Banking Reform Helps Reduce Budget Burden

According to the Vietnam Development Report 2006, the surface, the repayment rate of credits is strong in Vietnam. But this is to a large extent due to loan rescheduling. For instance, Vietnam Bank for Agricultural and Rural Development (Agribank) does not count frozen loans as nonperforming, or overdue. Admittedly, as the seizure of collateral is difficult in practice, the rescheduling of non-performing loans (NPLs) is at times the only alternative left to banks. But then the mere servicing of interest payments becomes then a poor indicator of their quality. Differences in the treatment of rescheduled loans are one of the main reasons for the wide dispersion of estimates of NPLs in Vietnam.

 

One frequently used approach counts as nonperforming any loan whose service is overdue by 90 days or more, as well as any loan which is overdue by less than 90 days but has been rescheduled. At the risk of simplification, this is the approach underlying a series of independent audits conducted since the year 2000 on SOCBs, using international accounting standards (IAS). At the other end, Vietnamese accounting standards (VAS) ignore rescheduling. The main consequence of this difference in their underlying approaches is a widening gap between IAS and VAS estimates of NPLs.

 

The issuance of Decision 493, in April of 2005, is an important step towards bringing Vietnam’s loan classification and provisioning practices closer to international standards. Decision 493 is indeed in accordance with IAS, although it will take three years to move from quantitative assessments (based on the number of days the servicing of a loan is overdue) to more elaborate, qualitative ones. According to this decision, banks need to provide 5 per cent of the value of loans which are not current but are less than 90 days overdue. The provisioning rate increases to 20 per cent for loans which are overdue by 90 to 180 days, to 50 per cent for loans overdue up to 360 days, and to 100 per cent above the one-year threshold. Loans which are being properly serviced but have been rescheduled are treated as if they were 90 to 180 days overdue. Provisioning rates aim to capture the expected loss (in probabilistic terms) for loans in each category. They may not really match the recovery rate that can be secured in Vietnam, which in turn may require that Decision 493 be revised from time to time. But they are based on international experience, and can therefore be used to produce a rough estimate of the losses the four main State-owned commercial banks (Agribank, VCB, BIDV and ICB) can expect.

 

The initial application of this decision yields a share of NPLs which is much higher than under the old VAS, although not as high as the share estimated on the basis of IAS.

 

According to economic experts, Decision 493 is a step in the right direction. Applied consistently, it should exercise considerable pressure on State-owned commercial banks to improve the quality of their lending. Such pressure was diluted under the much more lenient VAS approach.

The direct fiscal burden from NPLs is represented by the gap between the resulting figure and the actual provisioning by State-owned commercial banks. Applying the provisioning rates established by Decision 493 to this estimated structure leads to an expected loss equivalent to 13.3 per cent of outstanding credit. Note that this is less than the share of NPLs in outstanding credit, reflecting the expectation that not all NPLs will be a total loss. This average provisioning rate was also applied to off-balance sheet items, assuming that such items amount to 15 per cent of the figures captured by the IAS audits. And to err on the side of caution, the total provisioning (including both balance sheet and off-balance sheet items) was increased by 50 per cent.

This is a considerable sum, which cannot be mobilized in just one or two years, even under a determined effort to reform the banking system. It is more reasonable to assume that the government will have to issue bonds and transfer the resources to the State-owned commercial banks as they restructure. This in turn, would increase the overall public debt of Vietnam. A debt sustainability assessment jointly conducted by the IMF and the World Bank tried to estimate by how much. It assumed that the issuance of government bonds to support State-owned commercial bank reform would be spread over the five-year period. The assessment also considered that the economy would continue growing as it did in recent years, at about 7.5 per cent per year. Macroeconomic management was supposed to remain prudent, with the budget deficit being gradually brought down, until reaching 1.5 per cent of GDP by 2010. Under these assumptions, public debt would increase from 41.2 per cent of GDP at present, to 48.7 per cent in 2010. By contrast, if State-owned commercial banks had been performing well, and not generated fiscal liabilities, the stock of public debt would remain almost stable, reaching 42.4 per cent of GDP in 2010.

 

These numbers are to some extent deceiving, because a considerable portion of Vietnam’s public debt is in concessional terms. This means that the net present value (NPV) of the long-term obligations faced by the government is considerably less than its face value. On a present-value basis, the burden of SOCB restructuring would increase public debt from 28 per cent of GDP at present to 35.2 per cent in 2010. This figure is high, but it remains manageable. Delaying SOCB reform, on the other hand, could result in a much heavier burden. Even maintaining a stable share of NPLs in total credit would translate into an increasing fiscal liability, because credit is growing much faster than GDP.