Sri Lanka: Selected Issues and Statistical Appendix

This Selected Issues paper on Sri Lanka reviews several issues that highlight both Sri Lanka’s accomplishments and their policy constraints amidst a protracted period of civil conflict and political instability. High intermediation costs have held back development of the financial sector and could also frustrate Sri Lanka’s quest for higher growth. The main constraints to achieving higher growth include the civil conflict, political instability, high fiscal deficits and inflation, and underdeveloped financial markets.

Abstract

This Selected Issues paper on Sri Lanka reviews several issues that highlight both Sri Lanka’s accomplishments and their policy constraints amidst a protracted period of civil conflict and political instability. High intermediation costs have held back development of the financial sector and could also frustrate Sri Lanka’s quest for higher growth. The main constraints to achieving higher growth include the civil conflict, political instability, high fiscal deficits and inflation, and underdeveloped financial markets.

III. High Interest Spreads in the Banking Sector—Causes and Remedies1

A. Introduction

1. Insufficient progress in bringing down financial intermediation costs could frustrate Sri Lanka’s quest for higher growth and poverty reduction. This chapter finds that financial deepening has stagnated and interest spreads have remained high (Section B). Among the main causes for this lack of progress, the chapter argues, are inefficient state banks in combination with poor competition in the banking sector (Section C). Some policy implications associated with the later results are discussed (Section D): If the banking sector were sufficiently competitive, inefficient banks would adjust or leave the market, this way contributing to a reduction in interest spreads. Insufficient competition, on the other hand, requires that the government actively promote efficiency in the banking sector, both as shareholder and supervisor. While the chapter makes a few suggestions to overcome weak competition, identifying the underlying root causes requires further research.2

B. Evidence of Weak Financial Intermediation

2. Progress in financial deepening has been disappointing over the last decade. As shown in Figure III.1 bank credit to the private sector in Sri Lanka is broadly in line with the level in low-income countries. However, bank credit to the private sector declined over the last eight years and Sri Lanka is far away from the level of lower middle-income countries. Low credit to the private sector is a significant impediment to growth. Unless the trend can be reversed, Sri Lanka is unlikely to achieve the medium-term target of boosting investment from 25 percent to 30 percent of GDP and increasing growth from the historical average of 5 percent to 7 percent.

Figure III.1.
Figure III.1.

Bank Credit to Private Sector, 1995–2003

(In percent of GDP)

Citation: IMF Staff Country Reports 2005, 337; 10.5089/9781451972542.002.A003

Source: World Bank, World Development Indicators.

3. Crowding out by the public sector explains part of the stagnation in private sector credit. As shown in Figure III.2, the lackluster performance of private sector credit is not due to a lack of funds, as deposits increased continuously since 1998. Bank financing of burgeoning fiscal deficits more than offset the increase in deposits, leading to a decline in private sector credit-to-GDP until 2001.

Figure III.2.
Figure III.2.

Sri Lanka: Origin and Destination of Funds, 1995–2004

(In percent of GDP)

Citation: IMF Staff Country Reports 2005, 337; 10.5089/9781451972542.002.A003

Source: IMF, EDSS.

4. In addition, high interest spreads have undermined financial intermediation. The interest spread pattern in Sri Lanka resembles the observations on credit to the private sector. While the level of the interest spread is broadly consistent with countries at the same level of development, the spread has increased somewhat over the observation period (Figure III.3). High spreads drive a wedge between the interest received by savers and paid by investors, render many investment projects unprofitable, and are likely to have contributed to the stagnation in private sector credit.

Figure III.3.
Figure III.3.

Sri Lanka: Interest Rate Spread, 2002–2004

Citation: IMF Staff Country Reports 2005, 337; 10.5089/9781451972542.002.A003

Source: CBSL Annual Report, various issues.

5. The maintenance of high interest spreads does not seem to have impaired the monetary transmission mechanism. Changing interest spreads can render the transmission of monetary policy impulses to the economy ineffective. In 2003, for example, the loosening of monetary policy could have been diluted by the increase in spreads that took place in that year. This does not seem to have happened, as illustrated in Table III.1. Interbank rates, treasury bill rates, lending rates, and deposit rates are all closely correlated with policy rates. On the contrary, the monetary transmission mechanism seems to have improved over recent years. A similar table in the 2002 FSAP showed a very weak correlation between policy rates and deposit rates. Consistent with a high interest spreads, however, the correlation between deposit rates and lending rates is weaker than most other correlations.

Table III.1.

Sri Lanka: Correlation Between Various Interest Rates, 2003 1/

article image
Source: CBSL, Annual Report 2003.

The data refers to 2003, because in 2004 there was little variation in policy rates.

Primary Market.

C. Causes of High Interest Spreads

State Banks

6. Inefficient state banks still account for a significant share of the banking sector. There are two state-owned commercial banks in Sri Lanka, the Bank of Ceylon and People’s Bank. Their market share in terms of assets (and deposits), while falling, still accounted for 45 percent in 2003 (Figure III.4). These banks are saddled with large NPLs, owing to poor management and government intervention in the past (Table III.2). In addition, their personnel and operating costs are well in excess of the sector average reflecting, among other things, overstaffing. As a result, their profitability is poor and People’s Bank’s level of capital inadequate—in fact, its capital has been negative for several years.3

Figure III.4.
Figure III.4.

Sri Lanka: Market Share by Type of Bank, 1998–2003 1/

Citation: IMF Staff Country Reports 2005, 337; 10.5089/9781451972542.002.A003

Source: Audited Accounts, Sri Lanka Commercial Banks.1/ Share of banks’ assets in total assets.
Table III.2

Sri Lanka: Banking Sector Indicators by Type of Bank, 1998–2004

article image
Source: Audited Accounts, Sri Lanka Commercial Banks.

Percent of total assets.

Percent of total income.

Percent of NPLs.

Total capital in percent of risk-weighted assets. Only since 2002, data includes foreign banking units.

Profits as percent of end-of-period assets.

7. Remaining inefficiencies and past legacies of state banks prevent a reduction in their interest spreads. Personnel costs continue trending up, consistent with the observed pattern in interest spreads. In addition, new management teams have tried to bring the return-on-assets more in line with the sector norm and started to address the banks’ NPL problems. As a result, profitability, capital adequacy ratios, and provisioning for NPLs improved considerably over the last years. However, to the extent that state banks started to operate on commercial principles, the underlying weaknesses became apparent and interest spreads increased (Figure III.5).

Figure III.5.
Figure III.5.

Sri Lanka: Interest Margin of State Banks, 1999–2004

Citation: IMF Staff Country Reports 2005, 337; 10.5089/9781451972542.002.A003

Source: Audited Accounts, Sri Lanka Commercial Banks.1/ Net interest income to average assets, percent.

Lack of Competition

8. There are few alternatives to commercial bank financing. This limits competition and keeps spreads high. As shown in Table III.3, competitors to commercial banks hold 50 percent of Sri Lanka’s financial assets. However, alternative sources of financing for the private sector are much smaller. More than half of specialized bank assets, or 7 percent of total financial assets, belong to the National Savings Bank, a state bank that by design invests most of its funds in government securities. The same is true for the provident funds, which comprise 23 percent of financial sector assets. Hence, real competition to commercial banks is confined to 20 percent of the financial sector.

Table III.3.

Sri Lanka: Financial Sector, 2004 1/

article image
Source: CBCL Annual Report 2004.

Excluding CBSL.

9. Concentration in the banking sector is in line with international standards, and falling. As illustrated in Figure III.6, Sri Lanka’s Herfindahl Index—a common indicator of concentration and, thus, lack of competition—is broadly in line with the financially much more advanced EU countries. If specialized banks were included, on the grounds that they compete with commercial banks on some products (e.g., time deposits), Sri Lanka’s index would be even smaller. Also, banking sector concentration has been falling over the last couple of years. This finding is robust to the choice of index (Herfindahl versus M-Concentration, that is, the market share of the M largest banks) and the definition of market (deposits, loans, assets, all banks, commercial banks).

Figure III.6.
Figure III.6.

Country Comparison: Herfindahl Index for Total Assets, 1998 and 2003 1/

Citation: IMF Staff Country Reports 2005, 337; 10.5089/9781451972542.002.A003

Sources: Audited Accounts, Sri Lanka Commercial Banks; Report on EU Banking Structure, 2004.1/ The Herfindahl index is the sum of the squares of the market shares of all banks competing in the market.

10. However, price competition seems to be largely absent. Table III.4 tries to establish a link between the level of interest spreads, averaged over 1998–2003, and gains in market share between 1998 and 2003. For this purpose, 18 commercial banks for which data are available were grouped according to the level of their interest spread and the average gain in market share was calculated for each group. The largest gains in market share were realized by banks with medium interest spreads, while banks with the lowest spreads actually lost market share over the period under consideration. The finding of limited price competition is robust to the definition of market (assets, loans, deposits), the choice of price (interest spread versus interest margin) and the number of groups (two versus three groups).

Table III.4.

Sri Lanka: Correlation Between Average Interest Spread and Increase in Market Share, 1998–2003 1/

article image
Source: Audited Accounts, Sri Lanka Commercial Banks.

Sample size: 18.

Percentage points.

11. Low operating costs translate into higher bank profits, rather than lower interest spreads. Possibly reflecting a lack of competition, efficient banks do not offer more competitive rates, but instead drive home larger profits. This is illustrated in Table III.5, which shows average interest spreads and returns-on-equity by operating cost quintiles. As operating costs fall, profits increase, but interest spreads do not narrow. Only when return-on-equity reaches 30 percent, bank customers start benefiting from lower spreads.

Table III.5.

Sri Lanka: Average Interest Spreads and Profits by Operating Cost Quintiles 1/

article image
Source: Audited Accounts, Sri Lanka Commercial Banks.

2001–2003 averages (or latest 3-year averages available) of 15 commercial banks.

Operating costs to income, percent.

Net interest income to average assets, percent.

Yield on loans and advances minus deposit interest to average deposits, percent.

Banks with negative equity excluded from the sample.

12. Similarly, lower NPL-ratios are not passed on to consumers, but increase bank profits. Banks with lower NPLs need to provision less. In a competitive environment this would translate into lower spreads, but in the Sri Lankan context it increases bank profits. Table III.6, shows a strong negative correlation between NPLs and profits in a cross section of banks. Interest spreads, if anything, increase with lower levels of NPLs.4

Table III.6.

Sri Lanka: Average Interest Spreads and Profits by NPA Quintiles 1/

article image
Source: Audited Accounts, Sri Lanka Commercial Banks.

2001–2003 averages (or latest 3-year averages availbale) of 15 commercial banks.

Nonperforming assets to total assets, percent.

Net interest income to average assets, percent.

Yield on loans and advances minus deposit interest to average deposits, percent.

Banks with negative equity excluded from the sample.

13. Profit margins of efficient banks appear high by international standards. Large profits within a specific sector are another indication of weak competition. Figure III.7 compares the return-on-assets and the return-on equity of some of the most profitable banks across Asia. Sri Lanka turns out to have the second most profitable banks across the region irrespective of the measure of profitability. It is problematic, of course, to compare returns across countries; investors would require higher returns in countries that are more risky. Figure III.7 controls for this factor, by comparing (all) Asian countries with the same country risk, based on the rating of the Economist Intelligence Unit.

Figure III.7.
Figure III.7.

Country Comparison: Average Return of the 20 Percent Most Profitable Banks 1/ (In percent)

Citation: IMF Staff Country Reports 2005, 337; 10.5089/9781451972542.002.A003

Sources: Bank Scope; Audited Accounts, Sri Lanka Commercial Banks; Economist Intelligence Unit.1/Asian Countries with the same EIU country risk rating in 2001–2003. Data for Sri Lanka averaged over 2001–2003. Data for other countries latest year available.

Reserve Requirements and Taxes

14. Statutory reserve requirements cannot explain the unfavorable interest spreads. As shown in Figure III.8, reserve requirements in Sri Lanka are close to the median for a sample of countries. In addition, reserve requirements have been gradually lowered, from 15 percent in 1993 to 10 percent at present and, hence, cannot account for the observed stagnation in interest spreads.

Figure III.8.
Figure III.8.

Country Comparison: Statutory Reserve Requirement, 2004 (In percent)

Citation: IMF Staff Country Reports 2005, 337; 10.5089/9781451972542.002.A003

Source: IMF internal database, 2004.

15. Taxation may have a part in explaining high interest spreads. In 2002, the government introduced a debit tax of 0.1 percent on the value of any withdrawal from a bank account. Monthly debits not exceeding Rs. 20,000 (about $200) are exempted. In 2003, banks became subject to a value added tax of 15 percent. Many countries exempt the banking sector from VAT, because the assignment of value added to the borrower and lender involves high administrative costs.

D. Some Policy Implications

16. In the absence of competition, the problem of high spreads will not disappear on its own. In a competitive market, the efficient banks would offer more favorable rates and over the medium-term force ailing banks to adjust or leave the market. As laid out in Section III, such a laissez-faire approach is unlikely to work in Sri Lanka under the current circumstances. Instead, the government in its role as shareholder and supervisor has to ensure that banks pursue policies that, at least over the medium-term, are consistent with low interest spreads.

17. Reforms in the state-owned banks need to continue to bring down the average spread. The increase in state banks’ profitability and capital is encouraging and shows that the new management teams adhere to firm budget constraints. However, the upward trend in personnel costs will exert further upward pressures on interest spreads. Given powerful unions, both People’s Bank and Bank of Ceylon plan to shrink their staff through natural attrition. However, in the case of People’s Bank, this may be insufficient to bring personnel costs in line with private domestic and foreign banks. The government, as shareholder, should also ensure that the stock problems of the past do not recur, including by allowing management to run the banks on strictly commercial terms. In this context, it is encouraging that state banks have ceased to lend to ailing SOEs without government guarantees.

18. Banking supervision needs to be strengthened, to restore and maintain the health of banks not owned by the government. Despite the haircuts on collateral valuation introduced in January 2004, Sri Lanka’s provisioning guidelines do not conform to best practices. Loans overdue between 3 to 6 months do not have to be provisioned against and full provisioning is required only after 11/2 years, compared to a 1 year international norm. Provisioning in line with international best practices and assuming that the current collateral valuation is appropriate would bring several large banks close to, or below, the statutory capital ratio (in addition to People’s Bank, which currently operates on the basis of a letter of comfort).5 Also, more supervisory power needs to be vested in the CBSL, including through changes in the legal environment.

19. Further research is needed on the causes of weak competition. This chapter has taken a largely agnostic view as to why competition in the banking sector is relatively weak. State-owned banks may have a competitive advantage in attracting deposits due to implicit government guarantees (or explicit ones in the case of the National Savings Bank); information asymmetries between incumbent and contestant banks may allow the former to retain the bulk of high-quality borrowers; or sunk costs associated with the branch network may reduce the contestability of the market. Further research is needed to discriminate between these explanations.

20. However, several measures could increase competition in the short run. The authorities have already taken a useful step by mandating in 2003 that banks publish financial statements on a quarterly basis and display their interest charges and exchange rates for the general public in all bank braches and outlets. However, there is scope for further improvement. The adoption of a single banking license for commercial and specialized banks would do justice to the fact that these banks already compete de facto on several products, such as time and saving deposits. In staff’s view, consistent enforcement of prudential regulations and uniform treatment across banks—supported by more supervisory autonomy for CBSL—would enhance fair competition and administrative justice. Finally, the eventual divestiture of state-owned banks would create a level playing field by removing implicit government guarantees.

21. The government should contemplate measures to bring down the transaction costs of financial intermediation. A central and computerized registry for moveable property would help to collateralize loans and contribute to bring down transaction costs. The authorities may also want to carefully assess the impact that a planned deposit insurance scheme would have on transaction costs. The beneficial effects on competition of deposit insurance could be limited since state banks would continue to be perceived as safer for large deposits, which will not be covered by the scheme. In addition, the insurance may remove some of the discipline currently imposed by depositors on bank risk taking.

Reference

  • Bank Scope. Available on CD-LINK, via JOLIS.

  • Central Bank of Sri Lanka, 2004, Annual Report (Colombo).

  • Central Bank of Sri Lanka, 2003, Annual Report (Colombo).

  • Central Bank of Sri Lanka, Annual Report (Colombo, various issues).

  • Economist Intelligence Unit, Country Risk Service. Available on CD-LINK, via JOLIS.

  • European Central Bank, 2004, Report on EU Banking Structure, November (Germany).

  • World Bank, World Development Indicators. Available on CD-LINK, via JOLIS.

1

Prepared by Erik Lueth.

2

The main data used in this chapter are from commercial banks’ audited accounts. The analysis focuses on commercial banks on the grounds that specialized banks account for only 20 percent of banking sector assets (excluding the central bank).

3

A five-year restructuring plan for People’s Bank was formulated in late 2004. The bank will be recapitalized on a staggered basis based on the achievement of performance indicators.

4

The negative correlation between NPLs on one side and interest spreads and margins on the other may be a statistical artifact. The assets and loans used in the calculation of interest margins and spreads, respectively, include NPLs. If NPLs cease to earn interest, which is to be expected, the interest spreads and margins of banks with high NPLs would be understated. Notwithstanding, the strong correlation between profits and NPLs, both averaged over three years, suggests a lack of competition.

5

The simulation assumes that current provisioning follows CBSL guidelines. To the extent that banks’ provisioning exceeds these guidelines, the downward adjustment of the capital adequacy ratio would be smaller.

Sri Lanka: Selected Issues and Statistical Appendix
Author: International Monetary Fund