The Euromoneyformula for calculating its emerging market bank (Emba) credit rating uses eight variables. These are: net loans to total assets, non-performing loans to gross loans, return on average assets, overhead costs to net income, deposits to net loans, inflation, disclosure and operating environment. The data for these financial ratios was taken from BankScope, a product of the rating agency IBCA. Our thanks to IBCA for letting us use the data. The data was a mixture of 1995 and 1996 figures and a “1” in the left-hand column of the country tables (second from left in the top 50) shows that 1996 data was used. The financial and rating information used was the latest available on July 28 1997.
The model was tested for its correlation with Moody’s financial strength ratings on a sample of 224 banks from 40 emerging markets, all of which had a Moody’s bank financial strength rating and total assets in excess of $1 billion. By using a multiple regression model, it was possible to test the formula to discover how well it correlated with Moody’s financial strength ratings, to find which variables had the most and the least explanatory power and to insert other variables to see how this affected the outcome (see main story). It was discovered that return on average assets and country risk had the most explanatory power of all the variables.
The raw values of the different explanatory variables were replaced by their decile rankings in the total group of banks to give consistent numerical scores. These were compared with Moody’s financial strength ratings by turning the ratings into figures using the following scale: A = 1, B+ = 2, B = 3, C+ = 4, C = 5, D+ = 6, D = 7, E+ = 8 and E = 9. The model gave a 63% correlation with Moody’s financial strength ratings. In only two cases in the original sample was the variation between the model’s results and Moody’s ratings as high as three notches, eg, D to C+ or C to E+. In 26 cases the variation was two notches and for the remaining 196 the variation was either one notch or the result was the same. The model was then applied to a broader sample of banks, many without Moody’s financial strength ratings. The results were ordered numerically and then converted into a Euromoney rating using the Greek letters alpha, beta, gamma. Banks with scores in the range 4.500 to 4.699 (a low score denotes a high predicted rating) were ααα, followed by αα for the next 0.200, and so on using α, ßßß, ßß, ß, ΔΔΔ, ΔΔ, Δ.
Banks from countries that do not have data on inflation or political risk were eliminated as were banks where too few ratios are available, for example in Russia. If only some ratios are unavailable for a particular bank, that bank is assigned the average decile value for the entire universe for those ratios.
The original list of 642 banks was pared back by taking out branches and most wholly owned subsidiaries of foreign banks, but leaving in banks with foreign shareholders where it was considered that the bank was essentially a domestic institution. With the recent wave of Latin American acquisitions in mind, it was decided to leave in banks that, arguably, remain domestic institutions for the time being even though they are 100% foreign owned, such as Argentina’s Banco Roberts and Brazil’s Bamerindus, both 100% owned by HSBC.
Even with the surprisingly good results, the model should be used with caution* and can never be more than a first step in a fundamental credit analysis. Further company-specific analysis, also including more subjective factors, should definitely be taken into account. However, the model is useful in giving a first impression of the global relative financial strength of a bank based on important financial variables. The model was devised by Deutsche Morgan Grenfell’s credit research department in London.
* Euromoney‘s Emba ratings are statements of opinions not statements of fact and the company does not have any liability for any loss or damage arising from actions or decisions taken using the information.
