Gunter Deuber and Marion Muhlberger of DB Research -

Key views

-- Asset quality deterioration in Central and Eastern Europe to continue until at least mid-2010. As NPL ratios usually lag the business cycle by at least 6 months and the credit boom will continue to unwind, non-performing loan (NPL) ratios are expected to reach their peak levels only next year.

-- Peak NPL levels to reach 10-50%. Based on IMF data of previous episodes of credit booms in emerging markets and country-specific trends we come up with the following estimates for peak NPLs: 5-10% in the Central Europe-3, 15-25% in the Baltics, 15-20% in Southeast Europe, 25-30% in Russia, 35-40% in Kazakhstan and 30-45% in Ukraine.

-- Further capital injections likely. As highlighted by central bank stress tests the sharp deterioration in asset quality will lead to recapitalisation, acquisition of equity stakes by the EBRD, mergers and potential increases in state ownership.

Focus shifts from funding risks to asset quality

Concerns about a retrenchment of Western parent banks or an external funding crisis have eased over recent months. But now almost all CEE countries have entered recession, leading to sharply rising unemployment, declining property prices and lower incomes. As a consequence, asset quality has become the main factor to watch. As loan quality lags the business cycle and the credit boom will continue to unwind (see chart 1), we expect NPL ratios to continue rising over the next few months.

Forecasting peak NPL levels: If history is any guide...

Variables like unemployment, real interest rates, debt levels, income and the output gap have been proven to explain NPL levels in empirical models. But in the case of CEE, methodological changes, structural breaks and short time series make it very difficult to forecast "peak NPL levels" based on those variables. IMF data on past episodes of credit booms that were followed by banking sector stress provide at least rough guidance of how far NPLs might rise (see chart 2). Having those historical levels in mind we now take a look at country-specific trends.

Baltics close to peak: Credit quality is deteriorating among all loan categories. In Estonia, NPLs started to rise already in 2007. In the past, NPLs in Estonia followed recessions and unemployment without a time lag so that we expect them to be close to their peak level around 15% by now. In Lithuania and Latvia we expect NPLs to peak at 15-20% and 25%, respectively, in late-2009. The rise in NPLs has been limited in Czech Republic and Hungary, while NPLs might peak above the 5% level late in 2010. Poland has experienced the strongest relative rise in overall NPLs among the CE-3, which is worrisome as credit also continues to grow strongly. Moreover, NPLs in Poland have been mainly hit by credit quality deterioration in the corporate portfolio (40% of total loans), while the threat of deterioration in the household segment is increasing on the back of the slowing economy. Therefore, NPLs might peak around the 10% level in late 2010/early 2011. Moreover, there is evidence that banks are restructuring bad loans pre-emptively in the CE-3, and that a sale of problematic loans took place in Hungary.

Southeast Europe consumer credit boom starting to backfire: In Romania and Bulgaria GDP and credit growth only started to slow down significantly at the beginning of this year. Thus the rise in NPLs has not been too strong yet. But given the high share of consumer loans in total loans (40-50%), the high debt burden of households (30% of monthly income) and the severity of the economic downturn (GDP growth to drop from 7% to -8% in Romania and 6% to -5% in Bulgaria), we expect asset quality to deteriorate sharply over the next 12 months (ie. a NPL ratio of 15-20%).

CIS "evergreening" masks asset quality deterioration: Kazakhstan was hit by the crisis as early as August 2007, but NPLs saw the steepest increase only in May this year. This is due to several reasons: First, some banks are in debt restructuring talks and thus had to disclose their portfolio quality. Second, the February 2009 tenge devaluation hit unhedged borrowers. And third, real estate prices declined sharply in 2008. S&P estimates that loans under stress will continue to grow and may reach 50% of private sector credit over the next 12 months. The Russian 4.4% NPL ratio has to be interpreted carefully. It only reflects missed payments, as banks are allowed not to report the full amount of the loan outstanding until the maturity date. In addition restructured loans are also not reflected in NPL figures. Thus, the "real" NPL ratio may be at least 5% higher as some bad loans are rolled over to postpone their appearance on the balance sheet ("evergreening"). Estimates for "real" peak-level NPLs are around 30%. In Ukraine NPLs are rising rapidly as the 50% hryvnia depreciation has hit unhedged retail customers. Moreover, the severe output slump (close to 20% GDP drop in first half) has started to affect loan quality in the corporate sector, which accounts for two-thirds of total loans. According to the IMF's definition NPLs already reached 24% in March. Thus, despite large-scale loan restructuring processes going on, NPLs could even reach past crisis levels of above 50% over the next few quarters.

Further capital injections likely

Stress testing by CEE central banks show that solvency ratios would decline close to or even below regulatory minimum-levels in case of substantial asset quality deterioration (ie. NPLs of 20% or higher). Thus we expect (further) capital increases, both for foreign and domestically owned banks. In the CIS, stress tests on an individual bank level are not publicly available yet. According to the IMF, Russia's central bank officials "do not see the need for systemic stress tests of banks". In Ukraine a more detailed and timely banking sector monitoring is to be launched in September. In the CIS, we expect deteriorating asset quality to lead to bank mergers (especially in Russia), recapitalisation, acquisition of equity stakes by the EBRD, the establishment of distressed assets funds (à la Kazakhstan) and, as an ultimate measure, an increase in state ownership. Moreover, in the hardest hit countries of the whole region a more centralised way of restructuring private sector debt (eg. generally binding haircuts instead of bilateral agreements) should help to increase transparency.

Send comments to The Editor