Ranking methodology for Islamic banking sectors – modification of the conventional CAMELS method

The state of banking systems is an important issue. The purpose of this paper was to test whether the well-known CAMELS microeconomic methodology, generally used for ranking banks, is applicable to evaluating Islamic banking systems. The hypothesis was tested by implementing a method for a particular case, public, free data – from 2013 till the first quarter of 2018 – on Islamic banking systems from the “Islamic Financial Services Board” (IFBS) database. As expected, modifications were necessary. First, because of the lack of data (in Islamic databases, no data refer to the management (“M”)), and second, to avoid the subjectivity of the five-degree method and to reach more sensibility. Thus, a hundred-level (standardized) rating system was introduced – “CAELS 100”, where “100” refers to the levels. The other part of the methodology – creating a simple average of the (now level 100) rating of raw indicators to get the letters of CA(M)ELS in the relevant period – remained unchanged. After the data cleaning, only six countries (Bahrain, Egypt, Kuwait, Oman, Turkey, and the United Arab Emirates) were able to participate in the analysis.The result showed that Egypt, Turkey and Kuwait were the best ones respectively. Thus, it was concluded that this “CAELS 100” methodology is suitable for evaluating Islamic banking systems.
AcknowledgmentThe research was supported by the project “Intelligent specialization program at Kaposvár University”, No. EFOP-3.6.1-16-2016-00007.


INTRODUCTION
The rating of bank systems is an important issue. There are techniques for that, but most of them were developed for conventional banks and have not been used to rate Islamic bank systems.
The interest-free banking system began about fifty years ago when the first Islamic bank was founded in Dubai in 1975. This type of banking has widely spread to several non-Arabic (Pakistan, Malaysia, Indonesia, Turkey) and even non-Islamic countries like the USA and the UK (Karapinar & Dogan, 2015). One of the largest markets for Islamic finance is in Indonesia. In 1992, Bank Muamalat was established and the government improved banking regulations there.
After the 2008 financial crisis, more attention was paid to Islamic banking, as these banks had almost no 'toxic' assets as they run safer operations than conventional banks (Széles, 2015). The research question of this paper is whether the CAMELS microeconomic bank rating methodology is suitable for evaluating Islamic banking systems.

LITERATURE REVIEW AND ANALYSIS
The topic of Islamic banking is still poorly represented in the European literature. Similarly, in Islam countries, there are few publications referring to "conventional" banking.
Islamic banks must operate under the Islamic principles of Sharia'h rules, paying interest is prohibited. According to Islam, money is just a simple instrument, it has no value by itself. It is merely used to measure the value of things as the principles of the Muslims' holy book the "Holy Quran" and "Sunnah" tell. Islamic finance emphasizes partnership and cooperation. The institutions, firms, and tools base their operations on interest-free transactions and profit and loss sharing. The parties share the risks, returns, and losses. Tabash and Dhankar (2014) pointed to the double importance of Islamic banking that comes from the remarkable growth and stability during the crises.
The Islamic banking sector is dynamically increasing. The data from the free database of the Islamic Finance Service Board (IFSB) show the growth rate between 2013 and 2018 is about 50%, considering total assets (Table 1). A similar tendency can be found for other indicators in the same table.
CAMELS -the methodology intended to usewas introduced in 1979 by the US banking supervisors to analyze the financial performances of banks. It was adopted by the North America Bank to know the financial and managerial reliability of commercial lending institutions. There are several other techniques for analyzing banks' performance, but this is the most spread-up one, according to the literature (Baka et al., 2012). It is "a useful tool to examine the safety and soundness of banks, and help mitigate the potential risks, which may lead to bank failures" (Dang, 2011, p. 2) even after the banking crisis (Dang, 2011, p. 16 CAMELS is a subjective grading method that uses six criteria, the acronym comes from Capital Adequacy, Asset Quality, Management, Earnings, Liquidity, and Sensitivity to Risk. This model as- sesses the overall condition of a bank, its strengths and weaknesses. The composite ranges of the CAMELS rating system consist of five groups: • Rating 1 (composite range 1-1.49): is a strong position, good working in every respect, resistance to external economic and financial disturbances, no cause for supervisory concern.
• Rating 2 (composite range 1.5-2.49): shows a satisfactory position, is stable and can withstand business fluctuations well, supervisory concerns are limited to extent that findings are corrected.
• Rating 3 (composite range 2.5-3.49): fair position, financial, operational, or compliance weaknesses ranging from moderately severe to unsatisfactory, easily deteriorates if actions are not effective in correcting weakness.
• Rating 4 (composite range 3.5-4.49): marginal position, the immoderate volume of serious financial weaknesses, without correction, high potential for failure, without correction these conditions could develop further and impair future viability.
• Rating 5 (composite range 4.5-5.00): unsatisfactory position, high immediate or near term probability failure, without immediate corrective actions, liquidation is likely to be lost.
The literature review proved the CAMELS methodology can be implemented for ranking Islamic banking sectors of countries.

HYPOTHESIS AND METHOD
The hypothesis of this paper is whether the CAMELS method, after modification, fulfils the needs for evaluating Islamic banking systems based on data from the ISFB database.
The methodology for testing the hypothesis will be the implementation of the (modified) method for a particular case. The test data come from the free, public IFSB (Islamic Financial Services Board) database -referring to the time period from 2013 till the first quarter of 2018. Rating the Islamic banks' systems of the available countries, will be a newby-product -result.

Relationship between the CAMELS and IFSB indicators
CAMELS has been invented for conventional banks, but the aim of this study is to investigate the systems of Islamic banks. The difficulty occurred because the indicators are not the same in these two banking systems. All indicators of the IFSB database will be presented, but only those that can participate in a CAMELS-type analysis are described in detail. The names of the indices will remain original, thus, the numbering of the indicators in the analysis will not be monotonous ( Table 2).
Capital Adequacy Ratio (C) measures the safety and stability of banks. The equity capital shows the financial situation of a bank and allows one to write off losses if something goes wrong. CAR determines the ability of a bank to meet the obligation on time and other risks such as credit risk, etc. All core indicators correspond to equivalent IMF Financial Soundness Indicators (FSIs), except for Net Profit Margin and the Cost to Income ratio, which are commonly used banking indicators (IFSB, 2019b).
In most countries, the calculation of the capital adequacy ratio is regulated according to the Basel (I, II and III) recommendations. According to Basel II, the capital covers three types of risk, namely: credit risk (the risk of loss due to a counterparty defaulting on a contract), market risk (the risk of losses on on-and-off-balance sheet positions arising from movements in market prices, interest rates, and exchange rates), and operational risk (the risk of the non-perfect operation of the banking system).
For Capital Adequacy, both indicators are directly proportional (it is denoted by "+" after the short name of the indicator.) Asset Quality (A) is the second area of a CAMELS analysis. Its main area is lending quality. Lending activities are particularly important for banks, so it is essential to analyze the quality of assets in terms of a bank's successful operation and efficiency. Classified loans, especially non-performing loans (NPL), indicators in conventional banks, and non-performing financing (NPF) are mainly analyzed in Islamic banks. The NPL ratio provides information about the level of non-performing loans in the total loan portfolio. Non-performing assets are usually bad debts that are in default or near to be in default.
All Asset Quality (A) indicators are inversely proportional -denoted by "-" after the short name of the indicator.
The evaluation of the Management (M) for conventional banks is mainly based on the share price and the income-cost ratio of the relevant bank. Theoretically, it is possible to collect this information for every bank and take part in the analysis. However, given the bank systems of countries, this technique is impractical, unachievable due to the huge number of banks and the predicted lack of data.
In the IFSB database, there is no official indicator referring to Management. A possible methodological explanation can be the following. The role of the management and the attitude of customers, owners -there are many governments owned, supported banks -is different in Islamic banks. Also, the share price consideration (inter alia because of religious causes) is also different in Islam. This way, the performance of management in the two banking systems is really incomparable.
Some information about the management is involved in other indicators, like capital adequacy and earnings. Thus, even if letter M is avoided, the ranking is based on the performance of the management as well.
Earnings (E) is necessary for banks to generate sufficient earnings to stay in the market for a longer period. The profitability indicator refers to the management effectiveness. Return on Equity (ROE) shows how equity produces profit. It shows the efficiency, profitability of a bank, how efficiently the bank uses its capital. Return on Assets (ROA) gives information about a bank's assets. ROA avoids the volatility of earnings linked with unusual items and measures the bank's profitability.
The Net profit margin is equal to a bank's total interest income minus total interest expenses. The cost-to-income ratio is calculated by dividing the operating expenses by the operating income generated i.e. net interest income plus the other income. Three indices of Earnings are direct till the third one inverse proportional, as the lower the cost to income index, the better the operational efficiency of the bank.
Liquidity (L) is the ability of a firm to convert its financial assets into cash most rapidly or in quick succession. The indicators in the group of liquidity answer the question of how much a bank can fulfill its short-term liabilities using its current assets. The liquidity indicator shows how fast a bank's financial instruments can be converted to cash without losses. The liquidity indicators give information to what extent it can meet its shortterm liabilities with short-term assets. The higher the index value, the more liquid a bank can be considered. The liquidity rate was counted by using cash, central bank deposits, loans to other banks, and the sum of securities compared to the balance sheet total. There is not enough information on the indicators of the new Basel III system, LCR (liquidity coverage ratio), and NSFR (Net stable funding ratio) for the relevant period of 2013-2018. Both liquidity indicators are directly proportional, since the growth of liquidity means an upward trend.
Sensitivity to risk (S) consists of the interest, operational and financial risks, like changes in interest rates, foreign exchange rates, and prices. It affects a bank's earnings. Of course, in the case of Islamic banks, there is no interest risk. In the IFSB database, there are three indicators for this field. CP17 refers to the Net foreign exchange open position to capital (see below). CP18 refers to the "Large exposures to capital" and CP19 for the Growth of financing to the private sector for sensitivity to risk. Due to the lack of data, the last two had to be deleted and only the first CP17 remained. This index is inversely proportional, since the lower it is, the better the bank's position.

CAELS 100 new methodology
At the beginning of the study, it became obvious that the original CAMELS methodology should be modified and somehow improved for the following reasons. First, there were no data for the performance of the management in the IFSB database -thus, the "letter M" had to be left out, as was explained earlier. The second reason was the lack of sensibility, the third one -the method allows subjectivity. The last reason is a well-known property of the CAMELS analysis, which might be an advantage, if the evaluator wants to add some subjectivity, but a disadvantage when objectivity is the target of research. Given the proportionality of the criteria, a hundred-level evaluation was introduced that can be considered ratio (percentage, %) or standardization. It avoided the subjectivity and solved the lack of sensitivity problem. If a variable is directly proportional, the maximum got 100, the minimum 0, and vice-versa -if inverse proportional, the minimum got 100 and the maximum 0, or the composite indicators were the simple mathematical average of these standardized values created. To handle the remaining lack of data situation, "values of the letters" were created using the adaptive average technique. Available data were used and the missing ones were left out in the average construction -without any weight, simple mathematical average. For example, letter C (for Capital) is a simple mathematical average of CP01, CP02, and CP03 indicators (later, the indicator CP03 will be left because of the lack of data).
After having the CAELS average, the ranking of countries can be created, since it is part of the original CAMELS methodology.
The free of charge database of IFSB (Islamic Financial Services Board) was used for the investigation. That is an available comprehensive systematic collection of the Islamic banking data. The focus of the study was only on the countries with the Islamic banking systems, not on the Islamic windows. There were fifteen countries involved in the analysis: Bahrain, Brunei, Egypt, Indonesia, Iran, Jordan, Kuwait, Lebanon, Malaysia, Nigeria, Oman, Pakistan, Sudan, Turkey, and the United Arab Emirates. Due to the lack of data, it was necessary to delete not only some main or sub-indicators, but also some countries.
The database had 19 available indicators. The time series started with the average for 2013, continued quarterly until the first quarter of 2018, which amounted to 18 time-series data.
The original three-dimensional data cube contained 19 * 15 * 18 data (19 indicators, for 15 countries for 18 time periods). Two of them -CP11 (Capital to assets (balance sheet definition), Tier 1 capital, Total assets) and CP12 (Leverage (regulatory definition), Tier 1 capital, Exposure) -related to the leverage of the banking system and were not part of the CAMELS methodology, thus they were omitted first. Five additional indicators had to be omitted because of the lack of huge amount of data: • indicator CP3 "Common Equity Tier 1 (CET1) capital to RWA"; • indicator CP15 "Liquidity coverage ratio (LCR)"; • indicator CP16 "Net stable funding ratio (NSFR)"; • indicator CP18 "Large exposures to capital"; • indicator CP1. "Large exposures to capital".
LCR and NSFR have recently required indicators by the Basel III system, so it is obvious that there was no data for them.
In the raw data table of the CAELS 100 analysis (Tables A1-A6 in the Appendix), the names of indicators remained the same and were used in the IFSB database. Thus, one can relate it easily to the original IFSB data columns.
Fortunately, the withdrawal of these five indicators from the analysis did not make significant difficulty, as the technique of creating the "letter average" only from the available ones was implemented. In every group, at least one sub-indicator remained.
In After omitting the indicators and countries, there was even some particular lack of data, which are listed with the methodology for processing them below. The cleaned data are in Tables A1-A6 of the Appendix. The CAELS raw data of six countries are listed, with the original numbering from the IFSB database also with the proportionality of the variable ("+" or "-" directly or indirectly proportional).
The CAELS averages were added as a new column (Remark: as only one variable refers to sensitivity, it is titled with S average as well -instead of duplicating the column). Into the row of proportionality in the average columns, there are "N.A." written, as the proportionality was not applied for these variables. If for a certain average all of the raw variables taking part in the average are with the same proportionality, the relevant sign ("+" or "-") appears in brackets, just as information not used for anything. The cells of the missing date remained empty, and the averages were created without them. The details of the implemented techniques are below: • In the case of Bahrain, there was no data available for the Net foreign exchange open position to capital in the Sensitivity to Risks group. It was handled by creating the CAELS average from 5 criteria instead of 6.
• Also for Bahrain, the total column of CP14 "Liquid assets to short-term liabilities" was practically absent. For such cases, the CP13 Liquid assets ratio made the average of "L". For the period of 2017Q4, the situation changed. There were data for CP14, but CP13 was missing. The creation of "L" has always been consistent with data availability.
• In the case of Egypt, there is no data for CP5 "Net non-performing financing (net NPF) to capital", so the average of the two remaining asset quality indicators has to be created.
• In the case of Egypt, CP14 "Liquid assets to short-term liabilities" has no value, the average for "L" was created based on CP13.
• For Oman, six data were missing for column CP06. For these time periods, an average was generated without these values.

RESULTS
Time averages of CAELS 100 indices are presented in Table 3. Based on them, a ranking of countries can be compiled. It is in the last column.
Looking at Despite the stability of a banking system, which is an important issue, this extremely high value refers to a very low risk-taking of banks in Oman. Looking beyond this fact, investigation of the original time series data of the indicators (CAR and Tier 1 capital to RWA) and also constructing the graph for them are needed (Figure 1). It was found, that the original indicators used were extremely high at the beginning of the period (81%), and it reached a value of 15%, which is a common value in other countries.
The maximum values of other countries are 22% and 21%, the minimal are 11% and 7% for CAR and Tier 1 capital to RWA, respectively. The variance of these indicators for Oman is bigger than 22%, while the others are smaller than 2%. To sum up, the high average for Oman is due to the high value in the past, now they have reduced their CAR and Tier indicators to the general level for the region.  In contrast to the capital adequacy ratio, there is a significant data dispersion in the asset quality, profitability and liquidity indicators. In terms of asset quality, Oman also plays a leading role with the best value (89.25%). It is followed by Egypt and Kuwait, with a score of about 70%, while the remaining three countries are below 58% (Remark: these values are scores on the 100-level ranking system, not the original values of the indices).
With regard to profitability, Egypt is the first and Oman is the last. It is not surprising, since Oman operated in the most risk-avoiding way, more than enough secure, thus, the country has the weakest profitable banking system. The performance of other four countries is very close to each other, from 53% to 55%, they produced almost the same relative profitability.
As for liquidity, Egypt leads the field with 53.28%, Turkey follows with 35.05%, and Kuwait, Oman, and Bahrain are in the middle, 16% -21.5%. The worst liquidity situation is in the United Arab Emirates with a 7.1% relative value.
For the last value, which refers to Sensitivity, a two-fold situation occurred. While the United Arab Emirates has 34.6% and Oman has 53.2% -they are at the bottom of the ranking -the other four countries have rather high points in the range of 88%-94%. Figure 2 shows the CAELS based performance of the countries of the Islamic bank sector (There is no value for variable "S" for Bahrain, so the line is above the letter L. Figure 2 better shows how countries' scores match and how close -even the most sensitive CAELS 100 indices. This fact will form the basis of the grouping.

DISCUSSION
The final result of the CAELS 100 new method can be seen in Table 3. In the before the last column, one can find the standardized average for the period 2013-2018, on the basis of which the countries are ranked. Based on this result, four groups can be created, as some of the points are very close to each other. This means that the performance of the banking system is nearly at the same level: • Egypt entered the first "group", ranking first with an average relative score of 58.22%.
Source: Own calculation based on the IFSB data.  Egypt is the first in three variables: "A", "E", and "L". It ranks second in terms of "S" (Sensitivity to risk) just one relative point than Turkey. But Egypt is the last in the indicator of "C" (Capital adequacy), which indicates stability or risk-taking by banks. It can be stated that Egyptian banks are taking risks and are successfully coping with this, given this time period. Their success is evidenced by the high values of other indicators.
Perhaps a more detailed investigation into the reasons for the Islamic bank system's particular performance will be carried out, but apart from the page limit, the authors do not consider themselves empowered to analyze the detailed banking and economic policies of these countries.
In summary, it can be said that the hypothesis -CAMELS can be used to rank Islamic bank systems of countries -can be accepted, with the remark that methodology modification is needed, for example, deleting the "letter M" refers to management and creating a 100-level evaluation.

CONCLUSION
Evaluation and comparison of banking systems is an important issue not only for conventional but also for Islamic banks. In the banking analysis literature, the use of the microeconomic CAMELS methodology is very common to evaluate banks. In this paper, this way was not used, but CAMELS was implemented at the macro level for the aggregated indices of countries with Islamic banking systems. This idea with a hundred-level evaluation and the interpretation of the management indicator make this publication a novelty, uniqueness.
Hypothesis testing based on free access data, IFSB data, contains aggregated data of Islamic bank sectors of countries.
The conclusion from this study is that CAMELS -after some modification -can be applied to rank Islamic bank sectors. The modified technique can be called "CAELS 100" because the letter "M", an indicator referring to management, had to be deleted, since there was no data for it in the IFSB database. The name "100" refers to the level of grading. It is much more sensitive than five grades of the original CAMELS methodology. These were the novelties in the methodology.
As an additional conclusion of this study, a ranking of selected Islamic banking systems was compiled. The selection was based on data availability.
Egypt has the best Islamic banking system. The medium level: Turkey, Kuwait, and Oman, and the worst of all is in the United Arab Emirates and Bahrain. These groups were created because the indicesdespite this more sensitive methodology -were very close to each other. The ranking of Islamic banking sectors of these countries for the period 2013-2018 is also a novelty of this publication.