At the end of 2022, banks provisioned between 10-30% for USD-denominated government securities (ISBs and SLDBs), though this figure could rise to 60% by the end of 2023 if the true loss is to be reflected
Financial stability and credibility cannot be achieved by communicating only intentions, nor by avoiding difficult discussions, but only by implementing concrete policy mechanisms
Create and capitalize a Financial Stability Fund that is authorised to lend and inject capital to banks rapidly in case of an imminent bank failure
Sri Lanka is entering into a new, dangerous phase in its economic recovery. Although macroeconomic stabilisation has made progress, financial stabilisation needs immediate attention. The combination of covid-era loans and crisis-era loans with high-interest rates all coming due, and in an environment with volatility and uncertainty in loan quality, it is important that policies be put in place now, well before any crisis emanates from the banking system.
The current neglect of this arises from the on average 3.3% buffers over the required ratio of banks. Although these buffers, on average, are sufficient to absorb broad-based loss-sharing from restructuring domestic debt, there are two significant risks. First, the provisioning of non-performing loans according to pre-crisis metrics and the need for additional provisions to International Sovereign Bonds (ISBs) and Sri Lanka Development Bonds (SLDBs) on bank balance sheets.
If the impairment coverage rate of non-performing loans (NPLs)/Stage 3 loans reaches 60%, as is the case with most major Russian banks currently, then the buffer of the Tier 1 ratio over the stipulated regulatory minimum capital thresholds will be close to zero. If we consider the effect of all Stage 2 loans being categorized as Stage 3 (a portion of this is to be expected as firm distress takes several months to materialize into persistent loan delinquencies), an additional coverage of 40% would be required on these loans and the capital buffers become negative. Whilst the risk will be different across banks, a systemic event will mean that all banks will be impacted. Macroprudential measures are needed. Our research analysis focuses on a cohort of nine domestically significant commercial banks, introducing additional impairment coverages based on three scenarios: (1) Impairment coverage of 60% on stage 3 loans, which we identify as the general case. (2) Additional 30% impairment coverage on USD-denominated Government securities. We assume that the banks are currently maintaining at least a 30% provisioning on USD-denominated Government securities. (3) Stage 2 loans being categorized as Stage 3. The current impairment coverage stands around an average of 10% and 47% on Stage 2 and Stage 3 loans, respectively. In line with the Expected Credit Loss calculation framework adhered to by banks, there exists a subtle demarcation between the categorization of loans as stage 2 and stage 3, determined by their respective payment due schedules. We consider all these scenario-based augmentations and calculate the impact of buffer capital banks are maintaining.
At the end of 2022, banks provisioned between 10-30% for USD-denominated government securities (ISBs and SLDBs), though this figure could rise to 60% by the end of 2023 if the true loss is to be reflected. The effect of this provisioning will bring significant negative capital ratios in many banks.
The total potential effect on Tier 1 capital ratios is stark and requires immediate attention. Since early 2022, I have been arguing that financial stability needs to be taken seriously with concrete policy measures. “Debt restructuring and a path to confidence and hope”, 14 January 2022, DailyFT (Sri Lanka).
“Sharing the cost of the crisis: The need to restructure domestic debt together with an ‘Asset Management Company”, 19 April 2022, DailyFT (Sri Lanka).
The desirability of domestic debt restructuring, October 2022, Verite Research Background Note,
“Restructuring Domestic Debt Must be Fast, Credible, Comprehensive”, 8 June 2023, Daily Mirror (Sri Lanka): Link here] These include:
Quadruple the deposit guarantee limit from LKR 1.1 million to LKR 4.4 million. Increase of the Maximum Compensation Payment under the Sri Lanka Deposit Insurance and Liquidity Support Scheme (SLDILSS). Central Bank of Sri Lanka (cbsl.gov.lk)] to reflect the depreciation of the rupee, inflation and create a significant threshold above which most depositors will not panic. The Russian central bank doubled deposit guarantee limits in 2014 and has not experienced any significant domestic financial instability event since.
Create and capitalize a Financial Stability Fund that is authorised to lend and inject capital to banks rapidly in case of an imminent bank failure. For example, the government of Ghana has committed USD750 million to a similar fund, and the Greek government established the Hellenic Financial Stability Fund in 2010.
Establish an Asset Management Company (AMC) or “bad bank”. It is a special purpose vehicle that buys non-performing loans from bank balance sheets, thereby removing the uncertainty (and need for precautionary provisioning), as well as allowing longer periods of time for restructuring of the loans. This is particularly important given the large amount of non-performing loans to state-owned enterprises on bank balance sheets. [Banking sector exposure to the Central Government and SOEs amounted to rupees. 7.7 trillion and represented 40.1% and 512.2% of total assets and capital funds of the banking sector, respectively, at end September 2022].
Financial stability and credibility cannot be achieved by communicating only intentions, nor by avoiding difficult discussions, but only by implementing concrete policy mechanisms. Failure to implement these policies brings serious risks to the recovery of the country. The vulnerability of the banking system may be exacerbated when leverage and lending increases in the months that follow due to improved macroeconomic conditions and loosening of policy. On the other hand, if the risks outlined in this article are not of significant concern to warrant action, the argument that the banks needed to be precluded from DDO (Domestic Debt Optimization) becomes puzzling: it cannot be that we have “Schrodinger’s Banks”, banks that are too fragile to absorb DDO but not fragile enough to take the standard precautionary measures that were taken in other countries in similar situations.
Prof M. Udara Peiris is an Associate Professor of Economics at Oberlin College and a Global Academic Fellow at Verité Research. He can be reached via firstname.lastname@example.org.