ANNEX Observance of Financial Sector Standards and Codes: Summary Assessments
This section contains summary assessments of Moldova’s adherence to the Basel Core Principles for Effective Banking Supervision and the Code of Good Practices on Transparency in Monetary and Financial Policies. The detailed assessments have been undertaken in May-June 2004 as part of the Joint World Bank/IMF FSAP and have served as a basis for the summary assessments.
The assessments have been used to gauge the risks and vulnerabilities as well as the developmental needs of the financial sector. For the areas covered by the assessments that have not yet been fully aligned with these international codes, the FSAP team made recommendations to the authorities, who expressed their willingness to implement these suggestions as soon as feasible and hence to further strengthen the institutional and legal basis of the financial system.
Moldova is making considerable progress in observing the international standards and codes that were assessed. The NBM, in charge of banking supervision and monetary policy, also strives for transparency and openness in its policies and procedures. The overarching issue in the two standards assessed is the lack of proper protection of central bank staff for actions taken in the conduct of their official duties.
In banking supervision, assessors noted the need for further strengthening of regulations with respect to the transparency in ownership and the possibility for the NBM to conduct fit and proper evaluation of owners of banks. They recommend requiring that banks conduct a thorough examination of the ownership relations of their clients. Regulations are also needed in the areas of country, market, and reputation risk, as well as for deduction from the capital of lending to connected parties. Supervisory practices should develop further in the direction of risk-oriented supervision. Finally, steps need to be taken—including laying down the legal basis—to undertake consolidated supervision, both domestically and globally.
In general, the NBM practices a fairly high degree of transparency in most aspects of its operations, including banking supervision. The main areas for improvement relate to more clearly spelling out the primary goals of monetary policy and the conditions and limitations on the direct financing of the government by the NBM
APPENDIX: Stress Tests on the Moldova Banking System
70. Sensitivity analyses on the Moldova banking system revealed credit risk to be the major source of vulnerability. Stress tests included:
(i) a deterioration in credit quality due to shocks to external demand for agricultural products (owing to the country’s high dependence on export of agricultural products);
(ii) an increase in oil and gas prices that has a direct effect on credit portfolios of banks exposed to this sector, and an indirect ripple effect on banks’ exposure to the corporate sector;
(iii) an appreciation of nominal exchange rates (as the inflow of foreign workers’ remittances has a strong effect on exchange rates) with additional credit quality implications;
(iv) direct and indirect (credit quality) effects of an increase in domestic lending rates;
(v) a scenario with a combination of interest rate and exchange rate risks; and
(vi) the risk of illiquidity of government securities in anticipation of liquidity risks.
Among market risks, the banking system seems more vulnerable to high (yet plausible) rates of exchange rate changes, rather than to interest rate movements through deterioration in credit quality. In conducting the stress tests, balance sheet data on all 16 banks for end-December 2003 have been used; the results are summarized in Table 4.
71. In assessing individual banks’ vulnerabilities to various types of credit risk, the tests assume a loan migration from higher to the lower credit-classification categories in response to a shock—the corresponding increase in loan-loss provisions subtracted from both regulatory capital and risk-weighted assets yields the new (adjusted) CAR. Given the high baseline CAR (Table 4)8, a significant effect of a shock is perceivable only if we assume some migration from category A (‘standard’) and a substantial migration from category B (‘under-supervision’), to categories C, D, and E. The first credit risk stress test pertains to an overall shock that deteriorates the quality of loan portfolios of all banks equally. In particular, the resulting NPL (as a percent of gross loans) is assumed to match the average NPL experienced during the 1998-1999 currency crisis—31 percent of gross loans on an average between September 1998 and December 1999. For purpose of the subsequent tests, this generic credit risk will be considered as a second round (and, in most cases, the maximum indirect) effect of an interest rate or an exchange rate shock. In order to achieve an NPL of 31 percent, it is assumed that there is a decline in loans (in each bank) from category A by 24 percent, and from B by 30 percent (with corresponding increases in C, D, and E to hold total loans constant). As a result of this loan-migration, and the subsequent increase in provisions, two banks (one of them holding 14 percent of total assets of the banking system) move below the minimum CAR of 12 percent. The CAR of the overall banking system decreases by 8.6 percentage points, but the banks in group 3 (except for one) are unaffected by this shock owing to very small credit portfolios.
72. Two sectoral credit risk tests were conducted: (i) a shock to external demand for agriculture products assumes a migration of agricultural loans from categories A and B by 24 percent and 30 percent respectively, to categories C, D, and E, as a result of which, the NPL of the banking system rise to 13 percent of gross loans; and (ii) a fuel price shock assumes a primary round of migration of energy-sector loans from categories A and B by 39 and 50 percent respectively, and redistributed in C, D, and E, and a secondary round migration of overall corporate loans by 24 and 30 percent from A and B respectively to the other categories (using the generic credit risk assumptions mentioned earlier), which results in an NPL of 22 percent for the overall banking system. Sectoral credit risk tests indicate that the banking system is vulnerable to fuel price shocks, but resilient to moderate agricultural shocks. For the fuel shock, the asymmetric migration of loans considered above pushes two banks (one of them systemically important) below the 12 percent minimum CAR.
73. Exchange rate risk is the effect of exchange rate changes on the local currency value of banks’ assets and liabilities as well as off-balance sheet items. The direct balance sheet effects of exchange rate shocks are limited by modest open foreign exchange positions that are well within the regulatory limit. However, owing to a large share of credit extended in foreign currency, the banking sector is exposed to indirect effects of credit quality deterioration due to sudden large exchange rate shocks. Credit quality of exporters will worsen with real appreciations. With NBM regulations restricting foreign exchange lending to importers—who are not naturally hedged with income in local currency—credit quality of the latter would tend to deteriorate following depreciations. It is assumed that, for the purpose of the sensitivity analysis of exchange rate movements, banks are equally exposed to exporters and importers—for both exchange rate appreciations and depreciations, the effect on credit quality is symmetric. The banking system is vulnerable to large—for example, 40 percent—exchange rate appreciations and depreciations only when accompanied by higher credit risks. Asymmetric migration of loans pushes two banks below the minimum CAR, reducing banking system CAR by almost 10 percentage points.
74. Changes in interest rates affect net interest income (NII) that in turn influences profits, and through retained earnings, capital of the banks. An additional effect of interest rate changes is on credit quality—an increase in nominal interest rate is assumed to have a real effect on the solvency of borrowers from the banking system. Historical data (post 1998-99 crisis) suggest that interest rate changes have a stronger effect on NPLs than do exchange rate changes, even though the size of interest rate shock considered is much smaller. The banking system has positive repricing gaps (the difference between values of assets and liabilities to be repriced at each horizon) for most horizons—an interest rate increase would tend to raise CAR for the overall banking sector due to higher NII. The negative repricing gaps, which a few banks face, are small and do not result in significant capital losses in case of interest rate increases. The banking system in Moldova is not considered to be very vulnerable to interest rate shocks, unless it is accompanied by very large credit quality effects. In such a case, overall CAR reduces by 8.5 percentage points with two banks falling below the minimum CAR.
75. A combination of different types of shocks was considered to be a result of a crisis scenario—sudden depreciation of 40 percent of the USD and euro exchange rates, and 20 percent of that of the Russian ruble prompted by presumably a sudden decline in the flow of foreign workers’ remittances, accompanied by an increase in borrowing rates. Assuming NPLs to rise to the maximum level experienced during the 1998-99 crisis—39 percent NPL achieved by 32 and 40 percent migration of loans from A and B respectively into C, D, and E—the CAR of the banking system would decline by 11 percentage points, and two banks would fall below the 12 percent minimum CAR. If an additional effect of illiquidity of government securities in anticipation of a liquidity crisis is included in the scenario, the total effect will be very significant. Given the thin market for such securities and the large holdings by banks, the price of such securities would decline eroding considerably the realization value of such assets. Assuming that banks are able to recover only 50 percent of the value of such securities, the banking system would lose an additional 5.4 percentage point of CAR—that of group 1 would decline to 9.2 percent. Moreover, 6 banks, 4 in group 1, and 1 each from groups 2 and 3 would be pushed below the 12 percent minimum.
These owners are domiciled in Bahamas, Belize, Cayman Islands, Cyprus, Dominican Republic, Panama, etc.
The foreign currencies used are mostly the euro and the U.S. dollar.
The authorities, however, maintain that these banks are financially responsible for their subsidiaries only to the extent of the equity put up by them.
This recommendation is more stringent than normally required, but is justified by the lack of proper insurance supervision coupled with the large number of existing insurance companies, which put in doubt the soundness of the banks’ full investment in the insurance sector.
From July 2004, the NBM introduced foreign currency reserve requirements for foreign currency deposits, at the same time decreasing the rate from 12 percent to 10 percent. The process of switching reserve currencies will take place gradually, by shifting one percentage point every month and completing in 10 months, so as not to release additional leu liquidity at once.
The full assessment of AML/CFT, which is required in the context of the FSAP, is scheduled to be conducted in early 2005 by MONEYVAL, the FATF-type regional body for Europe. MONEYVAL will prepare an assessment of compliance with standards and codes following their mission.
CAR is high because some banks function as “pocket” banks, as described in the main body of the report and have a very low deposit base.