
Peter G.
Warr
.
2002
.
Vulnerability to a Currency Crisis: Lessons from the Asian Experience
.
The World Economy
25
(
1
):
33
57
.
Bandarage
,
Asoka
.
2009
.
The Separatist Conflict in Sri Lanka: Terrorism, Ethnicity, Political Economy
.
London
:
Routledge
.
Bastian
,
Dharisha
.
2014
.
The Agony of Aluthgama
.
Daily Financial Times
(Colombo), 17 June
.
Biziouras
,
Nikolaos
.
2014
.
The Political Economy of Ethnic Conflict in Sri Lanka: Economic Liberalization, Mobilizational Resources, and Ethnic Collective Action
.
London
:
Routledge
.
CBSL (Central Bank of Sri Lanka)
.
2013
.
Annual Report 2012
.
Colombo
:
SBSL
.
Collier
,
Paul
,
Martina
Kirchberger
, and
Mans
Soderbom
.
2013
.
The Cost of Road Construction in Developing Countries
. Centre for Study of African Economies Working Papers.
Oxford
:
University of Oxford
.
Cronin-Furman
,
Kate
.
2015
.
Sri Lanka's Surprise Political Transition
.
Washington Post
, 12
January
.
De Silva
,
Chandra R.
1978
.
The Politics of University Admissions: A Review of Some Aspects of University Admission Policy in Sri Lanka, 1971–1978
.
The Sri Lankan Journal of Social Sciences
1
(
2
):
85
123
.
De Silva
,
Kingsley M.
1986
.
Managing Ethnic Tensions in Multi-Ethnic Societies: Sri Lanka 1880–1985
.
New York
:
University Press of America
.
De Silva
,
Kingsley M.
2012
.
Sri Lanka and the Defeat of the Tigers
.
Delhi
:
Penguin Books India
.
Dougherty
,
Christopher
.
2007
.
Introduction to Econometrics, 3rd edition
.
Oxford
:
Oxford University Press
.
Economist
.
2015
.
Sri Lanka's Election. 17 January, pp
. 10,
27
28
.
Eichengreen, Barry
.
2006
.
The Next Tsunami? Preparing for a Disorderly Correction of Global Imbalances
.
Asian Economic Papers
5
(
2
):
1
6
.
Fernando
,
Laksiri
.
2015
.
A Partial, But a Great Victory for Democracy
.
Sri Lanka Guardian
(Colombo), 30 April
.
Fischer
,
Stanley
.
2004
.
IMF Essays from a Time of Crisis
.
Cambridge, MA
:
MIT Press
.
Goodhand
,
Jonathan
.
2012
.
Sri Lanka in 2011: Consolidation and Militarization of the Post-War Regime
.
Asian Survey
52
(
1
):
130
137
.
Gunasekara
,
Tisarane
.
2015
.
Rajapaksa Beats Racist Tom Tom: What the Statistics Really Show
.
Sri Lanka Guardian
, 15
January
.
Haggard
,
Stephan
, and
Robert R.
Kaufman
.
2012
.
Inequality and Regime Change: Democratic Transition and the Stability of Democratic Rule
.
American Political Science Review
106
(
3
):
495
515
.
Holt
,
John Clifford
, editor.
2011
.
The Sri Lanka Reader: History, Culture, Politics
.
Durham, NC
:
Duke University Press
.
IBA (International Bar Association)
.
2013
.
A Crisis of Legitimacy: The Impeachment of Chief Justice Bandaranayake and the Erosion of the Rule of Law in Sri Lanka
.
London
:
IBA
.
IMF (International Monetary Fund)
.
2011
.
Sri Lanka: Second and Third Review Under the Stand-By Agreement
.
Washington, DC
:
International Monetary Fund
.
IMF (International Monetary Fund)
.
2013
.
Sri Lanka: First Post Programme Monitoring Discussion
.
Washington, DC
:
IMF
.
IMF (International Monetary Fund)
.
2014
.
Article IV Consultation and Second Post Programme Monitoring Discussion: Staff Report
.
Washington, DC
:
IMF
.
Jayasuriya
,
Laksiri
.
2013
.
The Hybrid Regime in Post-Civil War Sri Lanka
.
International Studies
49
(
3
):
437
447
.
Jeanne
,
O.
, and
R.
Ranciere
.
2011
.
The Optimal Level of International Reserves for Emerging Market Countries: A New Formula and Some Applications
.
The Economic Journal
121
(
555
):
905
930
.
Jennings
,
Ivor
.
1954
.
Politics in Ceylon since 1952
.
Pacific Affairs
27
(
4
):
338
352
.
Kaplan
,
Robert D.
2009
.
Center Stage for the Twenty-First Century: Power Plays in the Indian Ocean
.
Foreign Affairs
March/April
:
16
32
.
Kumarage
,
Amal
.
2014
.
The Real Cost of Highway Development: Who Has Got the Numbers Right?
Sunday Times
,
21
December
.
Kuran
,
Tmur
.
1991
.
Now Out of Never: The Element of Surprise in the Eastern European Revolution of 1989
.
World Politics
44
(
1
):
7
48
.
Langston
,
Joy
.
2006
.
Elite Rupture: When do Ruling Parties Split? In
:
Electoral Authoritarianism: The Dynamics of Unfree Competition
, edited by
Andreas
Schedler
, pp.
57
76
.
Boulder, CO
:
Lynne Rienner
.
Law & Society Trust
.
2012
.
Mega Development in Hambantota
.
Colombo
:
Law & Society Trust
.
Lee
,
Jong-Wha
, and
Cyn Chong
Park
.
2009
.
Global Financial Turmoil: Impact and Challenges for Asia's Financial Systems
.
Asian Economic Papers
8
(
1
):
9
40
.
Kenny
,
Charles
.
2009
.
Transport Construction, Corruption and Developing Countries
.
Transport Review
29
(
1
):
21
41
.
Manor
,
James
.
1989
.
The Expedient Utopian: Bandaranaike and Ceylon
.
Cambridge
:
Cambridge University Press
.
Ministry of Finance, Sri Lanka
.
2013
.
Annual Report 2012
.
Colombo
:
Ministry of Finance
.
Rajapatirana
,
Sarath
.
2013
.
Export Growth and Appropriate Macroeconomic Policy: A Necessary Nexus
. In:
Way Forward for Sustained Growth
, edited by
A. D. V
.
De S. Indraratna
,
S.
Vidanagama
, and
A.
Wijesinghe
, pp.
16
71
.
Colombo
:
Sri Lanka Economic Association
.
Ranaweera
,
Lanka
,
Supannee
Kaewsutthi
,
Aung Win
Tun
,
Hathaichanoke
Boonyarit
,
Samerchai
Poolsuwan
, and
Patcharee
Lertrit
.
2014
.
Mitochondrial DNA History of Sri Lankan Ethnic People: Their Relations within the Island and with the Indian Subcontinental Populations
Journal of Human Genetics
59
(
1
):
28
36
.
Roberts
,
Michael
.
1986
.
Ethnic Conflict in Sri Lanka and Sinhalese Perspectives: Barriers to Accommodation
.
Modern Asian Studies
12
:
353
376
.
Sarvananthan
,
Muttukrishna
.
2015
.
Elusive Economic Peace Dividend in Sri Lanka: All That Glitter Is Not Gold
.
Geojournal
(forthcoming, published online: 13 March 2015)
.
Scott
,
David
.
2008
.
The Great Power “Great Game” Between India and China: The Logic of Geography
.
Geopolitics
13
(
1
):
1
26
.
Shastri
,
Amita
.
1997
.
Government Policy and the Ethnic Crisis in Sri Lanka
. In:
Government Policies and Ethnic Relations in Asia and the Pacific
, edited by
Michael E.
Brown
and
Sumit
Ganguli
, pp.
129
163
.
Cambridge, MA
:
MIT Press
.
Shourie
,
Arun
.
1974
.
Sri Lanka's National Accounts: Notes for Regression Enthusiasts
.
Marga Quarterly Journal
2
(
3
):
21
42
.
SLDCS (Sri Lanka Department of Census and Statistics).
2015
.
Household Income and Expenditure Survey 2012/13
.
Colombo
:
SLDCS
.
Snodgrass
,
Ronald R.
1999
.
The Economic Development of Sri Lanka: A Tale of Missed Opportunities
. In:
Creating Peace in Sri Lanka: Civil War & Reconciliation
, edited by
Robert I.
Rotberg
, pp.
89
108
.
Washington, DC
:
Brookings Institution Press
.
Tambiah
,
Stanley J.
1997
.
Levelling Crowds: Ethnonationalistic Conflicts and Collative Violence in South Asia
.
Berkeley
:
University of California Press
.
Tanzi
,
Vito
, and
Hamid
Davoodi
.
1998
.
Corruption, Public Investment and Growth
. In:
The Welfare State, Public Investment and Growth
, edited by
Horifumi
Shibata
and
Toshihiro
Idori
, pp.
41
60
.
Berlin
:
Springer
.
Teorell
,
Jan
.
2010
.
Determinants of Democratization: Explaining Regime Change in the World
,
1972
2006
.
Cambridge
:
Cambridge University Press
.
Transparency International
.
2014
.
Sri Lanka National Integrity System Assessment 2014.
http://www.transparency.org/whatwedo/publication/sri_lanka_national_integrity_system_assessment_2014.
Accessed June 2015
.
Wayland
,
Sarah
.
2004
.
Ethnonationalist Networks and Transnational Opportunities: The Sri Lankan Tamil Diaspora
.
Review of International Studies
30
(
3
):
405
426
.
Weerakoon
,
.
2015
.
Sri Lanka: The Incumbent Has No Hope of Winning Legitimately
.
Sri Lanka Guardian
,
January 7
.
,
Laurence
.
1990
.
Political Explanations of Macroeconomic Management: A Survey
.
World Development
18
(
8
):
1133
1146
.
Wriggins
,
Howard
.
2011
.
After Forty Years
. In:
The Sri Lanka Reader: History, Culture, Politics
, edited by
John
Clifford Holt
, pp.
607
617
.
Durham, NC
:
Duke University Press
.

## Notes

*

We gratefully acknowledge valuable comments received from the three discussants, Gamini Herath, Fredrik Sjöholm, and Siow Yue Chia, and other participants at the Asian Economics Panel Meeting held at Sunway University, Kuala Lumpur, 23–24 March 2015, and from Zahabia Adamaly, Sarath Rajapatirana, and Michael Roberts. Thanks are also due to Fahad Khan for his help with the econometrics.

1

There is historical and archeological evidence that the ethnic identities of the Sinhalese and the Sri Lankan Tamils evolved over time through the intermingling of many waves of migrants from different parts of India. According a recent study of genetic relations among the main ethnic groups in Sri Lanka, there is no clear genetic separation between these two ethnic groups (Ranaweera et al. 2014).

3

The five members who voted in favor lost the seats they contested at the first election (1948) held under the new constitution (Shastri 1997).

4

Under agreements between India and Sri Lanka signed in 1964 and 1974, about three-fifths of the Indian Tamil population was repatriated to India and others were given citizenship in the 1970s and 1980s.

5

Discontent within his own ranks about his indecisiveness in delivering the “Sinhala-only” and “favored status of Buddhism” promises resulted in Bandaranaike's assassination in 1959 by a Buddhist monk. In an authoritative biography of Bandaranaike, James Manor (1989, 320) concludes, “The riots … [and] the polarization of society along linguistic lines … virtually guaranteed that his bold move to the left would not have led nowhere. … [T]he assassination has obscured the wasted opportunities of his premiership by distracting attention from the disintegration of the political experiment that he had undertaken as Prime Minister. His murder and the legend which grew up around him therefore did more to revive the fortunes of that experiment than could have anything that he might have done had he survived.”

6

It is important to note that social and political tensions in post-independence Sri Lanka were not confined to the Tamil youth in the north; there were two traumatic episodes of uprising of the Sinhalese youth in the south in 1971 and in 1998–99 (Bandarage 2009).

7

For the first time in Sri Lankan history, the change of government in 1994 did not result in a major change in economic policy. In a significant departure from its past history of adherence to state-led closed economy policies, the new People's Alliance government continued and even built on the pro-market policies of the previous regime. Despite the continuing civil war, by this time the gains from reforms had been impressive enough to win bipartisan support.

8

The manner in which the final stage of the war was fought has subsequently generated international concern over war crimes and investigations by UN bodies.

9

At the presidential election held on 26 January 2010 Rajapaksa received 57 percent of total votes (17 percent margin over the main opponent); at the parliamentary election held on 20 April 2010 the United People's Freedom Alliance led by Rajapaksa won 144 of the 225 seats.

10

Total employment in the public sector increased from around 900,000 (10.4 percent of the total labor force) in 2005 to over 1.2 million (14 percent) in 2012 (Ministry of Finance 2013).

11

This was dramatically demonstrated in the murder of leading journalist Lasantha Wickrematunge in January 2009 (Holt 2011). Following his murder, the “independent” media in the country began to self-censor. The editorial written by Wickrematunge (entitled “And They Came for Me”), with instruction for it to be published in the event of his murder, has been reproduced in Holt (2011, 715–20).

12

See http://rsf.org.

13

The per-kilometer cost of constructing highways in Sri Lanka in 2005 appeared to match the global levels. But during more recent years, the cost in Sri Lanka was 5 to 15 times higher than even the upper-bound estimators reported in global studies.

14

Interestingly, all these non-bidding projects (except one—a US\$ 845 million project awarded to a local contractor) were funded from Chinese sources and awarded to Chinese contractors.

15

In January 2014, the DCS sacked the Acting Director of its National Accounts Department after he revealed that the CDS, in his absence, revised up the 2013 first quarter growth rate from 5.4 percent to 6 percent (Aneez and Sirilal 2014).

16

The only notable exception has been the export-oriented ready-made garment industry, which had already been well integrated within the global apparel value chain as a producer of upmarket apparel products (lingerie and fashion casual wear) thanks to trade-cum-investment liberalization reforms in the 1980s and 1990s.

17

The figure is from the IMF World Economic Outlook database.

18

Gross official reserves net of short-term (less than one year) contractual payment obligations and foreign exchange swap arrangements with domestic banks.

19

Under the Bretton Woods system of fixed exchange rates combined with binding controls on capital flows, the worst situation that could be imagined relating to balance of payments management of a country was that it could lose access to trade credit, which normally matures in three months (Athukorala and Warr 2002).

20

This ratio is the single most empirically supported indicator of a country's vulnerability to currency crises (Jeanne and Ranciere 2011).

21

Total losses of public corporations increased from Rs 8 billion (0.3 of GDP) in 2005 to Rs 191 billion (2.5 percent of GDP) in 2012 (Ministry of Finance 2013).

22

Another seemingly important reason was that the President's trusted astrologer had predicted indubitable victory if the elections was held on the specified date! (Economist2015)

23

Apparently Sirisena's candidacy was orchestrated using satellite phones to escape the Rajapaksa regime's oppressive surveillance (Cronin-Furman 2015).

24

See Kuran (1991), Teorell (2010), Haggard and Kaufman (2012), and the works cited therein.

26

The data are compiled from the sources listed in Tables 2 and 3.

27

The fixed-effects estimator is not applicable because the data series on the ethnicity variables are time-invariant.