Journal Issue
Share
Article

Sri Lanka: Ex Post Evaluation of Exceptional Access Under the 2009 Stand-By Arrangement

Author(s):
International Monetary Fund. Asia and Pacific Dept
Published Date:
September 2014
Share
  • ShareShare
Show Summary Details

Introduction and Background

1. The completion of Sri Lanka’s Stand-By Arrangement (SBA), approved in July 2009 affords a timely opportunity to perform an Ex Post Evaluation (EPE). Following two extensions, the SBA was successfully completed in July 2012, marking the longest engagement with the largest single facility that Sri Lanka has had with the Fund. Program discussions were initiated in early 2009 just prior to the ending of the country’s decades-long civil conflict. They were juxtaposed against the exceptionally fragile global environment in the aftermath of the collapse of Lehman Brothers, and the onset of the global financial crisis. This particular combination of circumstances—a reintegrating domestic economy amid a “once in a century” global financial crisis—made this program subject to more economic uncertainty than usual. The overall assessment of the program contained in this EPE, therefore, attempts to take these complexities into account.

2. This report is structured as follows. The first section provides the background and context for the program. The following section describes the broad strategy of the program, as it was initially envisaged. Section III describes program outcomes. Section IV examines some central questions with respect to the program’s successes, shortcomings, and its consistency with exceptional access policy. The last section concludes, discussing possible lessons for the future. The views of the Sri Lankan authorities regarding the program can be found in the Annex.

A. Build-up of imbalances

3. Leading up to its crisis, Sri Lanka’s economy was overheating, with manifest external and fiscal vulnerabilities. Fueled by strong private domestic demand, and increasing government spending, the economy grew above trend from 2005–08. Annual growth averaged 6.7 percent in this period, with capacity utilization showing strains on the supply side. Inflation (average) shot up from 11 percent to 22.6 percent over the same period. The current account deficit ratcheted up from 2.7 to nearly 10 percent of GDP, while the overall fiscal deficit hovered around or over 7 percent of GDP. From 2004–07, the external financing component rose from 2.4 to 3.7 percent of GDP, much of it on short maturity. The stock of public debt doubled between 2000 and 2008, and despite its high growth rates, Sri Lanka’s debt-to-GDP ratio was high relative to other emerging economies.

Selected Vulnerability Indicators(In percent of GDP, unless otherwise noted)
Sri Lanka

(2008)
Median, Sample of 48 Emerging

Economies (end-2007)
AllNon-Investment

Grade 1/
External sector
Gross reserves (in percent of short-term debt)31154.3221.2
Total gross external debt44.940.139
Current account balance−9.8−3.7−5
Gross external financing requirement 2/8.714.311.2
Public sector
Overall balance−7−1.5−1.3
Public sector gross debt81.13942
Of which: Exposed to exchange rate risk 3/46.71622.3
Source: IMF staff estimates.

Based on conformity with at least two major rating agencies—Moody’s, Standard &Poor’s, and Fitch.

Current account balance plus maturing debt.

Debt in foreign currency or linked to the exchange rate, domestic and external.

Source: IMF staff estimates.

Based on conformity with at least two major rating agencies—Moody’s, Standard &Poor’s, and Fitch.

Current account balance plus maturing debt.

Debt in foreign currency or linked to the exchange rate, domestic and external.

4. Despite a decision to float the rupee years earlier, de facto, the exchange rate was still tightly managed. In January 2001, the authorities made the decision to move the rupee to a floating regime. Yet, as reflected in several Article IV reports (e.g., 2008, 2007, 2005), as well as in past program documents (ECF initiated in 2003), the commitment to maintaining two-way flexibility of the exchange rate wavered. Coming into the 2009 SBA, the exchange rate had once again been tightly managed. High domestic interest rates and the de facto pegged exchange rate policy attracted “carry trade”-related portfolio flows into Sri Lanka’s fixed income market, as was prevalent in other emerging economies, particularly from 2003 onward. At the same time, rapidly rising inflation was undercutting competitiveness, leading to an estimated rupee overvaluation between 7 and 16 percent in real effective terms by 2008.

5. Fiscal imbalances built as revenues were on a long-term declining trend. Revenues declined by some 10 percentage points of GDP between the 1980s and the time the 2009 SBA was negotiated. Broad underlying factors, such as gradual trade liberalization and the civil conflict, undoubtedly contributed to falling tax revenues and grants. Yet tax policy and administration were also hobbled by other factors including a narrow tax base, reliance on customs revenue, ad-hoc measures to increase revenues, and poor tax compliance. Previous programs envisioned strategies to address this problem—including through rationalization of tax incentives, and a broadening of the tax base—but met with limited success. Sri Lanka also received comprehensive Technical Assistance on tax and revenue administration reform from 2001 to 2003, but the recommendations were not implemented.1

Government Revenues

(In percent of GDP)

Sources: Ministry of Finance and Planning; Central Bank of Sri Lanka.

B. Crisis trigger and the request for assistance

6. As with many other emerging economies, the need for Fund assistance was triggered by the onset of the global financial crisis. In the aftermath of the collapse of Lehman Brothers, Sri Lanka experienced a sudden stop of short-term financing from international markets, and broader capital outflows driven by redemptions of foreign holdings of government securities, inability to roll syndicated loans over, and sizable external debt repayments. Between the last quarter of 2008, and the first quarter of 2009, the central bank intervened heavily in the currency markets in an unsuccessful attempt to stem the tide of currency depreciation, costing it some 60 percent of its foreign currency reserves. Notwithstanding a rebound in capital flows following the end of the civil conflict in May of 2009, reserves remained dangerously low. Fissures in the domestic financial system were also becoming apparent with the failure of an unregulated finance company, and deposit withdrawals from a systemically important bank.

Gross Official Reserves

(In millions of U.S. dollars)

Source: IMF staff data.

7. The Fund was cognizant of the risks of going into a program. The program request documents underscored the need for a “sharp departure” from the tendency to tightly manage the exchange rate. Unsuccessful efforts at fiscal consolidation in the past were noted, as well as the possibility that the quality of the fiscal adjustment might be affected by the difficulty in raising revenue. Nevertheless, these risks were weighed against the alternative of not having a program, and the high likelihood of a much more socially disruptive balance of payments crisis, and a decision was eventually made to proceed with the program.

Program Targets, Strategy, and Conditionality

A. Program targets and strategy

8. The main program objective was to re-build reserves and restore external viability. This was to be achieved by anchoring the program in reserve targets. These targets were set such that estimated purchases of reserves would require a depreciation of the exchange rate, thereby helping to restore competitiveness. An upward net international reserves (NIR) adjustor was put in place for flows into the local bond market and Eurobond related flows, so that improvements in the external position were not driven by government borrowing. The program also included a clause requiring consultations by the authorities with Fund staff on the appropriate response, in the event that large FX supply-demand imbalances emerged that could threaten an overly disruptive depreciation. The target by the end of the 20-month program was to lift import coverage of reserves to 3.7 months from under 1 month’s cover. The program assumed no further access to commercial external loans during the program period, a slight increase in financing from official creditors to around $1.1 billion per year from an average of $0.8 billion in the period 2006 to 2009, and grants to fall to around $0.2 billion per year, reflecting Sri Lanka’s transition to middle-income status.

9. Monetary policy was anchored on reserve money targets aimed at controlling inflation and ensuring availability of credit. Given that Sri Lanka was not yet ready to move to an inflation targeting regime, this approach was deemed appropriate—especially as it was accompanied by concrete measures to assist the authorities in transitioning to inflation targeting (e.g., by helping them to develop a model of inflation forecasting). For this approach to succeed, it was crucial that the commitment to a flexible exchange rate be maintained, to ensure that monetary and exchange rate policies did not work at cross-purposes.

10. The fiscal strategy focused on headline deficit reduction. Given the strong nexus between the fiscal and the external risks, the strategy—in contrast to ongoing Fund programs elsewhere at the time—involved pro-cyclical fiscal policy with an expected negative impulse at a time of what was seen to be rapid cooling in the economy.2 The path for the reduction in the fiscal deficit was in line with the authorities’ stated plans to reduce the central government’s deficit by end-2011 to 5 percent of GDP, from an estimated 7 percent of GDP in 2008.3 To this end, the program only contained limits on net domestic financing. While a broad adjustment path—slightly tilted toward the expenditure-side—was set out, most adjustment measures required to close the gap were left unspecified. A structural benchmark, however, was put in place for an interim report of a special tax review commission to recommend base-broadening tax measures to be submitted in time for incorporation into the 2010 budget. Eventually, the slow pace of more comprehensive tax reform forced the authorities to rely on some ad hoc lower-quality measures in the first year (raising nation building and excise taxes, and increasing trade-based taxes), postponing the structural heavy-lifting.

11. Curtailing the losses of large state-owned enterprises formed a part of the fiscal strategy. The program aimed to bring to balance the accounts of the two largest state enterprises, Ceylon Electricity Board (CEB) and Ceylon Petroleum Corporation (CPC). In previous years, both of these entities registered large losses: in 2008, their combined deficit reached 1.11 percent of GDP. With oil prices falling in 2009, the combined CEB/CPC deficit was expected to decline, and the authorities aimed at bringing the CEB and CPC accounts to balance by 2011. A two-pronged plan was envisaged to achieve this objective: (i) a move toward improved operational efficiency and lower cost electricity generation; and (ii) more flexible pricing to ensure that electricity and fuel prices reflect the costs. The Sri Lanka Electricity Act passed in March 2009 established an independent regulator for the electricity sector. In addition, a high-level committee was established to monitor the operations of CEB and CPC and provide recommendations on the necessary changes. The program set indicative targets for these enterprises in recognition of the fact that structural reforms in this area lay outside the Fund’s expertise.

12. The program proposed several measures to strengthen the financial sector. The key structural benchmarks introduced at the start of the program included (i) the recapitalization of a major bank (Seylan Bank); (ii) development of a contingency plan for orderly workouts of problem banks and financial institutions; (iii) submission to parliament of a revised Finance Company Act to clarify the legal authority of the CBSL in enforcing its regulations on all deposit-taking finance companies; and (iv) issuance of prudential regulations and guidelines to credit card companies and payment service providers. In addition, the program called for revisions to the Banking Act and other pertinent laws to improve bank resolution, and strengthen the definition of large exposures and related parties to better capture risks. Other structural benchmarks were introduced later, to establish a deposit insurance scheme and a regulatory framework for private pension funds.

B. Program conditionality

13. The fiscal adjustment targeted under Sri Lanka’s SBA is comparable to fiscal adjustment in General Resources Account (GRA)-supported crisis programs during 2008–11. As noted earlier, the degree of adjustment was consistent with the authorities’ own plans for deficit reduction. On average, the 2008–11 crisis programs envisaged a decline in the government deficit between t (starting year of the program) and t+4 by 4.1 percent of GDP, similar to the 3.9 percent of GDP decline projected in Sri Lanka. In contrast, for the broader 2002–11 sample of GRA-supported programs, the envisaged deficit reduction between t and t+4 was only 1.8 percent of GDP.4

Sri Lanka and Other GRA Programs: Government Deficit and Debt 1/

(In percent of GDP)

1/ The solid black line indicates program country averages, dark blue indicates the 25 and 75 percentiles and the light blue indicates the 10 and 90 percentiles, respectively.

14. The relative parsimony of structural conditionality appears to reflect the “macro-criticality” and streamlining requirements for SBAs emphasized at the time. Structural conditionality largely focused on the financial sector, and fiscal policy (Box 1). All financial structural measures at the start of the program were aimed at improving the supervision framework, the application of supervisory powers beyond the conventional banking sector, capitalization of weak institutions, and an improved bank resolution framework (Table 3). These were all germane to macroeconomic stability—particularly viewed against the climate of fear that pervaded financial markets at the time. Fiscal structural benchmarks included either the submission of annual budgets consistent with the program targets, or measures aimed at improving the weak tax revenue collection—a major source of fiscal vulnerability. The large CPC/CEB losses required more borrowing from state-owned banks, posing significant fiscal risks. Thus, the CPC/CEB-related structural conditions could also be viewed as “macro-critical.”

Table 1.Sri Lanka: Selected Economic Indicators, 2008–15 1/
20082009201020112012201320142015
Prel.Proj.
GDP and inflation (in percent)
Real GDP growth6.03.58.08.26.46.36.86.5
Inflation (average)22.43.56.26.77.57.46.96.4
Inflation (end-of-period)13.95.06.84.99.27.46.36.2
Core inflation (end-of-period)10.35.98.94.77.56.56.36.2
Public finances (in percent of GDP)
Revenue 2/14.914.514.614.313.013.313.814.2
Grants 2/0.50.30.20.20.20.20.2
Expenditure 2/22.624.922.821.419.719.319.219.2
Primary balance−2.9−4.0−2.0−0.5−0.2−0.2−0.40.0
Central government balance 2/7.0−9.9−8.0−6.9−6.4−5.8−5.2−4.9
Consolidated government balance 2/−8.1−10.3−8.4−8.6−8.6−7.0−5.9−5.3
Central government domestic financing7.15.14.53.54.14.43.53.2
Government debt (domestic and external)81.486.181.978.479.178.076.274.2
Money and credit (percent change, end of period)
Reserve money1.513.118.821.910.215.514.314.0
Broad money8.518.615.819.117.615.015.015.0
Domestic credit18.00.518.634.321.718.514.814.7
Private sector credit7.0−5.824.934.517.619.018.718.8
Public sector credit48.813.38.233.729.617.87.86.6
Balance of payments (in millions of U.S. dollars)
Exports8,1117,0858,62610,5599,7749,38510,10510,929
Imports−14,091−10,207−13,451−20,269−19,183−18,310−19,627−21,046
Current account balance−3,885−214−1,075−4,615−3,915−3,053−3,092−3,259
Current account balance (in percent of GDP)−9.5−0.5−2.2−7.8−6.6−4.7−4.4−4.3
Export value growth (percent)6.1−12.621.722.4−7.4−4.07.78.2
Import value growth (percent)24.7−27.631.850.7−5.4−4.57.27.2
Gross official reserves (end of period) 3/
In millions of U.S. dollars1,5804,8976,4105,7586,6776,9666,7156,834
In months of imports1.63.93.53.23.83.73.33.1
External debt (public and private)
In billions of U.S. dollars17.820.924.829.433.735.436.938.9
As a percent of GDP43.749.750.149.856.754.452.450.8
Sources: Data provided by the Sri Lankan authorities; CEIC Data Company Ltd.; Bloomberg L.P.; and IMF staff estimates and projections.

Projections reflect available forecasts as of the date of circulation to the Board. These projections have been subsequently updated.

The budget presentation now places grants above the line according to standard practice. Previously grants were classified below the line as a budget deficit financing item. The consolidated government balance includes the Ceylon Electricity Board and the Ceylon Petroleum Corporation. Expenditure arrears in the estimated amount of 0.5 percent of GDP in 2012 are assumed to be cleared in 2013.

Excluding central bank Asian Clearing Union (ACU) balances.

Sources: Data provided by the Sri Lankan authorities; CEIC Data Company Ltd.; Bloomberg L.P.; and IMF staff estimates and projections.

Projections reflect available forecasts as of the date of circulation to the Board. These projections have been subsequently updated.

The budget presentation now places grants above the line according to standard practice. Previously grants were classified below the line as a budget deficit financing item. The consolidated government balance includes the Ceylon Electricity Board and the Ceylon Petroleum Corporation. Expenditure arrears in the estimated amount of 0.5 percent of GDP in 2012 are assumed to be cleared in 2013.

Excluding central bank Asian Clearing Union (ACU) balances.

Table 2.Sri Lanka: Quantitative Performance Criteria (PC) and Indicative Targets (IT)
2009201020112012
End-Mar. Indicative TargetsEnd-Jun. Performance criteria on NER and Indicative Targets on Reserve Money and NDP
EBS/12/36 Adj.Prel.EBS/12/36Adj.Prel.
Quantitative performance criteria and indicative targets
Net international Reserves (NIR) of the Central Bank of Sri Lanka (CBSL) (floor, cumulative change from the beginning of the year, in million US$) 1/2/3/4/5/2,725921−1,061−684−415−251−670−574−324
Reserve money of the CBSL (ceiling, eop stock, in million rupees)303,500358,869442,436468,500485,802474,165462,319
Net domestic financing (NDF) of the central government from the banking system and the non-bank sector (ceiling, cumulative from the beginning of the year, in million rupees) 6/7/8/9/10/392,476242,259259,245230,814248,992300,478296,167
Continous performance criteria
Accumulation of new external payment arrears (ceiling, eop, in million US$)0000000
Indicative targets
Overall balance of the Ceylon Petroleum Corporation and the Ceylon Electricity Board (floor, cumulative from the beginning of the year, in million rupees) 11/−19,400−22,900−113,800
Memorandum items:
External debt service assumed under the program (cumulative from the beginning of the year, in million rupees) 7/141,914108,607132,411
Privatization proceeds to the central government in connection with the sale of central government assets (in million rupees) 8/000
Outstanding claims by the Bank of Ceylon on the central government (item VIII (e, 1) on the balance sheet of the Bank of Ceylon, in million rupees) 9/6,0381,5751,575
Foreign program financing assumed under the program (cumulative from the beginning of the year, in million US$)040
External commercial loans (including Eurobonds and syndicated loans) assumed under the program (cumulative from the beginning of the year, in million US$) 2/5001,0001,0000000
Cumulative net inflows into the Treasury Bill and Treasury Bond market assumed under the program (cumulative from the beginning of the year, in million US$) 1/1,345467232
Official external debt service assumed under the program (cumulative from the beginning of the year, in million US$) 3/8788741,072
Settlement of syndicated loans assumed under the program (cumulative from the beginning of the year, in million US$) 4/2252500000
Outstanding value (i.e., receivables) on swaps and forwards by CBSL (eop stock, in million US$) 5/2459707821,052782878

If the cumulative net inflows into the Treasury Bill market and Treasury Bond market is higher/lower in U.S. dollar terms than assumed under the program, the floor on NIR will be adjusted upward/downward by the cumulative differences on the test date (discontinued after December 2011).

If the amount of commercial borrowing (including Eurobonds and syndicated loans) is higher/lower in U.S. dollar terms than assumed under the program, the floor on NIR will be adjusted upward/downward by the cumulative difference on the test date.

If the amount of official external debt service by the central government in U.S. dollars is higher/lower than assumed under the program, the floor on NIR will be adjusted downward/upward by the cumulative differences on the test date (discountinued after December 2011).

If the amount of debt service on syndicated loans by the central government in U.S. dollars is higher/lower than assumed under the program, the floor on NIR will be adjusted downward/upward by the cumulative differences on the test date. The adjustor is introduced from end-December 2009.

If the outstanding value on swaps and forwards is higher/lower than assumed under the program, the floor on NIR will be adjusted upward/downward by the difference on the test date. This adjustor is introduced from end-March 2012. The outstanding value as of February 29, 2012 was $782 million. The downward adjustment in the NIR target will be up to a maximum of $ 600 million.

If the amount of external loans is higher/lower in rupee terms than assumed under the program, the cumulative ceiling on net domestic financing of the central government will be adjusted downward/upward by the cumulative difference in external loans on the test date. From end-December, external loans will be defined as external program loans and external commercial loans (including Eurobonds and syndicated loans).

If the amount of external debt service by the central government in rupee terms is higher/lower than assumed under the program, the ceiling on net domestic financing of the central government will be adjusted upward/downward by the cumulative difference in external debt service payments measured in rupees.

If the amount of privatization proceeds to the central government in connection with the sale of central government assets is higher/lower than assumed under the program, the cumulative ceiling on NDF of the central government will be adjusted downward/upward by the cumulative receipt/reimbursement of any privatization proceeds.

If the amount of outstanding claims by the Bank of Ceylon on the central government (item VIII (e, 1) on the balance sheet of the Bank of Ceylon) is lower in rupee terms than assumed under the program, the NDF of the central government will be adjusted upward by the difference on the test date.

The 2012 indicative targets on NDF include restructuring bonds amounting to Rs. 60,000 million issued to settle outstanding dues from state owned institutions to the Ceylon Petroleum Corporation.

Slight revisions to end-December 2010 and end-December 2011 data are based on unaudited results.

If the cumulative net inflows into the Treasury Bill market and Treasury Bond market is higher/lower in U.S. dollar terms than assumed under the program, the floor on NIR will be adjusted upward/downward by the cumulative differences on the test date (discontinued after December 2011).

If the amount of commercial borrowing (including Eurobonds and syndicated loans) is higher/lower in U.S. dollar terms than assumed under the program, the floor on NIR will be adjusted upward/downward by the cumulative difference on the test date.

If the amount of official external debt service by the central government in U.S. dollars is higher/lower than assumed under the program, the floor on NIR will be adjusted downward/upward by the cumulative differences on the test date (discountinued after December 2011).

If the amount of debt service on syndicated loans by the central government in U.S. dollars is higher/lower than assumed under the program, the floor on NIR will be adjusted downward/upward by the cumulative differences on the test date. The adjustor is introduced from end-December 2009.

If the outstanding value on swaps and forwards is higher/lower than assumed under the program, the floor on NIR will be adjusted upward/downward by the difference on the test date. This adjustor is introduced from end-March 2012. The outstanding value as of February 29, 2012 was $782 million. The downward adjustment in the NIR target will be up to a maximum of $ 600 million.

If the amount of external loans is higher/lower in rupee terms than assumed under the program, the cumulative ceiling on net domestic financing of the central government will be adjusted downward/upward by the cumulative difference in external loans on the test date. From end-December, external loans will be defined as external program loans and external commercial loans (including Eurobonds and syndicated loans).

If the amount of external debt service by the central government in rupee terms is higher/lower than assumed under the program, the ceiling on net domestic financing of the central government will be adjusted upward/downward by the cumulative difference in external debt service payments measured in rupees.

If the amount of privatization proceeds to the central government in connection with the sale of central government assets is higher/lower than assumed under the program, the cumulative ceiling on NDF of the central government will be adjusted downward/upward by the cumulative receipt/reimbursement of any privatization proceeds.

If the amount of outstanding claims by the Bank of Ceylon on the central government (item VIII (e, 1) on the balance sheet of the Bank of Ceylon) is lower in rupee terms than assumed under the program, the NDF of the central government will be adjusted upward by the difference on the test date.

The 2012 indicative targets on NDF include restructuring bonds amounting to Rs. 60,000 million issued to settle outstanding dues from state owned institutions to the Ceylon Petroleum Corporation.

Slight revisions to end-December 2010 and end-December 2011 data are based on unaudited results.

Table 3.Sri Lanka: Structural Benchmarks (SB)
Structural BenchmarksDateStatus
Recapitalization of Seylan Bank through a public share issuance.9/30/2009Implemented
A contingency plan for orderly workouts of problem banks and financial institutions will be developed by the CBSL.9/30/2009Implemented
Approval by the Monetary Board of a revised Banking Act and other pertinent laws and legislations that: (i) improve the bank resolution framework that more clearly defines the provisions for acquisition, and roles of the conservator and liquidator; and (ii) strengthens the definition of large exposures and related parties to better capture all material risks.9/30/2009Implemented
Submission by the tax review commission of an interim report, including on base broadening measures to be incorporated into the 2010 budget.10/15/2009Implemented with delays
Approval by Parliament of an interim budget for the first four months of 2010 consistent with program targets.12/15/2009Implemented
Develop a plan to address outstanding debts between the CEB, CPC and state-owned banks.12/31/2009Implemented
Issuance of prudential regulations and guidelines to credit card companies and payment service providers.12/31/2009Implemented
Submission to parliament of the 2010 budget consistent with program targets.4/30/2010Implemented with delays
Submission to the parliament of a revised Finance Business Act which includes clarifying the legal authority of the CBSL in enforcing its regulations on all deposit taking finance companies.5/31/2010Implemented with delays
Submission by the Presidential Tax Commission of a final report, including specific tax reform measures.8/31/2010Implemented with delays
Parliamentary approval of full-year 2010 budget consistent with program targets.8/31/2010Implemented
Issuance of regulations to establish a deposit insurance scheme.9/30/2010Implemented
Submission to parliament of the 2011 budget consistent with program targets, including income tax and VAT reform measures.11/30/2010Implemented
Cabinet approval of a regulatory framework for private-sector superannuation funds.12/31/2010Not met
Restructure the overdue obligations accumulated up to end-2009 by Ceylon Electricity Board to Ceylon Petroleum Corporation.12/31/2010Implemented with delays
Amend BOI regulations.3/31/2011Implemented
Amend the Strategic Investment Law and issue supporting regulations on defining and contracting strategic projects.3/31/2011Implemented
Cabinet approval of amendments to the Banking Act.3/31/2011Implemented with delays
Cabinet approval of amendments to the Petroleum Act.9/30/2011Implemented with delays
Submission to parliament of the 2012 budget.11/30/2011Implemented

Box 1.Structural Conditionality in the Program

Structural conditionality was initially focused on the financial sector and, to a lesser extent, fiscal measures. On the financial sector, structural benchmarks were aimed at addressing the key vulnerabilities identified by the 2007 Financial Sector Assessment Program (FSAP), and included the recapitalization of Seylan Bank and revisions to the Banking and Finance Company Acts to clarify the role of the CBSL and improve the prudential and regulatory framework. Fiscal conditionality required a submission to parliament of a 2010 budget consistent with the program goals, submission by the tax review commission of an interim report, and the development of a plan to address the debts of the CEB and CPC.

In the course of the program the additional benchmarks that were added focused on the same areas. They either built on the existing benchmarks (such as “Submission by the Presidential Tax Commission of a final report”) or refined the original benchmarks to achieve their goals (e.g., “Restructure the overdue obligations accumulated up to end-2009 by Ceylon Electricity Board to Ceylon Petroleum Corporation”). New fiscal measures were added on the 2010, 2011, and 2012 budgets and on amendments to Board of Investment (BOI) regulations and the Strategic Investment Law in an attempt to control the proliferation of tax exemptions. Further financial conditionality was also introduced, requiring Cabinet approval of amendments to the Banking Act and other measures to strengthen the banking sector, as well as a regulatory framework for private-sector superannuation funds.

Compared with other GRA programs, the number of structural benchmarks in Sri Lanka’s SBA was relatively parsimonious. On average, there were 2.5 structural benchmarks per review. This compares with average 6.1 structural benchmarks per review in SBAs during 2002–11, and an average of 4.3 benchmarks during the period 2009–11 when the streamlining of structural conditionality took effect. As for the composition of structural conditionality, Sri Lanka’s program was very similar to the average composition of GRA programs with high access, with fiscal and financial structural conditions each representing about one-third of total conditions.

Sri Lanka: Structural Conditionality

Program Outcomes

A. Growth and inflation outcomes

15. Growth and inflation generally turned out better than projected. Average real GDP growth during 2010–12 reached 7.5 percent, compared with a 5.8 percent average growth projection at the program approval. At the same time, inflation turned out to be significantly lower: the 2010–12 average was 6.8 percent, compared with the program projection of 10.2 percent. As mentioned at the start of this report, the exceptionally uncertain circumstances under which the program was initiated made it difficult to estimate the effect of the ending of the civil conflict on potential growth, and to calibrate policy judgments associated with forecasts of crucial macroeconomic variables. In this regard, it is also interesting to note that growth forecasts appear pessimistic in 2010 and 2011, perhaps reflecting a degree of caution embedded in program projections.5

Growth Forecast Errors, Sri Lanka vs. All Countries

Sources: IMF staff calculations.

1/ Plotted distribution includes all countries, percentile rank refers to all countries.

2/ Projections made in the spring WEO vintage of the preceding year.

Real GDP Growth

(In percent, yoy)

Source: IMF staff calculations.

Inflation

(In percent, yoy, average)

Source: IMF staff calculations.

16. The program may have supported stronger growth through confidence channels. Undoubtedly, the end to the civil conflict played a key role in re-integrating the economy. The country’s Northern and Eastern regions have experienced robust growth (averaging respectively some 15 percent and 8 percent for 2010–11, according to government estimates). However, given that they constitute just 10 percent of the total economy, the high overall growth rate cannot be attributed to re-integration alone.6 The potential growth rate of the overall economy has been estimated by Fund staff to be about one percentage point higher since 2003–08, reflecting capital deepening, and Total Factor Productivity (TFP) growth in equal measure, according to the 2013 Article IV staff report (Country Report No. 13/120). Viewed from the expenditure side, growth in the program period (relative to the preprogram years) is driven by final consumption, which has offset a stronger drag on growth from net exports—noteworthy for a program anchored on exchange rate flexibility. The connection between the program and strong growth outcomes may therefore be related to stronger confidence, a “peace dividend,” and an improving external environment.

Box 2.Main Areas of Slippages and Modification during Program Implementation

In the context of resilient growth, and large capital inflows, the first review was completed without incident, and the NIR target (adjusted up by $1¾ billion) was exceeded by a large margin, allowing the future targets to be revised upward. Toward the end of the year, however, it became clear that the fiscal targets for 2009 had been missed by a very wide margin: revenues fell short of projections by around ¼ percent of GDP, and together with expenditure overruns on capital spending (partly due to some 1 percent of GDP in earlier than programmed disbursements of external financing), interest payments, and recurrent expenditure resulted in significant breach of the end-December domestic budget borrowing ceiling. The deficit reached nearly 10 percent of GDP against a program target of 7 percent.

Under the combined second and third reviews completed in June 2010 the program was brought back on track: the fiscal targets were loosened by 2 percent of GDP in 2010 and 1.8 percent in 2011, reflecting the worse-than-expected initial position, and the goal of a 5 percent fiscal deficit was postponed by one year to 2012. The program was extended to end-2011 in recognition of the additional time needed to contain the fiscal deficit.

The third, fourth, fifth and sixth reviews took place without major slippages: a waiver was granted for a missed NIR target for end-December 2010, largely caused by payments necessitated by the cancellation of an oil-related credit line. Some structural conditionality was added, focusing on fiscal measures and containing exemptions. At the sixth review the frequency of program reviews was shifted from three-monthly to six-monthly, on the basis that the authorities’ fiscal performance in 2010 and the launching of the 2011 budget with tax reform measures had boosted the credibility of their policy commitments.

The sixth review had begun to sound the alarm on the need for sufficient exchange rate flexibility in the face of an increase in demand for imports, and suggested the need for monetary tightening if credit continued to accelerate. As the current account deficit increased very sharply in the course of the year, however, the authorities sold increasing amounts of foreign exchange in order to keep the exchange rate fixed. As a result, the program went off track in the summer of 2011, and the end-June and end-July PCs on NIR were missed by a large margin.

In early 2012 agreement was reached on bringing the program back on track: The measures included (i) the announcement of a credit growth target (under 20 percent), and a rise of 100 bps in the repo rate; (ii) adjustment in petroleum and electricity prices, and a package of measures to secure a 6.2 percent of GDP target for 2012; and (iii) an up-front adjustment in the exchange rate of more than 15 percent, along with a track record of reduced intervention. Following the implementation of these measures, the program resumed, and was extended to July 2012, though no new structural benchmarks were added. The eighth and final review was completed in July 2012, with the end-June NIR target met.

B. External sector outcomes

17. The program was initially very successful in improving the balance of payments position. As sentiment reversed with the end of the conflict and the advent of the Fund program, the country experienced significant capital inflows, to which the authorities responded appropriately by accumulating reserves. As a result of the NIR adjustor on external flows related to government borrowings, the end-September 2009 target was adjusted up by $1.7 billion. Yet given the improvements in the balance of payments the authorities were still able to exceed the target by a further $450 million. In the first half of 2010, the authorities were again able to exceed the NIR targets by a further $200 million. The initial (2009) improvements in the balance of payments position were largely driven by import contraction (due to lower growth, and a fall in import prices). The external position was also supported by strong official financial flows, which were almost one-third higher than envisaged in the program documents. Both multilateral (including the World Bank and Asian Development Bank) and bilateral (notably from China and Japan) project financing exceeded projections, with donor financing averaging $1.7 billion/year in 2009–11 compared with $1.1 billion projected in the initial program documents. However, relatively little budget support was forthcoming, in part reflecting the strained international political context. Stronger-than-expected interest by foreign investors also allowed the country to issue $0.5 billion in external debt in 2009 (and some $1 billion per year over the next three years).

Evolution of NIR and GIR

(In millions of U.S. dollars)

Source: IMF staff estimates.

18. Once the exchange rate began to come under pressure, however, it became clear that the authorities’ were once again tempted to defend the currency. From September 2010 onward the CBSL’s net absorption of foreign exchange began to turn into net sales; from that month until June 2011 net sales totaled $570 million. Yet during this same period, rather than allow the exchange rate to depreciate in line with the balance of payments pressures, the authorities oversaw a slight appreciation of the rupee against the dollar from 112.5 to 109.6. In the latter half of 2011 the pressure on the exchange rate to depreciate increased with the surge in imports. With net foreign exchange sales averaging $440 million per month in the second half of the year, the authorities missed the NIR targets by a large margin in their efforts to keep the exchange rate stable.

Foreign Exchange Intervention

(In millions of US dollars)

Source: Central Bank of Sri Lanka.

19. This led the program to pause for an extended period. The pressure on the rupee remained elevated (as evidenced by the level of intervention and the missing of NIR targets). The authorities appear to have been convinced that balance of payments pressures would ease, as they did in 2009, due to renewed capital inflows. In the event, these did not materialize, and the authorities eventually enacted a set of corrective policy measures in early 2012—allowing a significant depreciation of the exchange rate, tightening monetary policy, and imposing direct credit ceilings—to mitigate external pressures and cool the domestic economy (Box 2). The authorities’ willingness to undertake these measures was seen as reflecting renewed commitment to the program, and on this basis, the program resumed in March 2012.

20. Once the CBSL transitioned to a more flexible regime in early 2012, the program restored reserves to a more comfortable level. Over the course of the program, gross international reserves increased from $2.1 billion to almost $7 billion, the equivalent of over 3.6 months of prospective imports. This compares well with the original program goal of achieving reserves of 3.7 months of imports at the end of the envisaged program period. Although it remained slightly under the original projection for 2012 of four months of imports, this reflects the sharp growth in imports rather than a shortfall in reserve accumulation: the original projection for 2012 had been for less than $5 billion in gross reserves. The rupee also moved significantly during the latter part of the program: when the program was approved, the exchange rate stood at around 112 rupees per dollar; by the end of the program it had depreciated to around 133 rupees per dollar. By the end of the program period the CBSL’s intervention in the foreign exchange market had been greatly reduced, suggesting that the supply and demand for foreign exchange had been brought closer to balance.

Gross Official Reserves

(In months of imports)

Sources: IMF staff estimates.

Gross Official Reserves

(In millions of U.S. dollars)

Sources: IMF staff estimates.

C. Monetary sector outcomes

21. Monetary targets were largely met, but inconsistencies between monetary and exchange rate policies were a significant factor behind the program going off-track in 2011. In the initial phases of the program, monetary policy was eased as credit growth sharply fell. However, credit growth (partially spurred on by moral suasion) began to pick up—albeit from a low base—substantially by the second half of 2010, breaching the 3 percent “warning threshold” for growth in the credit to GDP ratio in early 2011. However, given the uncertainties in the monetary policy transmission mechanism, the judgment of the authorities through this period was that total credit to GDP could still grow in line with demand, as inflation pressures seemed relatively benign. However, with the CBSL heavily intervening to defend the currency toward mid-2011 (while the monetary policy framework was anchored on reserve money targets), monetary and exchange rate policies were working at cross-purposes—neither were attempts at reining in credit growth and cool the economy through moral suasion successful, nor were the balance of payments pressures (i.e., strong import growth) alleviated. The acceleration of credit continued through most of 2011.

Domestic Credit and Imports Growth

(In percent, yoy)

Sources: CEIC; and IMF staff calculations.

Private Credit to GDP Ratio 1/

(In percentage points, yoy)

Sources: CEIC; IMF staff calculations.

1/ The horizontal line shows a 3 percent “warning threshold” (September 2011, GFSR.)

2/ Computed using Annual GDP (CY).

22. Weaknesses in the monetary policy transmission mechanism were exacerbated by constraints on the ability to manage domestic liquidity. Liquidity management relied on overnight open market operations or on affecting conditions in the standing facility in which banks deposit their liquid reserves on an overnight basis. The inability to sterilize the changes in liquidity resulting from foreign exchange purchases through mid-2010 beyond overnight maturities contributed to excess domestic liquidity. Although these difficulties were noted—the Fifth review explicitly pointed to the need for improved liquidity management, as well as the need to expand the scope of available monetary policy instruments—there does not appear to have been further follow-through in these areas. Staff did hold extensive discussions with the authorities and MCM experts, especially on the reliance on overnight facilities, but the development of longer-term liquidity management instruments appears to have been viewed as a longer-term priority.

D. Fiscal sector outcomes

23. The program achieved a reduction in headline deficit and debt, but the adjustment was unbalanced. After missing the 2009 fiscal deficit target by almost 3 percent of GDP, the authorities brought down the deficit down to 6½ percent by 2012. Total public debt fell to below 80 percent of GDP after peaking at 86 percent of GDP in 2009. However, the composition of adjustment was significantly different from what the program had envisaged. Primary expenditure ended up almost 1 percent of GDP below target; and tax revenues continued to fall quite precipitously despite the continuously expected increases, underperforming the original program target by almost 4 percent of GDP in 2012. Spurred by favorable interest rate-growth dynamics, the headline debt-to GDP ratio declined some 6 percentage points of GDP from 2009–12, registering a slight rise in 2012 due to the weaker rupee.

Overall Fiscal Balance

(In percent of GDP)

Source: IMF staff estimates.

Tax Revenue

(In percent of GDP)

Source: IMF staff estimates.

Total Public Debt

(In percent of GDP)

Source: IMF staff estimates.

Primary Expenditure

(In percent of GDP)

Source: IMF staff estimates.

24. Measured in cyclically adjusted terms, targets for each year gradually weakened as the program evolved. The recalibration of the adjustment path during the second and third reviews did not fully “claw-back” past slippages during the remainder of the program period, although the fiscal deficit was to be reduced by 2 percentage points of GDP (or double what was originally programmed).7 Nevertheless, the accommodation of revenue underperformance and greater-than-expected expenditure reductions meant that the targeted impulse and withdrawal on the revenue and expenditure side, respectively, became incrementally larger, while the overall fiscal stance loosened relative to original set targets through the course of each year (text figure).8

Sri Lanka: Fiscal Stance

(Base year = 2005)

Source: IMF staff estimates.

25. The authorities’ will to reduce expenditures to meet fiscal targets in the face of falling revenues was commendable. Noninterest current spending bore the brunt of adjustment in the course of the program. Difficult political decisions were taken to control increases in public sector wages and to tax public sector employees. Expenditure restraint also spanned reductions in categories, such as security procurement and—to a lesser extent—subsidies. The authorities’ willingness to tackle these issues was regarded as demonstrating their commitment to the program.

Reductions in Expenditure, 2009–12

(In percent of GDP)

Source: IMF staff data.

Reductions in Primary Current Expenditure, 2009–12

(In percent of GDP)

Source: IMF staff data.

26. However, there may have been scope to shield priority areas—such as infrastructure investment—more effectively. Defense cuts almost exclusively affected procurement, leaving the wage bill virtually unchanged (due, in the authorities’ view, to political and social stability concerns surrounding a large-scale military demobilization). The reduction in subsidies and transfers was mainly the result of slightly lower subsidies to agriculture and public corporations (such as railways and postal services), as well as lower pension expenditure. However, the scope to reduce less productive outlays may not have been fully exploited—to the detriment of the investment budget. Total capital spending is now low compared to East Asian peers, and lack of infrastructure may act as a drag on growth, as noted in the 2013 Article IV consultation. Going forward, higher revenues and reductions of remaining slack in current expenditures are necessary to make room for investment and social spending.

Reductions in Subsidies and Transfers, 2009–12

(In percent of GDP)

Source: IMF staff data.

Reductions in Security Spending, 2009–12

(In percent of GDP)

Source: IMF staff data.

27. CEB and CPC losses increased substantially. Instead of falling to zero as targeted under the program, the combined CEB/CPC operational loss reached 1¾ percent of GDP in 2011, and was 2 percent of GDP in 2012. Initially, the reduction of the combined CEB/CPC loss proceeded faster than programmed, and the targeted zero loss in 2011 appeared to be within the reach. In 2009, the CEB/CPC loss fell to Rs. 23 billion (0.5 percent of GDP), compared to the Rs. 35.4 billion indicative target. In 2010, the loss reached Rs. 26.6 billion (0.5 percent of GDP), compared to the Rs. 35.7 billion (revised) and Rs. 30 billion (original) indicative targets. The end-2009 structural benchmark to develop a plan to address outstanding debts between the CEB, CPC and state-owned banks was also met. There were also some adjustments made to electricity and fuel prices. However, after such a promising start, the progress quickly reversed. Instead of falling to zero in 2011, the combined CEB/CPC loss widened by over 1 percent of GDP compared to 2010. This largely reflected exogenous factors, such as higher oil prices, and the shift in power generation from hydro, to more expensive thermal power. In addition, there were delays in meeting the structural benchmarks: the end-2010 structural benchmark to restructure overdue obligations by CEB to CPC, and the end-September 2011 structural benchmark of Cabinet approval of amendments to the Petroleum Act. Following further domestic price increases and introduction of fuel adjustment surcharge in February 2012, the eighth review projected a decline in CEB/CPC 2012 losses. Instead, reflecting a sharp rise in costs of power generation, the losses rose further by one-third.

Sri Lanka: Tariffs and Losses, 2007–12

Sources: Ministry of Finance and Planning; and IMF staff calculations.

E. Financial sector outcomes

28. The program had mixed success tackling financial sector weaknesses. On the positive side, the CBSL made important progress in providing supervisory guidance to banks on a comprehensive risk management framework, expediting the move to risk-based supervision, and strengthening the risk management system of banks. This was a significant achievement in bolstering macro-financial stability. Nevertheless, the 2012 FSAP Update (key recommendations of which can be found in Appendix II) noted that deficiencies still exist relative to large exposure limits and exposure to related parties, and concluded that a comprehensive program for consolidated supervision is still a work in progress. These were aspects that were included in the original structural benchmarks of the program. The Fund has emphasized that the Banking Law needs to pay greater attention to the bank resolution framework, consolidated supervision, and related-party lending issues. Also worth noting is that establishment of a regulatory framework for private pension funds (an end-2010 structural benchmark intended to contribute to longer-term capital market development) was put off as it became entangled with an initiative to create a mandatory private pension scheme that encountered opposition. In terms of banking sector vulnerabilities, the external liabilities of the banking system have increased notably—a factor cited in the recent lowering of the outlook on Sri Lanka’s sovereign rating by Moody’s. The NPL ratio has fallen (estimated at 4.7 percent at H1 2013, some 4 percentage points lower since the program’s inception), although the ratio of Tier-1 capital to risk weighted assets, has also fallen slightly since then to about 13 percent. Recently, the central bank announced the provision of liquidity support to finance companies as difficulties emerged once again at one such company. This underscores the need to strengthen regulation in this sector, again as noted in the 2012 FSAP update.

Key Questions

A. Did the program facilitate a “catalytic effect” for the country at a difficult time?

29. The program played an important part in ameliorating immediate external pressures at a difficult time. Even though the end of the civil conflict in May 2009 already began to turn around sentiment and reverse the capital outflows, reserves were still low at the time of the approval of the program in end-July at some $2.1 billion (compared with some $1.5 billion at end-May). By the time of the completion of the first review in October 2009, reserves had more than doubled to some $4.8 billion (five months of imports), with a large increase not only in portfolio flows to the bond market, but also in remittances and other flows, allowing the authorities to exceed reserve targets by a very comfortable margin. As a result, by the end of the year, it was clear that Sri Lanka had averted a potentially very disruptive external adjustment, and was able to meet its international trade needs and financial obligations smoothly.

30. The period of program engagement also marked a more durable turn in confidence. Market-based indicators point to the fact that this catalytic effect carried beyond the initial period of program approval. Sri Lanka’s EMBI Global spread narrowed sharply after program approval. In March of 2009, Sri Lanka’s spread was well over 1000 basis points wide of the overall EMBI Global spread. By the time of the approval of the program this spread differential narrowed to the vicinity of 300 basis points, and through the duration of the entire program averaged about 160 basis points. Just as striking is the fact that Sri Lanka’s spreads did not widen significantly, nor did the local currency bonds move appreciably well after the completion of the program. In fact, one major investment bank report read “…it is important to recognize that the impact (of IMF engagement) is now psychological…an assurance that the policy reforms…would be taken to their logical conclusion.”9 Sri Lanka benefitted also from sovereign rating upgrades, as well as the ability to issue progressively longer dated market debt, both domestically and internationally over the course of the program.

Sri Lanka and Global EMBI Spread

(In basis points)

Source: Bloomberg.

B. Was targeting reserves the right approach to restoring external viability?

31. Targeting international reserves was consistent with long-standing recommendations in favor of exchange rate flexibility. In the months running up to the program’s agreement, it had appeared evident that a depreciation of the exchange rate would be required to alleviate the drain on the balance of payments. With the end of the conflict, however, the external pressures began to alleviate and even reverse, making it less clear that a sharp depreciation would be warranted. In this context, the program aimed to rebuild reserves: to 2.4 months of imports (around $2.5 billion) by the end of 2009, and more than 3½ months of imports at end-2010 (around $4 billion).

32. The design of the reserve target was novel, and fulfilled the task of building reserves. The program attempted to ensure that reserve accumulation was not achieved via temporary inflows of capital which could be quickly reversed, thereby leaving the fundamental problems of a pegged and overvalued exchange rate unaddressed. The adjustor was therefore added to the target on Net International Reserves (NIR) for short-term inflows into the domestic bond market, as well as Eurobond issuance related inflows, in an attempt to ensure that the reserve buildup was generated through a current account adjustment rather than through external borrowing. While the program did not explicitly target a fall in the value of the currency, staff believed that a depreciation of the currency was likely to be necessary to attain the NIR targets, and macroeconomic projections were based on this taking place. In the event, capital inflows were much stronger than anticipated. With the benefit of hindsight, there may have been scope to build reserves even further.

33. However, issues related to competitiveness still need to be addressed, and external vulnerabilities remain high. The trade deficit tripled over the course of the program, suggesting that the underlying problems of competitiveness cannot be addressed solely through exchange rate flexibility—a point which is aptly recognized in the 2013 Article IV consultation. While some increase in the trade deficit would have been justified by the post-conflict situation and the reconstruction and investment needs, the $6.3 billion increase in the trade deficit dwarfs the $0.4 billion increase in FDI, leaving much of the deficit to be financed by debt.10 Perhaps reflecting the SBA’s relatively short-term focus, there was little progress on the underlying difficulties faced by the export sector (trade barriers, red tape, business climate, and infrastructure). Indeed, certain measures such as the passage of legislation (in November 2011) enabling expropriation of assets of “underperforming” enterprises—ostensibly to reclaim idle government land—may have dampened the enthusiasm of foreign investors to enter the country, as reflected by the coverage of this issue in market and other external reports.

34. Moreover, the commitment to exchange rate flexibility may be tested anew. After it depreciated sharply following the package of measures announced in early 2012, the currency once again appreciated (by around 6 percent) against the dollar between June 2012 and April 2013.11 Since mid-May 2013, as broader external sentiment toward emerging markets soured, other emerging Asian currencies have depreciated against the dollar by an average of some 10 percent, compared to the Sri Lankan rupee which has depreciated by some 4.5 percent amid market reports of moral suasion on currency dealers, as well as direct intervention in the currency markets.

C. Have debt-related vulnerabilities declined?

35. External debt sustainability has not improved as projected during the program. Between 2009 and 2011, external debt remained relatively stable (around 50 percent of GDP), but it jumped to 57 percent of GDP in 2012, about 10 percent of GDP higher than projected at the time of the program request. A sharp widening of the current account deficit in 2011 and 2012 was the most important contributor to the increase in external debt ratio (text figure). Favorable impacts from automatic debt dynamics (real GDP growth, currency appreciation) helped to keep the debt ratio down for some time, but eventually, the loss of foreign reserves led to currency depreciation, which contributed to the jump in the debt ratio in 2012. Non-debt-creating capital inflows (FDI) picked up somewhat in the second half of the program, but at 1.5 percent of GDP financed only about one-fifth of the current account deficit, with official and private sector borrowing covering the rest. Looking ahead, the still large share of longer-term concessional debt is likely to mitigate somewhat, the existing rollover risks, but the debt ratio remains sensitive to large currency depreciations. This underscores the importance of productivity and competitiveness-boosting reforms in achieving a steady reduction of current account deficit.

Contribution to External Debt Changes

(In percent of GDP)

Source: IMF staff estimates.

36. Fiscal vulnerabilities persist, and by some measures have risen, despite the improvement in headline deficits and debt. The ratio of debt to tax revenue increased from 6 in 2008 to more than 7 in 2012, while interest payments reached almost half of total tax revenues, from about a third in 2008. Debt service including the roll-over of short-term debt was more than double of tax collections in 2012. As tax revenues have declined, expenditure reductions will reach their limits, in particular given the need for growth-enhancing infrastructure investment. Debt reduction has largely been the result of favorable interest rate-growth dynamics—the actual contribution of the average interest rate to debt dynamics between 2009 and 2012 was substantially lower than projected at the time of the program. This reflects a number of factors: (i) primary fiscal consolidation; (ii) a captive domestic investor base; (ii) improved investor sentiment, which permitted longer maturity and lower interest debt to replace short-term higher interest rate debt more effectively than was (cautiously) envisaged at the time of program inception; and (iv) to a lesser extent, lower interest rates on external debt issuance, and the search for yield in a low interest rate global environment. Although the estimated reduction in the primary deficit was over 2 percentage points of GDP over the program period, it should also be noted that the initial program had projected a primary surplus by 2011. The original program’s targeted fiscal consolidation (and its composition) was sufficient in size and quality. If it were not for slippages in revenue collection, the effectiveness of the consolidation would have been stronger. Looking ahead, debt dynamics exhibit high sensitivity to a growth slow-down and/or a failure to sustain fiscal consolidation, which in turn could push interest rates up in a less supportive global environment.

Contributions to Debt Changes

(In percent of GDP)

Source: IMF staff calculations.

D. Why was the quality of the fiscal adjustment weaker than envisaged?

37. Quality adjustment measures for the program period were largely left unidentified. Of the 4½ percent of GDP in required adjustment in 2011, 2½ percent of GDP were spending reductions expected to come from security-related spending and reductions in transfers. However, there was a paucity of concretely specified measures in both areas. On the revenue side, a number of “less than ideal” measures were taken as prior actions to boost revenues quickly.12 These were intended to be phased out as the envisaged higher quality reforms aimed at broadening the tax base were to yield results. In the event, they constituted the bulk of identified measures in the program request.

Revenue and Expenditure Measures under the Program

(In percent of GDP)

Source: IMF staff estimates.

38. Greater emphasis could have been placed on critical structural fiscal reforms earlier in the program. At the time of approval, the program left the underlying structural fiscal efforts insufficiently spelled out, and to be implemented disproportionately in the outer years. Arguably, the time of the original request may have been most suitable to focus on higher-quality tax reform measures, for example, as the sense of urgency with respect to the program may have been stronger. The tumultuous external backdrop at the time of program approval may have limited the time available for more clear specification of measures at the outset. Nevertheless, this could have been done at the earliest opportunity after the program had been put in place.13 Also, the importance of strengthening tax collection was apparent, with weaknesses in auditing and enforcement explicitly mentioned in the program request along with the need for tax reform. Yet, no technical assistance on tax administration was sought until much later in the program. The inclusion of a dedicated FAD economist or expert to the team might have been helpful.

39. The design and execution of program conditionality could have better supported the implementation of high-quality adjustment measures. Some of the most important objectives mentioned in the program request—to broaden the tax base and rein in tax incentives administered by the Board of Investment—were folded into the general benchmark of the submission of the interim report of a tax commission. None of the (relatively few) structural benchmarks were elaborated in the TMU, and benchmarks were in some cases vaguely specified (e.g., “Amend BOI Regulations (by 3/31/11)”, and”Submission to Parliament of a 2012 Budget (by 11/30/11)”). Also, the program was repeatedly adjusted to accommodate delays in implementing structural benchmarks, and underperformance on the revenue side. The Special Commission on tax reform was significantly delayed, while new tax exemptions continued to be granted after the program was approved. Critically, the tax reform that was eventually adopted did not take on board all of the recommendations of the Special Commission in terms of its design and implementation. Tax revenues kept falling as a share of GDP throughout the program.

E. Why were tax revenues falling throughout the program?

40. All categories of tax revenues underperformed relative to projections, except for excises. The original program expected most of the tax revenue gains between 2009 and 2012 to come from the VAT, trade taxes and income taxes. In the event, those were the tax categories that also contracted the most (domestic VAT collections in particular), while excises far outperformed expectations set at the time of the program request.

Actual and Projected Changes in Tax Revenues, 2009-12

(Percent of GDP)

Sources: IMF staff calculations.

41. The erosion of tax revenues likely reflects shortcomings in tax reforms, as well as weak tax administration. The 2011 tax reform contained a number of positive elements, also recommended by the 2009 tax policy TA mission: the BOI regime of tax incentives was brought into the legal framework; import taxes were substantially streamlined; VAT rates were unified albeit at a low rate, and some “nuisance” taxes were abolished (Appendix I). These were all welcome steps in the right direction in terms of improving the transparency and efficiency of the tax system. However, it also fell short in a number of important respects: it saw a unification of the VAT at a low rate, and a reduction in corporate and income tax rates without a commensurate broadening of the tax base. A more strategic approach to sequencing the reforms gradually—by prioritizing base broadening first, carefully calibrating the pace of rate reductions, and carefully aligning revenue targets—might have made a difference. Instead of such a “phased approach,” as still suggested during the fourth review, significant up-front rate reductions were accompanied by little base broadening. Despite the elimination of some nuisance taxes, still other taxes and special regimes were maintained. Furthermore, while the nation building tax was extended to the retail and wholesale sectors, its rate and threshold were lowered, making it hard to administer in return for a low expected revenue yield. Yet, as previously noted, technical assistance on tax administration was not requested until the very end of the program.

42. The structural capacity of revenue administration is open to question, and the will to stop leakage needs to be stronger. If poor capacity contributed to the poor revenue performance, this would imply that targets may not have been realistic, and the structural measures not complete. Certainly, technical assistance on revenue administration should have been called upon in a timelier manner. Yet, there was a perception that deeper structural reforms were not appropriate under the SBA and its initial 20-month duration. The poor revenue performance may partly reflect the fact that this tension was unresolved. Also, even as efforts are being undertaken to bring firms that have enjoyed tax holidays into the tax net over time, new tax incentives and exemptions have been added each year (within the legal and budgetary frameworks), further eroding the tax base and complicating efficient tax administration.

F. Why were goals with respect to state-owned enterprises not achieved?

43. The short program time-frame, combined with unfavorable exogenous developments put the goals with respect to the state-owned enterprises out of reach. Prima facie, the losses reflected the failure to pass fully the costs to the prices. Average electricity tariffs remained well below the average costs, even as CPC continued to supply fuel oil at subsidized rates to the CEB, and diesel and kerosene costs were not fully passed to the consumers. However, a flexible price mechanism alone could not realistically eliminate the losses that were mounting in 2011 and 2012 despite several domestic price adjustments. Sri Lanka’s electricity tariffs were already relatively high by regional standards, and full cost recovery price would have pushed them higher. More flexible pricing needs to be accompanied by improvements in the cost structure of electricity generation. This, in turn, requires improved corporate and financial management, as well as further development of cheaper power generation (hydro, coal) to replace the more expensive thermal plants. Some of these aspects lie outside of the core expertise of the Fund and the design of reforms might have benefitted from sectoral expertise of other multilateral institutions (the Asian Development Bank’s 2013–15 country program aims to address these issues). Moreover, the sensitivity of CPC/CEB’s financial positions to exogenous shocks—fluctuations in world oil prices and rainfall affecting hydro generation—played a significant part in the underperformance of the state-owned enterprises. Two out of three structural benchmarks related to the CPC/CEB dealt with clearing the accumulated debt, and did not contain specific measures how to address the ongoing losses.

G. Could monetary policy have been tightened in 2011 to head-off external pressures?

44. Monetary policy was indeed slow to tighten, but tighter policy alone would have been insufficient. The Third and Fourth reviews were prescient to warn of a forthcoming rise in imports. The Fifth review (January 2011) made explicit mention of a need for vigilance on monetary policy, and there may have been a plurality of views within the country on the need for tightening. Fund staff made the call—correctly, in the view of this report—that it would be better to allow the eventuality of a weaker exchange rate to take a central role in curbing import growth, instead of relying on monetary policy tools alone. This was driven by two factors: Firstly, an overarching conviction that acting through the exchange rate was essential to bring monetary and exchange rate policies in alignment, and consistent with broader program goals. Secondly, there was also a concern—explicitly expressed to the authorities when the program was on pause—that absent the resumption of exchange rate flexibility, an “overreliance” on monetary policy tightening to cool credit growth may result in an unnecessary slowdown in the economy.

45. Increased efforts to improve liquidity management would have been likely to help curb credit growth. Liquidity in the banking system experienced a surge by late 2010, mainly as foreign exchange inflows were not fully sterilized. With the benefit of hindsight, it may have been possible to put in place a few measures that could have had a measurable impact on liquidity management in a relatively short time, notwithstanding the preexisting difficulties. As noted in the 2012 FSAP Update, these measures could have included the following:

  • In terms of liquidity monitoring, a measure of structural liquidity that excludes CBSL operations with a maturity of less than one month from bank reserves could have been employed. This can be viewed as a measure of the average level of underlying liquidity in the banking system. This may have helped to understand and manage liquidity in the banking system more effectively.

  • In terms of strengthening liquidity management, the CBSL could have moved to undertaking longer-term repo operations (not just overnight operations), generally with maturities of between 7 and 30 days, selected to offset future expected liquidity flows, and with longer-term instruments being used for managing structural liquidity. The CBSL could also, as an alternative, have undertaken outright sales/purchases of interest bearing securities, issued by itself or the government. The authorities are not in favor of this, because they view, from past experience, that the markets are mispricing these securities.

  • The statutory reserve requirement for banks could also have been modified by lengthening the maintenance period—a measure that the authorities now plan to undertake. A move from partial to full reserve averaging may also have been considered.

H. Was exceptional access justified?

46. The program envisaged access at 400 percent of quota (SDR 1.65 billion), to be disbursed in eight equal installments on the basis of quarterly reviews, with the final review scheduled for March 2011. While the 600 percent cumulative limit on access would not be breached, the original schedule envisaged access of 200 percent in the first year of the program and hence required exceptional access.14 This exceptional access was justified in the program documentation by the need to ensure a substantial increase in resources available to the central bank in light of the steady depletion of foreign currency reserves, the potential for nonresident deposit outflows, and fragile confidence in the domestic currency.

47. The four criteria for exceptional access were met, ex ante:

  • The member is experiencing or has the potential to experience exceptional balance of payments pressures on the current or capital account resulting in a need for Fund financing that cannot be met within the normal limits. With gross reserves having fallen to $1 billion in the months before program approval, equivalent to only around one month of imports, the country was clearly experiencing severe balance of payments pressures. Given the broader context of the global financial crisis a large program was felt to be appropriate to restore confidence. With the information available at the time, the forecast recovery in the balance of payments was appropriately conservative. The amount of financing in the program appeared justifiably driven by an abundance of caution.

  • A rigorous and systematic analysis indicates that there is a high probability that the member’s public debt is sustainable in the medium term. If the authorities’ commitment to fiscal consolidation were maintained, the analysis demonstrated the sustainability of the country’s debt burden under plausible and conservative assumptions for growth (a reversion toward its historical mean), inflation (significantly below observed levels during the run-up to the crisis), and interest rates (averaging nearly twice the historical average on foreign exchange denominated debt). The stress tests were clear in indicating vulnerability of the baseline projections to fiscal slippages, as well as to large one-time shocks to the exchange rate.

  • The member has prospects of gaining or regaining access to private capital markets within the timeframe when Fund resources are outstanding. With foreign investment in the domestic bond market already returning at the time the program approved, this criterion was clearly satisfied.

  • The policy program of the member country provides a reasonably strong prospect of success, including not only the members’ adjustment plans but also its institutional and political capacity to deliver that adjustment. Even though the prior track record of the government (in power since late 2005) was seen as favoring policies to spur faster short-term growth, the ending of the civil conflict was seen as providing a chance to chart a new course for macroeconomic policies. Given the government’s stated commitment to the fiscal targets—which were consistent with their own ambitions for deficit reduction—and the fact that it had a substantial parliamentary majority, Sri Lanka’s prospects for achieving the program’s goals appeared to be strong.

48. The ratio between Fund financing and adjustment also appeared broadly appropriate at the program’s inception. The proposed schedule implied IMF disbursements were to account for around two-thirds of the increase in international reserves in 2009, half of the increase in 2010, and forty percent of the 2011 increase. This struck a balance between the need to accumulate reserves, while accommodating necessary expenditures in the post-conflict situation. A more front-loaded disbursement profile would have been difficult to justify in light of the relatively back-loaded nature of the reform program, with significant fiscal measures remaining to be spelled out at the start of the program.15

49. Within six months of the inception of the program, however, the need for significant modifications became apparent. Two factors were responsible for this: First was the significantly worse-than-anticipated fiscal outcome in 2009 that caused the program to go off-track, and second, the more favorable external context. The reversal of flows that came with the end of the conflict and the start of the Fund program had eased the BOP pressures significantly. With the strong balance of payments pressures abated, the need for exceptional access was now less strong.

50. The phasing of disbursements and program duration were adjusted appropriately to reflect changed circumstances. When agreement was reached on completing the second and third reviews in the summer of 2010, it was also decided to extend and re-phase the program to reflect the fact that while exceptional access was no longer an imperative, more time would be needed to address the fiscal situation. The program was extended to end-2011; the remaining 300 percent of quota under the original program was divided into nine equal disbursements, with two disbursements in June 2010 following the completion of the combined second and third reviews.

Conclusions and Possible Lessons

51. On economic grounds, Sri Lanka’s need for assistance in 2009 was evident. The balance of payments pressures were acute. The climate of heightened uncertainty in the immediate aftermath of the global financial crisis, as well as the end to the country’s civil conflict presented extraordinary circumstances. Fund support was necessary and crucial for the country to wrest the initiative presented by the moment, and to avoid a severe shock at a singular point in its history. The exceptional uncertainty of the time justified the request for exceptional access, which in the view of this report, was granted in accordance with Fund practices. As circumstances changed, the program was re-phased such that exceptional access did not occur, de facto. This was also appropriate.

52. Viewed through the lens of averting a highly disruptive external shock, Sri Lanka’s program was successful. The program provided a catalytic effect to confidence at a crucial time when the country was in need of assistance. The primary goal of the authorities in seeking Fund assistance—smoothing the economy’s path through the external shock—was achieved. Moreover, the program was associated with a period of strong growth and low inflation as the economic benefits of the improved security situation kicked in, and the external environment became more favorable. The Fund was also cognizant about the risks to the program at its inception, and appears to have engaged constructively with the authorities during periods when the program ran into difficulty, eventually bringing the program back on track. This marked a departure from other recent Fund engagements with Sri Lanka, and merits recognition. For their part, the Sri Lankan authorities see the program very positively, attributing its successful completion to high “ownership” they had in setting the program targets and designing its underlying policies (see Annex).

53. Viewed through the broader prism of the achievement of other program objectives, success was partial.

  • The program was ultimately successful in rebuilding reserves and the design of the NIR target contained a novel element that can be associated with this success. However, vulnerabilities with respect to external debt have not receded. The commitment to a flexible exchange rate regime may be further tested, while external competitiveness issues need to be addressed.

  • On the fiscal front, thanks to effective expenditure control by the authorities, the program did succeed in lowering fiscal deficits. However, an adjustment strategy based almost exclusively on expenditure cuts will inevitably run into its limits. The life of the program saw revenues as a share of GDP fall precipitously. This marked reduction in revenues—placing Sri Lanka below most of its lower-middle income peers—constrains the government’s ability to finance much needed (and growth inducing) capital expenditure. Moreover, the drop in revenues has raised overall fiscal vulnerability.

  • The program was also unable to meet its goals with respect to bringing the state-owned enterprises into balance. This is in part due to the short program timeframe, which may not have been suited to achieving the goals, the lack of Fund expertise to reform corporate governance, as well as exogenous factors, such as rainfall, and global oil prices.

  • Finally, the program period was associated with some success on strengthening of the financial sector—notably a move to risk-based supervision, but progress still needs to be made on other fronts, as noted by the 2012 FSAP update (see Appendix II).

Tax Revenue, 2011

(In percent of GDP)

Sources: IMF staff estimates.

54. With hindsight, some lessons can be drawn:

  • More attention should be paid to the country’s track record in devising the program strategy. This report began with the observation that Sri Lanka’s fiscal revenues had been on a long-term decline—a malady that previous engagements had not tackled. Under these circumstances, more attention could have been paid to underlying revenue objectives, and not just to program targets that prioritized headline deficit reduction.

  • Large scale fiscal reforms with an uncertain yield should be made as “water tight” as possible, and phased in strategically. While not seeking to underplay the difficulties associated with putting a program together at a time of crisis, the realism of the program goals should (a) be supported with credible bottom-up measures that are clearly specified; and (b) be marked by strong ownership of measures by the country authorities. For example, in Sri Lanka’s case, tax reforms could have been sequenced such that the base was broadened to support revenue collection before enacting large rate reductions. Political considerations (including but not restricted to risks of reversal) may have driven the speed with which rate reductions were put in place. However, if cutting tax rates is politically easier than broadening the base, durably reclaiming lost revenues could be an uphill struggle.

  • Monetary and exchange rate policies need to operate together under a more coherent framework. As noted in the 2013 Article IV Report, consideration could be given to flexible inflation targeting in a manner that is consistent with the CBSL’s growth, inflation, and financial stability objectives. This will take time and study, but could be valuable in helping Sri Lanka make a decisive break from repeated patterns of building up imbalances—by allowing domestic liquidity conditions to remain consistent with domestic and external stability, while allowing the exchange rate to act as a buffer against external shocks.

  • More attention should be paid to addressing structural capacity gaps. Several areas where the program was not fully successful appear to share the common problem that there was not enough structural capacity to deliver. This appears to be true not only for tax reforms, but also with respect to the ability to sterilize foreign exchange purchases, monitoring of banking system liquidity, reform of the governance of state-owned enterprises, and financial sector reform. This necessitates more careful specification of structural benchmarks with realistic measures to go with them, and just as importantly, a fuller consideration of the necessary technical assistance that is needed. Allowances should also be accordingly made when judging the issue of whether structural conditionality is adequately streamlined, or perhaps too parsimonious.

  • Program duration and the ambition of program goals should match. There is always a difficult balance to strike between ambition and realism. Programs that are short-term by construction (either due to the preferences of the Fund or of the country authorities) may be better suited to a surgical focus on emergency needs and the enactment of credible and identified adjustment measures to build a track record, trust, and ownership—after this, structural deficiencies can be more methodically tackled through subsequent longer-term engagement. On the other hand, if it is judged that adequate buy-in exists to enact bold structural reforms during a time of urgency, then program length needs to be based on a realistic assessment of the technical difficulties, and the need for assistance in overcoming them. Put differently, to the extent that Sri Lanka’s program is indicative of broader experience, it may be useful to consider the possibility of keeping a sharper distinction between SBAs and EFFs—particularly if in some circumstances, it may be difficult to put together a well designed multi-year EFF under emergency circumstances. In this respect, it is interesting that other recent EPEs have also pointed to potential mismatches between program duration and structural capacity to deliver on ambitious goals (Greece)—also in similar areas such as VAT reform (Pakistan), emphasizing the importance of tax administration technical assistance (Pakistan), and state-owned enterprise reform (Pakistan).

  • The flexibility afforded by Fund policy to reassess the appropriateness of the financing facility could be more fully exploited. After Sri Lanka’s SBA had gone off-track following the first review, the time of the second and third reviews may have presented an opportunity to re-engage with the authorities on the basis of an EFF. This might have given more scope to deal with the many structural issues. This possibility was not favored by Fund staff—based on the judgment that an exceptional access SBA as approved is better to be completed as such, particularly if it has gone off-track. In the event, Sri Lanka’s program was extended twice, eventually lasting three years, but with the “legacy” of having started as an SBA, and with the attendant continuity of program goals, targets, and objectives. Despite indications from the authorities that they may have been open in 2010 to a longer-term engagement in principle, it is entirely possible that a concrete agreement may have been difficult to reach. Nevertheless, with an eye to the future, consideration to increased flexibility in thinking about these issues may be warranted.

Annex. Views of the Authorities

The views of the Sri Lankan authorities were sought during a visit to Colombo from July 22–31, 2013. Meetings were held with the Central Bank of Sri Lanka, the Ministry of Finance and Planning, the Board of Investment, the Department of Inland Revenue, the Customs Revenue Department, the Ceylon Petroleum Corporation (CPC), and the Ceylon Electricity Board (CEB). Consultations were also held with broader stakeholders, including with the private sector, local economists, other multilateral institutions, and the political opposition. The views of the authorities are summarized below.

Overall views on the SBA, and Fund involvement

The authorities viewed the SBA as a successful program, designed and executed against a difficult global backdrop. It represented the longest engagement with the Fund, and the first program in recent history to be successfully completed. This was viewed as an important achievement, strongly signaling the authorities’ commitment to pursuing sound macroeconomic policies. The authorities noted that the end of the civil conflict represented a unique point in the country’s history. The policy course taken resulted not only in stronger growth and lower inflation, but for the first time in recent history, a continuous reduction in the fiscal deficit. The Fund program, in the view of the authorities, deserves a share of the credit for these positive developments. The primary reason for this success, as seen by the authorities, is that they had adequate ownership of the program. The country team was credited with a genuine understanding of their broad vision, keeping program design consistent with “home-grown” policy choices that best matched the country’s social and political realities. The authorities observed that the experiences gained over the period of the SBA will serve them well into the future, and stated that they continue to view the Fund as a useful partner and technical advisor.

The major criticism from the authorities was the length of time between negotiations which began in February of 2009, and the final approval of the program in July of that year. In the authorities’ view, the essential elements were put in place by end-March 2009, but the subsequent delay was unnecessarily protracted, due to political reasons, even in the face of demonstrable economic need for assistance. In addition, the authorities also expressed the view that although the overall size of the program financing was adequate, more front-loaded phasing was warranted. Finally, the authorities emphasized their view that it would have been helpful to have had less frequent changes in the composition of the country team.

Views on program implementation and outcomes

The authorities noted that the Fund program facilitated a catalytic effect on confidence, not just in 2009, but in the years that followed as well. Even though conditions began to turn around following the end to the civil conflict, the authorities recognized that continued engagement through the SBA beyond the immediate crisis in early 2009 represented an important longer-term signal of credibility to domestic and international investors, and enabled the sovereign to access international bond markets at longer maturities and lower yields, along with higher sovereign ratings.

The central bank felt that targeting reserves was the right approach to restoring external viability. The authorities pointed to the success of the program in raising reserves to a comfortable level early on, which in their view also gave them a cushion during the period in 2011 when balance of payments pressures re-emerged. According to the central bank, targeting an up-front devaluation, as suggested by the Fund initially, would have unnecessarily upset confidence while worsening public and external debt dynamics. In contrast, the design of the program’s Net International Reserve (NIR) target allowed the eventual course of flows to provide for a natural accretion of reserves, and the determination of the exchange rate.

Regarding debt-related vulnerabilities, the authorities noted that the public debt-to-GDP ratio fell over the course of the program, and credited the primary adjustment, improved investor confidence, as well as skillful debt management for this. While emphasizing strong long-term commitment to further debt reduction, the fiscal authorities believe that they would be able to comfortably pursue a public investment target of about 6 percent of GDP. As long as recurrent expenditure could be contained to match revenues, and underlying growth is healthy, the debt dynamics would continue to improve, in their view. The authorities are aware that revenue declines, and the progressive substitution of long-term concessional debt with market financed external debt, affect overall perceptions of debt sustainability. However, they view these as manageable challenges. The fiscal authorities took pride in the progressive reduction of the deficit following the initial slippage. This success, in their view, was a product of the broader recognition that the details of the expenditure-revenue mix should be left to domestic policy choices that adapt to the political and social realities of the country. The main near-term priority was to entrench a belief early in the country’s post-conflict rebuilding years that the persistent fiscal deficits of the past would be resolutely lowered with each year, while the stage was simultaneously set for longer-term fiscal structural reforms to bear fruit. Over the program period, both of these goals were broadly achieved in the view of the fiscal authorities.

The fiscal authorities attributed the decline in revenues primarily to lower than envisaged import-related revenues—particularly following the measures that were taken to contain balance-of-payments pressures that arose in 2011. They also viewed lower imports as having indirect “spillovers” to domestic revenue collections, through its effects on revenue generating domestic economic activities. They also observed that more incremental and phased approach to lowering tax rates commensurate with base broadening, had been tried unsuccessfully in the past. The authorities viewed an approach of incremental rate reductions as carrying potential risks of reversal, even if the path of rate reductions were clearly preannounced. Therefore, in their view, the rapid reduction of rates was the more pragmatic approach to take. The authorities expressed optimism that efforts to expand the tax base will eventually bear fruit, as existing tax exemptions start to lapse, while tax compliance is also envisaged to increase. The fiscal authorities welcomed technical assistance from the Fund to address revenue administration weaknesses, and also anticipated that technological improvements and other modernization efforts that are being undertaken at Inland Revenue and Customs Departments would lead to higher rates of compliance.

There is a clear recognition—not just by the authorities, but in society more broadly—that the reform of state-owned enterprises remains a significant work in progress. There was a general concurrence of views that the initial program’s indicative target of setting the combined CEB and CPC balance to zero by end-2011 was ambitious. For the former, the time needed to reduce the weather-dependence of financial outcomes was acknowledged to be significantly longer than originally envisaged in the program. Representatives from CEB expressed hope that within the next year, the mix of energy generation will change durably as new power plants come on-stream. For both enterprises, the authorities noted that sensitivity to political considerations would require price adjustments that are incremental rather than rapid. Nevertheless, the authorities noted that the program was helpful in keeping the issue of SOE reform in public focus. The authorities believe the price adjustment mechanisms are now delivering expected results, and that the process of establishing a mindset that these enterprises must be made viable, is now entrenched.

Regarding the re-emergence of balance-of-payments pressures in 2011, the authorities expressed the view that this was the natural manifestation of the expectation of a peace dividend. Broadly, the authorities felt it was important to let the growth momentum that began in 2010 sustain, and recognize the general sentiment that the war-time repression of the economy was ending. In that context, they felt that the stronger than envisaged reserve accumulation of the preceding period afforded some cushion to lean against depreciation pressures, and that this was the appropriate time to make some use of this cushion. However, as these developments further intensified, they felt that the time had come by early 2012 to re-establish exchange rate flexibility, and send stronger signals of tightening that included a credit growth ceiling, upward adjustment of policy interest rates, imposing higher duties on the import of vehicles and a few other commodities, and increasing the administratively determined prices of petroleum products.

Acknowledging the presence of structural capacity gaps in key areas such as revenue administration, and the corporate reforms of the state-owned enterprises, the fiscal authorities noted that they may have been open to a recalibration of the program in 2010 towards a longer-term engagement with more structural components—provided that the nature of the attendant conditionality matched domestic social and political realities. Fund engagement could have served as a public commitment mechanism for the necessary reforms, if an agreement could have been reached on the elements of such a reconstituted program.

Appendix I. Tax Reform Measures in Sri Lanka: 2009–12
Tax areaMain TA recommendationsReflection in program languageMain elements of 2011 tax reform
2nd and 3rd Review; LOI of June 19, 20104th review; LOI of September 14, 2010
VAT1. Abolish NBT within 2 yrs and to compensate expand VAT base/raise rate if necessaryNeed to simplify VAT/PIT-have sought TA; consider reducing the # of taxes and broadening their basesTake steps to streamline VAT and Income taxes: broaden tax bases and unify and reduce rates; suggested “phased approach” with rate reductions only as base broadening has progressed (to reduce need to rely on NBT to compensate for any revenue shortfalls)NBT extended and reduced to 2 percent (from 3); threshold reduced to broaden base; no phased approach or compensating measures through NBT; “base broadening measures have created welcome room for rate reductions”; NBT to be replaced by VAT over medium term
2. Confine exemptions to health, education, social services, property and basic food; fold commodities under SCL into VAT.Maybe replace regional turnover tax with VAT and eliminate two-thirds of exemptionsExempting VAT on petroleum “and other sectors” (telecom, local software, leasing)
3. Expand VAT to wholesale/retail trade AFTER VAT is simplified; the base expanded; and tax administration is strengthened. The threshold would need to be increased.Maybe extend VAT to retail and wholesaleRTT replaced with NBT and extended to retail and wholesale
4. 20 % rate for banks and luxury goods should be reduced to standard rate.Unify VAT rates; any reduction in rates should be compensated with an increase in NBTVAT unified at 12 percent
5. VAT withholding should be repealed or limited to nonresidents.
Excises1. No room to further increase tobacco, alcohol, motor vehicles and petroleum rates-adjust with inflation.Reduced duties on motor vehicles by 25%; increased taxes on alcohol and tobacco
2. Bring small cigarette nmanufacturers into tax net.
3. Excise imported and domestic alcohol in the same manner-merge excises departments in IRD and customs.
4. Repeal excises on products other than those under 1.
5. Exclude VAT from base for ad valorem excises.
Other indirect taxes1. Repeal debits tax and compensate through VAT/ITRepealed
2. Repeal CIFGL and increases income tax withholding contacts from 1 to 3 percent.Exemptions for special projects of MoF
3. Impose single levy on mobile phones.Removed from VAT, two new levies
4. Repeal SRL and adjust rates.Repealed
5. Repeal stamp duties at national level.Transferred to provinces
Trade taxes1. Reduce total tax burden on imports; simplify the rate structure; offset revenue loss through VAT base broadening and rate.Reduced #of tariff bands;Simplified regime; surcharges eliminated; lower rates on MVs; temporary tax exemptions on essential goods reversed; higher taxes on alcohol and tobacco4 bands
2. Abolish PAL and EDB and increase duty rates accordingly; abolish RIDL and increaes excises on motor vehicles; abolish SCL and include goods in VAT.Reduce exemptions; incorporate import surcharges in duties; essential import commodities subject to unified SCLEliminate nuisance taxesSCL only to apply to selected imports; RIDL abolished
Income tax1. Make base as broad as possible
2. Redraft law to simplify it.PIT down to 24; taxes on income from alcohol, tobacco and casinos at 40 percent; higher threshold
3. Quantify in-kind benefits better (fringe benefits are taxed)Exempt motor vehicle benefits up to threshold
4. Repeal all tax exemptions on employment income, including public servants.Civil servants and nonresidents included in the tax base
5. Include capital gains (from asset sales, stocks, real estate etc.) into taxable
income
6. Repeal exemptions for interest income.
7. Tax all income from capital at the same rate.
Tax all sectors at the same rate
8. Maintain income theshold; reduce brackets; and consider reducing PIT rates as the base is broadened.
9. remove CIT concessions and consider reducing basic rate gradually as revenue considerations permit.CIT down to 25 from 35CIT down to 28; agriculture included; but new/extended exemptions
10. Repeal all other corporate taxes, including for quoted companies; small businesses; SRL on undistributed profits; and ESC.ESC with 4-band structure and higher threshold
11. Retain final withholding tax on dividends and strengthen provisions.
Tax incentives1. Reduce tax expenditures delivered through the tax system.Reduce ad hoc tax holidays nad duty exemptions; caps on new tax holidays and limits on BOI exemptions will be consideredRationalize investment incentive regime with fewer tax concessions; concentrate on large foreign projects
2. Prepare account of all tax expenditures, their purpose and their cost.List all goods that will be subject to import duty regardless of importer’s BOI status (1/3 of imports now duty-free)Negative list for BOI imports included in budget
3. Subject them to regular review, just like direct expenditures.All projects with concessions will be gazetted; large projects require parliamentary approval
4. Incorporate all tax incentives into IRA; change operation of BOI.BOI asked to refrain from granting new concessions until new regulations are finalizedBOI repurposesd and all concessions only in tax laws/strategic investment laws; no longer BOI discretion
5. Move away from tax holidays and create more level playing field with lower uniform rate of CIT.No renewal of tax holidays; cancelation of approved but not started BOI projects
6. Any incentive to investment should be general and in the form of accelerated depreciation or investment tax credits linked to the amount of investment.5-yr tax holidays for invesntemnets between $3-10 million; lesser concessions for smaller projects
7. Targeted incentives should be used sparingly to address market failures.Option of strategic investments, but subject to NBT and ESC
8. IRD to enforce tax filing by all tax holiday companies.
Tax administrationWorking with ADB to simplify processes, the management information system, HR development, tax audit, and IT use.Tax administration improvement important by simplifying procedures; improving management information systems; and strengthening audits (SR); “tax policy and administration will need to be priority”No specifics
Sources: IMF Country Reports; FAD TA Report “Reform of Direct and Indirect Taxes,” November 2009
Sources: IMF Country Reports; FAD TA Report “Reform of Direct and Indirect Taxes,” November 2009
Appendix II. Sri Lanka—Key Recommendations for Financial Stability from the 2012 FSAP Update
Systemic risk monitoring, cross-sectoral linkages, and macro-prudential policy
  • Strengthen the macro-prudential framework by formalizing roles and responsibilities of members of the Inter-Regulatory Institutions Council (IRIC).

  • Map ownership structures of all financial groups and identify significant ultimate beneficiary owners.

Banking sector soundness
  • Require annual bottom-up stress testing by all systemically important banks.

  • Initiate a program for a large systemically important state-owned bank to increase capital and liquidity buffers over the medium term to enhance its resilience to various shocks.

  • Continue to improve corporate governance at state-owned banks.

Nonbank Financial Institutions
  • Harmonize the regulations of banks and deposit-taking nonbank financial institutions in order to diminish the possibility of regulatory arbitrage.

  • Eliminate possibility of providing ELA (Emergency Liquidity Assistance) to finance companies from SLDIS (Sri Lanka Deposit Insurance Scheme) funds; if desired, replace by inclusion of finance companies in ELA under Section 86 of the MLA (Monetary Law Act).

  • Remove the requirement that all state assets and contracts should be insured with state-owned companies.

  • Continue to enhance the supervisory capacity of the IBSL (Insurance Board of Sri Lanka), including through cooperation with IFIs.

  • Supervise offshore activities of insurance companies.

  • Develop an overall pension strategy and introduce a robust supervisory framework for all pension funds.

Supervision and regulation
  • Revise current hiring practices to effectively recruit candidates with specialized skills.

  • Ensure full adoption of all Pillars of Basel II Accords.

  • Expand definition of related parties.

  • Revise statutory provisions on large loans to international norms.

  • Develop a program of consolidated supervision.

  • Introduce legislation to facilitate information exchange between both domestic and foreign financial sector supervisors.

Financial safety nets and crisis management
  • Develop a clearly defined and well documented financial crisis management plan for the entire financial sector.

  • Conduct periodic crisis simulation exercises.

Salient among the recommendations was the introduction of a unified Revenue Authority, which the authorities decided against.

In the event, growth turned out higher than expected, and reaching the program targets in 2009 and 2010 would actually have implied a neutral stance (see section on Program Outcomes).

The 2003 Fiscal Responsibility Act had aimed to reduce the deficit to 5 percent of GDP by 2006.

For the countries in the sample, see 2011 Review of Conditionality, Background Paper 4, Appendix 12, available via internet at http://www.imf.org/external/np/pp/eng/2012/061812d.pdf.

The authorities’ growth forecasts for this period were also lower than outturns, albeit less so than the Fund’s. It is also noteworthy that local analysts have expressed the view that the government statistical office’s methodology of deriving GDP estimates needs to be refined.

The Northern and Eastern regions have provided a boost to the estimated potential growth rate and, raising overall growth rate by a quarter of a percentage point per year, per the 2013 Article IV report (Country Report No. 13/120).

Given that the public debt ratio in 2010 was lower than programmed due to stronger growth, a partial “claw-back” was deemed adequate at the time.

The fiscal stance figure shows for 2009–12, in each year: (i) the original program targets for the revenue stance, the expenditure stance, and the overall fiscal stance; (ii) the final targets prior to the test date; and (iii) the actual outcomes. For each year, the overall stance (blue squares) became increasingly looser, reflecting a loosening of the revenue stance (red bars) less than offset by a tightening of the expenditure stance (green bars).

JP Morgan, March 2013.

The Gross external financing requirement rose from 13.3 percent of GDP in 2009, to 20.4 percent of GDP in 2012.

The 2013 Article IV Report (Country Report No. 13/120) estimated that the rupee was overvalued by between 4–8 percent as of the time of its publication.

Staff Appraisal, program request staff report (Country Report No. 09/310).

During the EPE mission to Sri Lanka, the fiscal authorities indicated that they may have been open to a more structurally based approach to the program after it came back on track in 2010, provided that they had adequate say in the design of the measures.

Sri Lanka’s program was small in comparison to other exceptional access programs that were approved around the same time: Latvia, Ukraine, and Romania, had access of 1200 percent, 800 percent, and 1100 percent of quota.

The 2011 Review of Conditionality noted that front-loaded disbursements tended to be associated with front-loaded conditionality (Background Paper 2: Design of Fund-Supported Programs, Box 8).

Other Resources Citing This Publication