90158 August 2014 – Number 131 WHA FRIM: A New Tool for Financial Risk Monitoring in MENA Pietro Calice1 Introduction: The impact of the global financial crisis on banking systems and banks in the Middle East and North Africa (MENA) has highlighted the importance of differentiating across countries and among financial institutions. While the region avoided systemic banking distress, the crisis had a stronger impact on countries in the Gulf Cooperation Council (GCC), where financial systems were more globally integrated and banks more overextended. Countries elsewhere in the region, including transition countries, weathered the crisis better. However, protracted economic weakness is likely to have a significant impact on several banking systems and banks, which can have profound consequences for A sound banking system is critical for access to finance for credit intermediation and ultimately economic growth small and large enterprises. (World Bank Photo Collection) and job creation. This underscores the need for better tools to monitor financial risk and vulnerabilities in To this end, we have developed a simple analytical tool MENA. which allows a preliminary assessment of the health of individual financial institutions in the MENA region. In Most importantly, the global financial crisis revealed particular, we develop a Bank Risk Indicator (BRI) and significant differences in the resilience of individual provide nearly-automated spreadsheet templates, FRIM, banks. This was largely the result of the quality of to facilitate the exercise. We show that the BRI can be management, corporate governance arrangements and useful for initial identification of individual bank banks’ business models, sometimes mitigated by pre- fragility. However, we also note its limitations and emptive policy actions taken by country authorities. For acknowledge the need to complement this analysis with example, the Kuwaiti authorities promptly recapitalized more rigorous quantitative and qualitative assessments. a domestic bank during the recent turmoil. In many cases, knowledge about the specific features Methodology and Data: The starting point for the underpinning individual financial institutions’ health analysis of vulnerabilities in MENA’s banking systems is has been crucial for identifying problems and informing the construction of the BRI, a proxy for insolvency risk policy action. This suggests that an early warning derived from a set of bank fundamental indicators. The system built on bank-specific fundamentals can provide BRI provides an indication of the probability of a useful monitoring tool, helping World Bank Group insolvency of an individual bank. It is calculated for each staff detect vulnerable financial institutions in the region bank in each MENA country at each point in time using to better inform policy dialogue and advisory activities. a multivariate logistic regression framework.2 1 Pietro Calice is a Senior Financial Sector Specialist in the Finance and 2 Markets Global Practice (GFMDR) of the World Bank Group. This K&L See Calice, 2014. “Predicting Bank Insolvency in the Middle East and Quick Note was cleared by Simon C. Bell, Practice Manager GFMDR. North Africa”, World Bank Policy Research Working Paper #6969. Bank insolvency is defined as a state where (car + roa) ≤0, are common in off-site surveillance models employed by where car is the bank’s capital-asset ratio and roa its supervisors. For example, since 1994 the Federal Reserve return on assets. The BRI is estimated by summing up uses an early warning system known as SEER or System five fundamental indicators which are statistically to Estimate Examination Ratings to monitor the correlated with the probability of bank insolvency in condition and performance of state member banks. The MENA. Indicators are taken from individual banks’ model estimates the probability that a bank will fail or income statements and balance sheets. To visually become critically undercapitalized within the next two represent the BRI, “traffic light” icon sets are years. subsequently generated to provide a snapshot of a particular bank (see below). BRIs are generated automatically by FRIM, an Excel- based tool, which downloads the requisite data for 385 The following financial variables are used for calculating individual banks from Bankscope and automatically the BRI: generates the corresponding risk measures3. BRIs are shown as icon sets in the form of traffic lights to easily  Size: log of total assets. This accounts for the fact differentiate among financial institutions in a particular that larger banks may be better able to diversify period and over time. A “traffic light” icon set is a their loan portfolio, both sectorally and graphical representation of data where the individual geographically, thus reducing asset risk and values contained in a matrix are represented as colors. ultimately the probability of insolvency. Typically, the green color is associated with the “best” value; the red represents the “worst” value; whereas  Capitalization: ratio of equity to total loans. A intermediate values are presented in yellow based on measure of the extent of leverage based on loans their proximity to the red zone. instead of total assets provides a sensible indication of the bank’s buffer stock which can The FRIM ranking is calibrated on the historical serve as a cushion to absorb loan losses since probability of insolvency observed in MENA. As the banking crises typically involve shock to the objective of FRIM is to develop an early warning system loan portfolio and MENA banks’ main risk is for bank insolvency, it is important to identify a constituted by lending risk. Higher levels of threshold above which the model would send a signal capital act as a buffer against financial losses that the bank is approaching the insolvency point. and are expected to reduce the risk of bank Therefore, banks whose BRI are above the in-sample insolvency. probability of insolvency will be denoted in red; banks whose BRI is approaching the in-sample probability of  Asset quality: ratio of loans to total assets and insolvency will be represented in yellow; and healthy ratio of loan loss provisions to net interest banks will be denoted by a green light. income. An asset mix characterized by a high proportion of loans relative to other earning Application of the BRI and FRIM: In this section we assets indicates a riskier asset portfolio. On the show the application of the BRI and the output from the other hand, high provisioning reflects high FRIM template. We begin by calculating the BRI of nonperforming loans (NPLs) and therefore MENA banks in 2014 and compare it with the BRI in higher insolvency risk. 2013 to demonstrate the tool (Table 1).4 An overview of the results suggests that:  Management quality: cost to income ratio. Higher values for this indicator signal relatively poor  For the sample of MENA banks where financial managerial quality and therefore higher data are available, the proportion of banks at insolvency risk. risk of insolvency (red light) decreased in 2014 compared to 2013 (from 2.2 percent to 1.6  Liquidity: ratio of liquid assets to deposits and percent). short-term funding. Low liquid assets compared to short-term sources of funding signal a  However, the share of banks approaching the stretched liquidity position which can ultimately risk of insolvency (yellow light) increased in jeopardize the bank’s solvency. 3 For more on Bankscope see bvdinfo.com Bank-level financial indicators are lagged by two years 4 Due to the two-year lag of bank-level financial indicators, the BRI for to generate the BRI in the MENA region. Two-year lags 2014 is calculated using 2012 audited financial statements while the BRI for 2013 is based on 2011 data. August 2014· Number 131 · 2014 compared to 2013, from 3.7 percent to 4.4 Based on end-2011 accounting data, we estimate the BRI percent. for Tunisian banks in 2013 as a first-pass analysis of their financial soundness. The “traffic lights” icon sets which  Tunisia is the country with the largest number are subsequently generated using FRIM shows clear of banks where the probability of insolvency is differentiation in the soundness of banks within the above or close to historically observed values. system. The results are consistent with the findings of the FSAP, notably that SOBs are among the weakest and Next we move to the case study of the Tunisian banking less resilient banks with the system along with a system as a backdrop for testing the usefulness of the smattering of smaller, less systemically important BRI. A joint IMF/World Bank Financial System Stability institutions (Table 2).6 Assessment conducted in 2012 found that the banking sector in the country faced significant challenges due to All in all, the back-tests show the general effectiveness of a weak domestic economy combined with the legacy of the BRI in highlighting banking system vulnerabilities the previous regime. These were reflected in and differentiating the soundness of banks within a deteriorating solvency, asset quality and profitability system. metrics across the board.5 Caveats: As with all synthetic indicators, the BRI Of particular concern was the weak performance of presents a number of caveats which have to be state-owned banks (SOBs) such as Société Tunisienne de considered when interpreting its information content. Banque and Banque de l’Habitat. In particular, their First is the issue of aggregation. The BRI is the sum of financial profile had deteriorated due to weaknesses in five financial indicators which are significant predictors underwriting and deficiencies in corporate governance of bank insolvency in the MENA region. As such, the standards which led to inappropriate lending to well- BRI provides an overall indication of the probability of connected borrowers. Against this background, the insolvency of a particular bank yet it can also hide Tunisia authorities were urged to take preemptive valuable granular information about specific aspects of actions to raise bank capital and a number of initiatives individual banks’ performance. in this respect are underway, including the establishment of an asset management company to Table 2 dispose of NPLs and a recapitalization of SOBs. Tunisia, BRI 2013 Société Tunisienne de Banque Table 1 Banque Nationale Agricole 2013 2014 Banque de l'Habitat Bank 1 Algeria 0 0 15 0 0 15 Bank 2 Bahrain 0 0 25 0 2 23 Bank 3 Djibouti 0 0 3 0 0 3 Bank 4 Egypt 0 0 27 0 0 26 Bank 5 Iran 0 0 13 0 0 10 Bank 6 Iraq 0 0 14 0 0 6 Bank 7 Jordan 1 0 17 0 1 17 Bank 8 Kuwait 0 0 15 1 0 15 Bank 9 Lebanon 0 0 34 0 0 32 Bank 10 Libya 0 0 7 0 0 7 Bank 11 Morocco 0 2 12 0 1 11 Bank 12 Oman 0 1 8 0 1 7 Bank 13 Bank 14 Palestinian Territories 0 0 4 0 0 4 Bank 15 Qatar 0 0 10 0 0 10 Saudi Arabia 0 0 12 0 0 12 Sirya 0 2 13 0 1 8 Second is the issue of in-sample bias. The BRI is Tunisia 5 4 9 2 5 9 calculated based on an empirical assessment of the United Arab Emirates 0 1 26 1 0 26 determinants of bank insolvency within the MENA Yemen 0 0 8 0 0 9 region and during a specific time period. The use of a Total 6 10 272 4 11 250 6 We anonymize private banks to avoid implicating any particular 5 Available at https://www.imf.org/external/pubs/ft/scr/2012/cr12241.pdf. institution. August 2014· Number 131 · global sample to produce the BRI would improve the banks’ financial soundness in the region, and an Excel - signaling value of the indicator. Related to the latter is based spreadsheet tool (FRIM) to facilitate its calculation the issue of cross-sectional comparability. The BRI does and presentation. A back-test based on actual not adjust for nuances associated with heterogeneity developments in the Tunisian banking system suggests across banks. FRIM should ideally be applied in a that the BRI is able to accurately differentiate banks homogenous environment, i.e. to institutions with according to their financial health. similar business models, ownership structures, similar regulatory and supervisory requirements etc. An Still, the signaling value BRI should be interpreted with analysis of the BRI should take these aspects into caution as there are a number of caveats attached to its consideration. use. In particular, the BRI is the result of the aggregation of a number of individual indicators so the relative Finally, given the definition of insolvency underlying performance of the latter should also be taken into the assessment analysis, the BRI is generally able to account in any analysis. Moreover, the BRI is calculated predict the occurrence of an insolvency event only when based on an assessment of the health of individual banks the latter originates from widespread losses on the asset within the MENA region, which implies some sort of side of a bank’s balance sheet, which leads to a selection bias. The differences in business models and progressive deterioration of the bank’s fundamentals. other intrinsic bank characteristics should also be Insolvency events, however, can also originate from the considered when interpreting the results. In addition, liability side of a bank’s balance sheet, such as a the definition of insolvency implied by the BRI is such traditional bank run or a wholesale bank run, as the that it is not able to signal the risk of insolvency recent crisis showed on a large scale and in GCC stemming from other-than deterioration in the bank countries as far as MENA is concerned. These point to fundamentals. Finally, it is important that the user has the importance of supplementing the signaling value of familiarity with the specific features of any banking the BRI with any other relevant quantitative and system under analysis and that any assessment is qualitative information. supplemented by any other relevant quantitative and qualitative information. Concluding Remarks: Recent episodes of global financial stress have underscored the importance of individual banks to the stability of their own or even the global financial system. Therefore, the analysis of the soundness of individual institutions based on their fundamentals has been established as a central component of surveillance and crisis management frameworks. This is especially important in the MENA region, which has shown significant differences in the relative health of financial institutions. Against this backdrop and to respond to the need of World Bank Group staff for a wider diagnostic toolkit, we have developed the BRI, an icon set in the form of a traffic light to enable preliminary analyses of individual August 2014· Number 131 ·