Financial System of Kyrgyzstan
Part II
3. Efficiency of banks in Kyrgyzstan
3.1. The Banking Sector
3.2. Bank Efficiency
3. Efficiency of Banks in Kyrgyzstan
This chapter is divided into two sections: (a) an analysis of the Kyrgyz bank sector as a whole, and (b) an analysis of the functional efficiency of the Kyrgyz banks as such.
Back to Top
Dimensions
The banking sector of Kyrgyzstan is very small - by any measures. As of the end of 2003, the total assets of the domestic banking institutions amounted to 9.0 billion soms, or $203 million, and the total equity to 1.9 billion soms ($43 million)(1). Some idea of how Kyrgyzstan compares in this respect with other former Soviet countries, most similar to it in terms of economic development, can be derived from the list of the 100 largest banks of the CIS (excluding Russia) published in the last year by the British magazine The Banker(2). The fourth place in this list is occupied by Kazkommertzbank, the largest Kazakh bank (also represented in the Kyrgyz market by its subsidiary Kazkommertzbank Kyrgyzstan ). As of the end of 2001, Kazkommertzbank's assets neared $1.3 billion, and its tier-one capital topped $127 million, while the total assets of all Kyrgyz banks were just $105 million and the total equity $32 million, which is about the size of the equity of one Ukrainian bank, Nadra ($34 million), ranked 15th by The Banker. The then largest Kyrgyz bank, AKB Kyrgyzstan, had less than $14 million in assets and $3 million in equity. Both Kazakhstan and Ukraine are, of course, much larger than Kyrgyzstan, but even in Moldova, which is almost equal to Kyrgyzstan in terms of economy size, only one, albeit the largest, bank, Agroprombank, with equity at $27 and total assets at $109 million, exceeded by size the whole banking sector of Kyrgyzstan in 2001.
Given the intermediary nature of the banking business, relative measures generally better reflect its dimensions than do absolute ones, especially in international comparisons. This approach, however, makes the underdeveloped state of the Kyrgyz banking sector all the more obvious. In relation to gross domestic product (GDP), the banking sector's average total assets made up only 10% in 2003. This is the level of modern, war-devastated and internally divided Iraq , whose banking system is usually described as rudimentary. In developed countries, for comparison, banking asserts frequently exceed GDP.
Getting back to the banking system of Kyrgyzstan , it should be noted that its development has been seriously retarded by the Russian financial crisis of 1998. During the next three years after it, the ratio of banking assets to GDP has decreased from about 10% to slightly more than 6%, that is, to its 1995 level. Only during the last two years, the banking assets have jumped by 93% to reach again the 10% level of GDP. In other words, quantitative characteristics of the domestic banking sector assets have been quickly improving of late; now we will try to find out what is happening with their quality.
Asset Quality
Banking assets can be divided into two large groups: earning (investment-account) assets and nonearning assets. The basic function of a banking system is to mobilize free financial resources and place them in earning assets, first of all, loans. Other, nonearning assets play an auxiliary role, helping banks to smoothly perform their main tasks, meet the reserve requirements, and maintain correspondent networks necessary to execute international payments. For banks, payments services are a source of noninterest income, which becomes especially attractive in conditions of increased macroeconomic volatility and decreased business activity, as the great majority of payments transactions are free from credit risks associated with lending and largely influenced by the condition of the economy.
If, however, in a desire to reduce the overall level of risk on its operations, a bank chooses to cut back its lending volumes, its profitability is bound to lower, as is its ability to pay competitive interest on deposits and dividend on its shares. This means that raising external financing in the open market becomes problematic. A bank can only afford to let such a development take place if, first, it is supplied with (relatively cheap) financing from other sources, and, second, profit maximization is not in fact its main objective. In such a mode function state-owned banks; to a certain extent, banks supported by international donor institutions; and banks connected with financial or industrial groups. In all these cases, a bank is, in effect, just a link in a broader system of interests, within which it is assigned some special tasks, while its own profitability is not considered a top priority. The latter circumstance gives such a bank a big, if not decisive, advantage over banks financed on a purely market basis, even if they are more effective functionally. At that, the largest distortions to the nature of competition in the banking sector are brought about when banks provided with financing of nonmarket or not quite market origin operate in the main banking markets.
This is an important factor in understanding the situation in the Kyrgyz banking sector, and the efficiency evaluation methodology that was applied to Kyrgyz banks under this study and that is described in the second section of the present chapter.
Banking is one the industries most sensitive to macroeconomic fluctuations. Interestingly, whereas in 1995-1999 the dynamics of the changes in the ratio of banking earning assets to total assets approximately followed those of the country's economic growth but with a lag of roughly one year, over the next two years these two trends have substantially converged, both in time and in the scale of yearly fluctuations. Together with the fact that the level of statistical association between the key volume indicators of the banking sector and nominal GDP has substantially risen over the last years, this can be seen as a sign of increased susceptibility of the domestic banking to the condition of the economy.
The highest level, 62%, of the earning-to-total asset ratio in the banking sector of Kyrgyzstan was recorded in 1998. By the end of 2002, it declined to 42%, and over 2003 rose to 49%. While in the last two years total assets have grown, as was already mentioned, by 93%, earning assets have increased only by 77%, and nonearning assets by 111%. In other words, the overall quality of the banking assets has substantially deteriorated. Notably, the higher growth rate of nonearning assets was almost totally due to just one their type, balances in correspondent accounts with other (that is, foreign) banks, which have increased by the whole 246%. Even more striking is that of the total amount of correspondent-account assets of local banks, 55% in 2002 and about 45% in 2003 was accounted for by only one AsiaUniversalBank. At that, this bank's share of the total earning assets of the domestic banking sector was only 20% and 13%, respectively. In total assets of domestic banks excluding AsiaUniversalBank, correspondent accounts with other banks make up on average(3) 13%. In AsiaUniversalBank this figure stands at 58% (67% in 2002). For comparison, in large Western banks this ratio rarely exceeds 3%, while in such Kazakh banks as Kazkommertz, TuranAlem and Halyk it is about 8% to 9%. Taking into account the differences in sizes, overall efficiency and economic environment, the 13% for Kyrgyz banks looks approximately normal. While the situation with AsiaUniversalBank can only be explained by this bank's preoccupation with payments business, an activity normally carried out by specialized clearing institutions.
Over the last five years, the following main changes have taken place in the composition of the banking sector's earning assets: the share of net loans has decreased from 75% to 61%, while the share of securities has risen from 16% to 26%. In other words, the level of risk inherent in earning assets has substantially lowered, as did their overall income-generating capacity. In addition, the ratio of loan loss allowance to gross loans increased from 4.6% to 7.2% (8.6% in 2002), testifying to deterioration of loan quality.
Thus, over the last five years the domestic banks' lending activity, relative to their overall operations, have substantially lowered: if in 1998 the average share of loans in the banks' total assets was 48%, by the end of 2003 it decreased to 36%. In Kazakhstan , for comparison, loans make up about 60% of total banking assets. The disposition to engage in lending varies in a very wide range among Kyrgyz banks. Most active in this respect is Ineximbank, with the average share of net loans in total assets for the last three years at 60.5%. Immediately following are AKB Kyrgyzstan (56.1%) and Ecobank (53.0%). In 2003 alone, the most active lender turned out to be Kazkommertzbank Kyrgyzstan , with 64.0% of its total assets in net loans. (In the parent Kazkommertzbank, this ratio stands at over 70%.) Kazkommertzbank Kyrgyzstan is also ahead of all other actively lending domestic banks in terms of the average annual growth rate in net loans for the last three years-73%, compared with 20% on average for all the banks. At the other extreme, after the practically non-lending Settlements and Savings Company (SSC) and Kairat, is Demir Kyrgyz International Bank that had only 1.6% of its total assets in net loans in 2003 and 0.8% on average for the last three years. By the way, in 1998 net loans exceeded 27% of Demir's total assets.
Securities are the main source of interest income chiefly for government-owned SSC and Kairat bank(4); in both of these institutions, securities (also of the government type) have accounted for about half of total assets and more than 90% of earning assets on the average for the last three years. Among private banks, the securities' share of total assets is the largest in AsiaUniversalBank (18%), Energobank (17%) and AKB Kyrgyzstan (12%). But unlike the other two banks, AsiaUniversalBank has its securities portfolio entirely composed of corporate paper. In this, by the way, AsiaUniversalBank is different from all other banks in Kyrgyzstan .
Financial Structure
Over the last five years, the average financial structure of banks in Kyrgyzstan has not changed much. The share of equity in the total banks' funding has remained practically at the same high level of about 28%. The share of deposits has increased from 52% to 58%. In the structure of deposits, the share of corporate demand deposits has increased from 47% to 54%, while the share of corporate term deposits has decreased from 8% to 6%, and the share of savings deposits from 45% to 42%(5). In the overall financial structure, that is, relative to total liabilities and equity, the share of savings deposits has decreased from 24% to 22%, and the share of corporate term deposit from 5% to 4%. In other words, today pay sources play a somewhat smaller role in the Kyrgyz banks' funding than five years ago.
A striking departure from this general picture is again, like in the case of the composition of assets, marked by AsiaUniversalBank in whose total deposits 53% (compared with 4% on average for other domestic banks), or 33% of total assets, is accounted for by demand deposits of banks and other financial institutions.
Profitability
Interest Income and Expenses
In the last five years, the average gross yield on earning assets (the ratio of interest income to average earning assets) of domestic banks has deceased from 33% to 19%, while inflation has lowered from 35.9% to 3.1%, that is, inflation has decreased to a much larger degree than has yield on earning assets. Taking into account that the share in total earning assets of less profitable items (securities) has over the same period increased, it follows that interest rates on loans must have been lagging behind inflation even by a larger margin than the overall yield on earning assets. In other words, the sharp decline in the inflation rate has allowed the domestic banks to impose their own anti-inflationary, say, tax on their borrowers. As a result, real gross yields on the banks' earning assets rose quite substantially, but not as much as did real interest rates on loans.
During the first two years of the five-year period considered, the average real cost of funds (as measured by the ratio of interest expense to average interest-bearing funds) remained negative. Over 2001-2002, it rose to 6.3%, and by the end of 2003 lowered to 4.4%. Since term deposits make up about 80%, on average, of the banks' total interest-bearing funds, the ratio of interest expense to interest-bearing liabilities can be used as a proxy for yield on bank deposits. As will be shown later, the differences between the domestic banks on this measure are very large; and the decrease in its general level (in real terms) that occurred in the last year indicates that the portion of the banks paying less on deposits than the average level is larger today than it was a year ago. Statistically speaking, the asymmetry of distribution of fund-cost data has shifted towards small values.
As a result of the above trends, the interest spread, that is, the difference between the interest rates earned and paid, roughly corresponding to gross profit margin of nonfinancial concerns, has become almost the same in nominal and real terms - about 11.5%.
The highest yields on earning assets are received by the following three banks: Ecobank (35% in nominal terms), DosKredo (33%) and KyrgyzKredit (30%). The least yielding are earning assets in the three non-private banks: the Bishkek branch of the National Bank of Pakistan (NBP, 3%), Kyrgyz Investment and Credit Bank (KICB, 7%) and Kairat (10%). Among private banks, yield on earning assets is the lowest in AsiaUniversalBank (11%).
The highest levels of interest expense relative to the amount of funds raised are displayed by Kairat (15%), DosKredo (15%) and KyrgyzKredit (11%); the least expensive funding is employed by NBP (0.4%), Demir (0.9%) and AsiaUniversalBank (1.9%).
In terms of interest spread, the top three banks are Ecobank (25%), KyrgyzKredit (19%) and Bakai (19%); the bottom ones are Kairat (-5.2%), KICB (1.5%) and NBP (2.7%), and, from among private banks, Kazkommertz Kyrgyzstan (8.5%).
Noninterest Income and Expenses
Over the period 1999-2003, the share in the banks' total revenues of their second component, noninterest income (income from services), has increased on average from 28% to 49%. At that, the range of variation of this ratio among banks has considerably widened. If in 1999, its largest value was 79%, and the smallest 11%, in 2003 the corresponding figures reached 89% (Demir) and 9% (Tolubai), respectively. Since these two banks, Demir and Tolubai, have been at the opposite ends of this ratio range all the last five years, they can be viewed as extreme examples of narrow specialization in terms of main revenue sources. From the banking point of view, Tolubai's business model is stronger because most of its revenues this bank generates in a way a commercial bank, strictly speaking, should-from lending, the most profitable banking activity. Although the level of noninterest income as low as it is in Tolubai is of course a drawback. There is another aspect to this problem: when a bank depends too heavily on noninterest income, its risk management capacity, the key managerial characteristic in banking, tends to weaken. In this respect, one of the most productive combinations of ratios is demonstrated by Ecobank, in which noninterest income accounts for 52% of total income, while earning assets yield 29% in interest income after loan loss provisions. For comparison, in Demir the second ratio stands at less than 13%. (In Tolubai, it is 28%).
The overall profitability of banks in which noninterest income accounts for more than half of total income tends to be lower than the average level. An exception from this rule, bar the already mentioned Ecobank, is Bakai, with 69% of total income made up by noninterest income, and profitability far above the sector average and higher than in both Ecobank and Tolubai. However, Bakai has the actual cost of funds at a level that is more than two times lower than it is in Ecobank and Tolubai.
Bakai and Ecobank also excel all other domestic banks in the efficiency of noninterest income production: on each 100 soms of average assets they generate 22 and 18 soms of noninterest income, respectively, while the sector average is only 8 soms.
In most cases, a bank's noninterest expense exceeds its noninterest income. Since the main portion of noninterest expense is made up by items classified as fixed expenses (such as salaries), the ratio of net noninterest expense (less noninterest income) to earning assets is used in banking as a measure of break-even point (yield). For manufacturing enterprises, for example, the break-even point is usually defined as the smallest volume of production required to cover fixed expenses. When output exceeds the break-even point, net profit is earned; otherwise, net losses are incurred. In banking, where percentages are more common, the break-even point is calculated as the minimal level of net interest yield on earning assets at which a bank finds itself in the black.
The average break-even yield in the Kyrgyz banking sector currently stands at about 7% (it was 9% a year ago). There are only three banks in Kyrgyzstan that have this ratio below zero, with noninterest income exceeding noninterest expense: Demir, Bakai and AsiaUniversalBank. In Demir's case, the absolute value of the break-even yield (29%) even exceeds the net interest yield on earning assets in the best-performing, on this count, banks, such as Ecobank (20%).
Overall Profitability
In the last five years, the average return on assets (the ratio of net income to average assets) in the domestic banking sector has even slightly decreased-from 2.2% to 2.1%, while relative to average equity net income has decreased from 10.1% to 9.1%, on average. With extremely large, by international standards, interest spreads, such values of the sector-average profitability ratios look very modest. They, however, disguise wide differences between the most and the least profitable banks in Kyrgyzstan .
The most profitable bank in Kyrgyzstan for the last two years in a row remains Bakai whose average return on assets for these two years stands at 8.5%, and return on equity at 52.4%. The least profitable, or rather most loss-making, bank, also for the last consecutive years, is DosKredoBank (-11.2% and -30.5%, respectively). The range of variation of the banks' profitability is, however, considerably narrower today than it was in the past. Apart from the jump caused by the Russian financial crisis of 1998, the general tendency observed for the last nine years is for the banks to steadily draw closer to each other in terms of profitability. This may be a sign that in spite of the remaining structural differences between individual banks, the banking sector of Kyrgyzstan is more homogeneous today functionally than it was before.
Over the last five years, the share of banks with above-average profitability has decreased on both ratios, but on return on equity it has declined much more than on return on assets. If in 1999, the share of banks earning more than the sector average was 77% relative to assets and 75% relative to equity, by the end of 2003 these figures lowered to 45% and 40%, respectively. In other words, it is getting more difficult for the domestic banks to earn competitive returns on assets than on equity. From the technical point of view, this is quite understandable, since the banks' assets have increased by a much smaller degree than has their equity. In economic terms, this tendency threatens to further undermine the already feeble economic profitability of the banks (see Cost of equity capital and economic profit below).
Concentration of Banking Markets
Over the last five years, the concentration of the customer loan market (excluding loans to financial organizations) has not practically changed: by the joint share of the three largest players (the CR3 ratio) it was 46% in 1999 and 47% in 2003. The largest lender in Kyrgyzstan in the last four years has invariably remained AKB Kyrgyzstan; this bank accounted for 25% of the loan market in 2000 and 20% in 2003. The second and third largest lenders are Ineximbank and Kyrgyzpromstroybank with, respectively, 14% and 13% of the loan market.
The concentration of the market of government securities has decreased, by CR3, from 69% in 1999 to 59% in 2003, with 29% of the market held by SSC (39% in 2002), 17% by Kairat and 12% by Energobank.
The leading position in the market of savings deposits, the concentration of which over the last five years has decreased from 49% to 34%, is also held by AKB Kyrgyzstan. This bank accounted for the largest shares in the saving deposit market during all the three last years, with 14% of the market in 2001 and 12% in 2003. Just slightly lagging behind AKB Kyrgyzstan is Energobank whose market share is close to 12%. At that, Energobank is more than 40% smaller than AKB Kyrgyzstan in terms of total assets. The third largest player in the market of savings deposits is SSC with 10%.
Even more assuredly than AKB Kyrgyzstan on savings deposits, Demir is leading on corporate demand deposits, with its share in their total amount in the banking sector now standing at 18% (29% in 2001). The second place is occupied by Kyrgyzpromstroybank (about 17%), and the third by, again, Energobank (10%). Thus, the joint share of these three banks in the total corporate demand deposits is 44% (the corresponding figure for 2001 stood at 60%).
Demir is also holding the lead in the market of banking service; its share in the total noninterest income of the domestic banks is 13% (24% in 1999). Other largest domestic banking institutions in this respect are Kyrgyzpromstroybank, SSC, AsiaUniversalBank and AKB Kyrgyzstan. The concentration of the market of banking services has decreased over the last five years from 53% to 34% (the largest value of this ratio over the period in question, 64%, was recorded in 2000).
The concentration of net assets (equity) has decreased from 46% in 1999 to 37% in 2003. The largest banks in Kyrgyzstan by net assets are KICB (15% of the banking sector total), AsiaUniversalBank (14%) and AKB Kyrgyzstan (8%).
In sum, the highest levels of concentration are observed at present in the markets of government securities and corporate demand deposits, the lowest in the market of savings deposits. In the current year, 2004, with the bank Kairat passing into private hands, concentration of the market of government securities is bound to decrease, following an inevitable reduction in the share of such items in the composition of this bank's earning assets. As of the end of 2003, with 99% of its earning assets in relatively low-yielding government securities, Kairat remained one of the least profitable banking institutions in Kyrgyzstan .
As concerns corporate demand deposits, the structure of this market can be substantially changed by, again, Kairat, if in its regained capacity as private bank it manages to reestablish itself in the market. The past record is encouraging: In 1998, Kairat (then Maksat) accounted for 18% of the total corporate demand deposits in the banking sector of Kyrgyzstan , or the same market share as is held by the current largest player, Demir.
On the whole, the banking sector of Kyrgyzstan is entering an interesting stage in its development. The unlucky attempt of the central National Bank of Kyrgyzstan to initiate consolidation of the sector by raising the minimum capital requirement has shown that the banking sector of Kyrgyzstan is not ripe yet for internal mergers. And now Kyrgyz banks are being steadily absorbed by their Kazakh counterparts. From this, by the way, Kazakh banks can gain much less than Kyrgyz ones-given the enormous differences in the banks' sizes, modest capacity of the Kyrgyz market, and less than brilliant condition of the banks being acquired. But Kazakh bankers seem to be inspired by much stronger optimism in respect of Kyrgyzstan , which could not be anything but pleasing if not for a feeling that this optimism is tinted by the herd instinct so well described in banking literature.
Incomparably better capitalized and aggressive Kazakh banks are well in a position to profoundly change the structure of the banking sector of Kyrgyzstan. Another question is how soon this may happen. It is hard to imagine that for such a bank as Kazkommertz, for instance, investments in the Kyrgyz banking sector could have strategic importance justifying large-scale, accelerated measures to increase its share of the Kyrgyz market. On the other hand, Kazakh banks are not very likely to remain content with their present position in the Kyrgyz market for long. We believe that the initial stage of consolidation of the Kyrgyz banking sector will start in the nearest two to three years. Which of the local banks will be first involved in this process will largely depend on the strategy to be chosen this time by the Kazakh banks. As numerous Western studies show, about 50% of mergers fail to achieve their main goal, that of increasing shareholder value. An explanation lies in the fact that in too many cases mergers and acquisitions are targeted at slowly growing but less expensive companies. And growth in this respect should be judged by the rate of increase not in accounting but economic profits, that is, profits less the charge for equity capital.
Cost of Equity Capital and Economic Profit
First of all, it should be made clear what is meant by cost of equity and why it is important to get a handle on its numerical value. The true economic soundness of a company is determined by its ability to generate profits at levels exceeding not only production and operating costs, but also the cost of capital. Capital can take two basic forms - borrowed and own, or shareholders' equity. The cost of borrowed capital is fully determined by the interest rate which for the creditor serves as the expected return on its loaned funds(6). The expected return on equity investments is just another name of cost of equity capital, or by general convention, cost of capital. A difference between a creditor and shareholder is that a creditor's claim to his expected returned can be defended in a court of law, whereas a shareholder can only hope that his actual return (dividend) will not turn out to be a disappointment. Before becoming a shareholder, a man, of course, tries to weigh the pros and cons; nevertheless, the intrinsic uncertainness of shareholders' returns remains, giving shareholders the right to expect higher levels of returns than on debt financing, that is, the right to a risk premium.
For quantitative evaluation of the cost of (equity) capital, a most commonly used method requires, first, to identify a risk-free interest rate to which at the second stage of the computation a market-risk premium is added. Yields on government securities, of one maturity or another, are usually taken as the risk-free rate. In Kyrgyzstan , as in other less developed countries, such an approach is not directly applicable because government securities, or debt obligations, are not quite free from the risk of default. This circumstance is reflected in the notion of sovereign risk which also includes country risk relating to general risks of conducting economic activity in given country. Sometimes these two notions, sovereign risk and country risk, change places, and country risk is construed as a more general term including the risk of default on government obligations-as in the rating of 185 countries of the world by the level of country risk published in the March 2004 issue of the Euromoney magazine, an abstract from which is presented in Table 3-1.
Over the period of September 2003 to March 2004, the assessment by the Euromoney magazine of the country risk in Kyrgyzstan has substantially lowered, unlike in other former Soviet republics, including Turkmenistan and Moldova . In Euromoney's view, doing business in Kyrgyzstan is as dangerous and unpredictable an occupation as it is in such African countries as Togo, formerly part of the Slave Coast, and Mali where only one third of children attend primary schools. Another striking feature is the extremely low level of access to bank finance in Kyrgyzstan , judged by the ratio of domestic credit to the private sector to GDP (we wrote about it in the previous issue of our journal). Euromoney also publishes country ratings in terms of economic projections for the nearest two years; by this measure Kyrgyzstan has been downgraded from the 136th rank to the 137th, while Kazakhstan, for instance, has risen from the 83rd rank to the 58th, and Moldova from the 140th to the 123rd. By the way, the International Monetary Fund (IMF) has lowered its economic growth forecast for Kyrgyzstan for 2004 from 5% to 4%.
Table 3-1
Country Risk
(Click toView Table)
Note: in its full form, the table includes other criteria, such as, for example, Debt indicators; the sum of the individual scores shown here does not therefore add up to the total score. Source: Euromoney, March 2004.
So, the country risk in Kyrgyzstan is very high, at least by Western estimates, and it obviously should be somehow included in the cost of capital. In practice, precise quantification of the country risk in conditions of undeveloped capital markets and absence of internationally comparable financial instruments, as is the case of Kyrgyzstan , becomes an extremely difficult task to fulfil. And such estimates as offered by the Euromoney magazine are impossible to use for the calculation of the country risk premium-they are designed for other purposes.
As concerns market-risk premiums, they are calculated on the basis of historical yields on equities and risk-free instruments, which approach is only applicable under conditions of developed securities markets and more advanced economy.
The simplest and most widely applicable method uses the 10-year U.S. government yield to which an inflation differential between the U.S. and a given country is added. Thus calculated, the local risk-free rate is then increased by 4.5% to 5.0% representing the assumed market-risk premium which in emerging markets is also extremely difficult (if at all possible) to calculate precisely. This approach is based on the assumption that investors have access to an international risk-free rate, which for the great majority of local investors in Kyrgyzstan is of course not true.
We therefore will employ a modification of this method in which the derived risk-free rate is increased by the standard deviation (a measure of variation) of the annual changes in real GDP for the period from 1992 to a given year. Our reasoning is based on the fact that in economic sense risk is the volatility, uncertainty of future outcomes, that is, the magnitude of their potential deviation from the mean, which statistically measured by standard deviation. The results of this exercise, the estimates of the cost of equity capital in Kyrgyzstan for the last five years, are shown in Table 3-2.
Table 3-2
Cost of Capital
(Click to View Table)
Now we can proceed with our evaluation of the value-creating capacity of the local banks by calculating their economic profits. Economic profit is defined as net profit after the charge for capital calculated as the product of the sum of shareholders' equity at the beginning of the year and the nominal cost of capital (expressed as a percentage). A company is said to be creating value if its economic profit is positive; otherwise value is destroyed.
Most of the banks in Kyrgyzstan do not create value, that is, they do not earn enough net profit to cover the cost of capital. Moreover, they remain in this condition chronically. There are only seven banks that had their economic profits in the black at least in the last, 2003, year. These banks are shown in Table 3-3.
Table 3-3
Banks Creating Value
(Click to View Table)
In terms of the value-creating capacity, the absolute leader is the bank Bakai whose economic profit was slightly negative only once over the last five years-in 2000. Among the rest of the local banks, closest to the leading group is the Bank of Asia with the economic loss of less than 0.9 million soms (about $20 thousand) in 2003-a sharp decrease compared with nearly 12 million soms ($267 thousand) in 1999.
Back to Top
Method of Evaluation
Under this study, the efficiency of the banks in Kyrgyzstan was evaluated with the help of a method known as data envelopment analysis (DEA).
DEA is a technique for quantitative evaluation of the efficiency of entities (organizations, programs, etc) that utilize resources to obtain desired outputs. DEA allows to use multiple inputs and outputs to arrive at a scalar measure of total efficiency/productivity. This gives DEA a significant advantage over traditional, mostly partial measures of productivity. Take, for instance, conventional bank ratings-the mechanics of their development are mostly based on selection of a set of financial ratios each of which is then assigned a certain score, or weight, depending on the value of the ratio, and the final rating is computed as the sum total of individual scores. One of the greatest weaknesses of such approach is that the score intervals according to which the relative significance of a given ratio value is determined are preestablished from outside the model by the analyst. This makes the whole system given to subjectivism. DEA is free from this main deficiency of conventional rating systems because the weights assigned to variables are determined inside the mathematical model, independently of the analyst.
In mathematical terms DEA can be described as a fractional linear programming model that can include multiple outputs and inputs without recourse to a priori weights (as in index number approaches) and without requiring explicit specification of functional relations between inputs and outputs (as in regression approaches). It should be remembered that DEA evaluates relative efficiency of an entity as compared with other similar entities.
DEA efficiency scores are computed as quotients of the sums of weighted outputs and inputs of each of the entities included in the data set being analyzed. Thus, the more and the less efficient entities are identified. DEA is carried out by maximizing the following mathematical model:
Ep = E uj Yjp / E vi Xip
subject to efficiency scores Ep 1 for all entities (p), and weights vi and uj assigned to outputs (Yjp) and inputs (Xip) being more than zero. The model is run repetitively for each one of the entities in the set, with the constraints held constant. In each case, the best combination of weights is computed-a combination that results in a maximal score under the given model parameters. Any other combination of weights will produce a lower score. This is another advantage of DEA, as it leaves no ground for an argument that an efficiency score could be higher if a different set of weights was chosen.
Most of traditional efficiency/productivity measures-such as, for instance, manufacturing output or, in banking, number of client accounts processed per unit of labor-are partial by nature. More comprehensive evaluations of efficiency/productivity are produced on the basis of the total factor productivity (TFP) approach. DEA is a most widely used TFP technique today. The range of types of organizations and activities to which DEA is applicable-and is actually applied-is very wide, from prisons, educational institutions and clinics to warehouses and oil companies. A separate field of DEA-based research is banking.
In banking, the notion of efficiency is not as clear-cut as in other areas of economic activity. Identifying proper variables to be used as inputs and outputs in the efficiency-evaluation model is therefore a task in itself. Deposits, for example, can be classified as either inputs or outputs. The approaches to solving this problem can be divided into two large categoriesproduction-oriented and intermediation-oriented. Under the production approach, emphasis is made of the ability of banks to produce and efficiently process various types of deposit and loan accounts. Outputs are measured by the number of such accounts, or the number of transactions per account, per unit of time, and inputs by total operating costs incurred in producing these outputs.
Under the second approach, banks are primarily viewed as financial intermediaries, and the technical efficiency, with which banks execute their day-to-day tasks, plays an auxiliary role. Outputs are measured by the value of loans and investments, and inputs by the cost of labor and capital. Deposits can be treated as either inputs or outputs.
The intermediation approach-employed under this study-is most widely used because it presents fewer data problems than the production approach. In conditions of underdeveloped financial markets, the intermediation approach has another advantage over the production approach-it allows to take account of non-market factors that can substantially distort the nature of competition.
Model Parameters
The following variables have been used as the parameters of the DEA model under this study:
- Outputs:
- average earning assets,
- average savings deposits,
- total income as the sum of interest income after provisions for loan losses and noninterest income.
- Inputs:
- average nonearning assets,
- overall costs as the sum total of noninterest expense, the charge for equity capital, and adjusted interest expense (the charge for interest-bearing funds).
Justification of the Choice of the Model Parameters
First of all, we would like to emphasize once again that the object of this study is the efficiency with which the banks in Kyrgyzstan perform their basic intermediary role, that of taking deposits and making loans. The combination of these two functions, plus liquidity that banks provide to their customers, is what makes them different from other types of financial organizations. These basic premises underlie the choice of parameters for the efficiency measuring model applied to the banks in Kyrgyzstan under this study. More specific explanations are given below.
Output Parameters
Lending is an activity that brings banks most profits but that, at the same time, exposes them to most damaging banking risks. In order to limit these risks, banks place part of their free funds in other earning assets, first of all, government securities. This allows banks to maintain their liquidity at safe levels so that they can meet their current obligations without much difficulty, which becomes especially important in times of deteriorating economic situation and their decreasing credibility with the public. With loans being the least liquid assets, banks need to have such assets as well that could be more easily converted into cash, if such a necessity arises. Liquidity of assets, however, is inversely related with their earning capacity, correspondingly affecting banks' profitability. But of two banks with similar structure of earning assets, a higher level of (interest) income will be earned by the one that better manages its loans. Based on these considerations, average earning assets and interest income after loan loss provisions have been included in the model output parameters.
Of all types of deposits, savings deposits meet the definition of bank product most closely. In addition, more than any other banking characteristic, savings deposits reflect the amount of social capital of a bank, a criterion most important in modern Kyrgyzstan , given the very low credibility of its banking system.
Input Parameters
Total assets reflect the value of all physical and financial resources used by a bank. But as one part of them, earning assets, has already been included in the output parameters, to avoid double counting here we include only their nonearning remainder.
Noninterest expense reflects the overall cost of running a banking business. The charge for equity capital, the second component of the cost input variable under the model, has been calculated as the product of the shareholders' equity at the beginning of the year and the percentage cost of capital (see Cost of equity capital and economic profit above). The last component of total costs, the adjusted interest expense (the charge for interest-bearing funds) requires a special explanation.
Interest Expense
The DEA methodology is based on the principle of comparability of the objects analyzed. In other words, this technique is only applicable to such entities that are engaged in basically the same activities and, importantly, function in equal conditions. It follows that the criteria by which they are evaluated should also be the same-by essence, not just by name. But when the market self-regulation mechanisms are not developed enough, in certain industriesincluding the banking among the first-structural distortions may exist that change the nature of competition and give certain firms advantages over the others.
In the traditional Porter industry model, all firms compete on purely economic, market-based terms, which in developing countries rarely holds in full. In the Coyne/Subramaniam model, this factor is taken account of by inclusion of several new dimensions, such as privileged relationships and co-dependent systems. As their name implies, privileged relationships exist between companies that treat each other in a special way based on family ties, friendship, common political interests, etc. Co-dependent systems refer to such industries where alliances and networks are important factors of competition. Such business webs may of course be vertically integrated, affecting whole sections of a supply chain, or even entwining different supply chains. Without going too deep into this issue, we would like to note here that in banking the existence of co-dependent systems may be manifested, in particular, by low cost of borrowed funds or deposits.
In normal conditions, with market forces well developed, the costs of similar types of borrowed funds and deposits in banks operating in the same markets should be approximately equal. Because commercial banks are supposed to obtain their funding on a market basis from enterprises, organizations and population. Substantial differences in the costs of similar funds can be economically justified only when there are large enough differences in maturity characteristics. Obviously, this is not the case of the banks in Kyrgyzstan . Nevertheless, by the actual average cost of funds, as measured by the ratio of interest expense to interest-bearing funds, Kyrgyz banks vary in an extremely wide range.
To a certain extent, these variations are explained by the fact that some banks in Kyrgyzstan have access to relatively cheap funding provided by international donor organizations, or, in the case of branches of foreign banks, by their head offices. In addition, a bank's overall cost of funds is influenced (sometimes, considerably) by other items, such as, for instance, "deposits of and loans from the government agencies". However, the use of such sources of funding should be viewed rather as a deviation or, say, a result of fortunate circumstances, than a normal practice. In any case, differences in the cost of funding, when not objectively warranted, can substantially distort the true functional efficiency of banks, and should therefore be adjusted to prevent their influencing the evaluation results. Let us add, in parenthesis, that data dispersion is much smaller in the case of the ratio of interest income to earning assets than on the ratio of interest expense to interest-bearing funds. This is understandable, because in the present-day Kyrgyzstan , opportunities for attaining profitability at levels exceeding industry averages are much narrower than for obtaining less expensive funding.
The adjustment of the charge for interest-bearing funds has been performed in following manner: the average value of interest-bearing funds of each bank has been multiplied by the banking-sector average cost of funds, as measured by the sector average ratio of interest expense to interest-bearing funds.
In this connection, an objection may be raised that such an approach may make look undeservingly better such banks that have the level of interest expense higher than the sector average due to poor management or other internal problems. To this, we can say the following: (a) an excessively high level of interest expense, as a reflection of poor management, cannot exist alone, without affecting other key aspects of a bank's performance; and all those banks (in the sample analyzed) that during the last three years have stably remained much above the sector averages for the cost of funds (DosKredoBank, in particular) are ranked the lowest by our model.
Evaluation Results
The efficiency of the banks in Kyrgyzstan has been evaluated for the last two years. Accordingly, only such banks have been covered that have remained in full-scale operation during both these years. Also excluded, have been banking institutions that do not engage in commercial lending, and the Bishkek branch of the National Bank of Pakistan (NBP)-because of the unclear situation with its equity. In this away a sample of 15 banks has been selected, whose resultant efficiency scores are shown in Table 3-4 below.
As can bee seen, the efficiency of the banks in Kyrgyzstan , as a whole, has somewhat increased over the last year. The largest gains in efficiency have been made by such banks as AKB Kyrgyzstan and the Bank of Asia-about 20% each. A considerable increase in efficiency (12%) has also been achieved by Energobank, the fifth most efficient bank in Kyrgyzstan . Except for Energobank and the absolute leader, Tolubai, the other three banks among the top five demonstrate decreasing efficiency, with the greatest loss (8%) recorded for Ecobank.
Noticeably, all the three most efficient banks in Kyrgyzstan , namely: Tolubai, Bakai and Ecobank, belong to the category of small banks-their joint shares are only 8% in the total assets of the sector and about 11% in its net loans. At that, these banks account for 20% of the total sector's savings deposits and 28% of its net profit.
Table 3-4
Kyrgyz Banks' Efficiency
(Click to View Table)
Analysis of the Results
Here we will try to show why the rankings in our efficiency scoring system have been distributed in this way and not another. We begin from asset quality.
Asset Quality
The ratio of earning assets to total assets in Tolubai, the most efficient bank in Kyrgyzstan by our evaluation, is much above the average level for the whole sample of 15 banks analyzed. In the other two most efficient banks, Bakai and Ecobank, this ratio is, on the contrary, lower than the sample average. In addition, over the last three years Tolubai has been considerably ahead of the rest of the banks in terms of stability of the share of earning assets in total assets which has invariably remained in it within a narrow range of 68% to 69%. Whereas in Bakai this ratio stood at only 47% in 2003, having increased though from 37% in 2001, and in Ecobank its has decreased over the same period from 74% to 42%. On the average for the whole sample, earning assets accounted for 54% of total assets in 2001 and 51% in 2003.
On the average for 2001-03, the share of government securities in total earning assets, as a measure of their liquidity, is 31% in Tolubai and 12% for the whole sample. Over 2003, however, this ratio has decreased in Tolubai from 38% to 15%, which is, nonetheless, still higher than the sample average for the same year (13%). In Bakai and Ecobank this ratio is substantially below the average-about 9% for each in 2003, and 8% and 7%, respectively, on the average for the last three years.
On the whole, quality of assets is one area where Tolubai excels most impressively. On the average for the last three years, the ratio of loan loss allowance to gross loans is 3.0% in Tolubai, while the sample average stands at 8.8%. By this measure, Tolubai is falling behind only the Bank of Asia (2.6%), but while the loan loss allowance relative to gross has even slightly increased in the latter from 2.8% in 2001 to 2.9% in 2003, with Tolubai this ratio has shrunk over the same period from 4.1% to 2.1%. Moreover, by the ratio of current-year loan loss provisions to gross interest income Tolubai is far ahead of all banks in Kyrgyzstan , including the Bank of Asia: in 2003, this ratio was 2.1% in Tolubai, 6.9% in the Bank of Asia, and 14.0% on the average for the sample. Bakai and Ecobank loan loss provisions made up, respectively, 15.2% and 17.3% of gross interest income.
In spite of the lower, on the whole, income-generating capacity of its earning assets (because of a smaller share in them of loans) compared with Bakai or Ecobank, Tolubai manages to earn larger interest income after loan loss provisions on its earning assets than any other bank in Kyrgyzstan . On the average for the last three years, this ratio is 31% in Tolubai, 27% in Bakai, 22% in Ecobank, and 22% on the average for the sample.
In sum, it can be said that as far as asset quality is concerned, Tolubai is a distinct and steady leader among the banks in Kyrgyzstan . The other two most efficient banks, Bakai and especially Ecobank, are considerably lagging behind Tolubai.
Growth
Tolubai, however, is outstripped by many banks in growth. Over 2001-03, the average annual rate of growth in assets was less than 8% with Tolubai, 21% on the average for the sample, 34% with Bakai, and 26% with Ecobank. Rapid growth of assets can be explained by new injections of equity capital, so to get a handle on what is called organic growth, we divide the average rate of growth in assets by the average rate of growth in equity. The resultant ratio, the average value of which for the sample is 1.03, is the highest with Energobank (1.14). Bakai and Ecobank have it at 1.09 and 1.10, respectively, while with Tolubai it is close to unity.
Noninterest Income
Another Tolubai's serious drawback, limiting its overall profitability, is a low level of noninterest income (income from services). On the average for the last three years, the ratio of noninterest income to deposits in Tolubai (3%, the smallest value among all the banks) is five times smaller than the sample average (about 15%), and more than nine times smaller than in Bakai (28%) which is in this respect far ahead of the majority of the other banks. In Ecobank this ratio stands at 25%.
Correspondingly, Tolubai is also falling behind all banks in Kyrgyzstan in terms of the share of noninterest income in total income. Moreover, since 1999 this ratio has decreased in Tolubai from 11% to 9% in 2003, while in Bakai it has grown from 35% to 69%, that is, nearly two times, and in Ecobank from 16% to 52%, or more than three times. On the average for the whole banking sector, the share of noninterest income in total income has increased from 29% in 1999 to 49% in 2003.
And this is rather bad than good-given the limited, if not miniscule, volume of formal financial intermediation in Kyrgyzstan . Because it testifies to a shift in the banks' activities towards income sources that are, strictly speaking, of secondary importance. If such a tendency strengthens and becomes entrenched, it may well retard the growth in the volumes of bank financing of the real sector, and, consequently, in the economy as a whole. In other words, when noninterest income increases to levels close to interest income, banks start to feel less necessity to develop their lending function. Especially damaging, such trends are in developing countries with their low or practically null access to alternative sources of debt finance.
The above-said does not mean, of course, that banks in Kyrgyzstan should give up earning noninterest income. But this is a question of measure: auxiliary activities can be welcome only as long as they do not weaken the main functions.
As to the optimal combination of lending and services in banking, there are three banks in Kyrgyzstan at present that have both the share of loans in total financial assets, and the share of noninterest income in total income at levels not lower than 90% of the respective banking sector averages, and at the same time are among the most profitable, value creating banks in the country (see Cost of equity capital and economic profit above). These banks are Ecobank, Energobank and Kyrgyzpromstroybank. The first of them, Ecobank, has 59% of its total financial assets in loans, and 40% of its total income in noninterest income.
But by return on assets, on average for the last three years, both Energobank (2.3%) and Kyrgyzpromstroybank (1.3%) are below the industry average (2.5%), whereas Ecobank (5.4%) is much above it. In this regard, Ecobank is substantially falling behind the leader, Bakai; however, taking into account that, as was not once noted before, the cost of funds in Ecobank is much higher than it is in Bakai, the latter's lead becomes less obvious.
The general trend among the banks in Kyrgyzstan is for the share of noninterest income in total income to remain less than 50% until the share of net loans in total financial assets falls below 46-48%. After this point (to which Kyrgyzpromstroybank is the closest), the share of loans starts to decline sharply, accompanied by almost as sharp an increase in the share of noninterest income.
Noninterest Expense
With the sector average for the ratio of noninterest expense to assets at 12%, Ineximbank, Tolubai and Demir each have it at about 9%. More efficient in this regard is only the Bank of Asia (8%). Fairness requires a mention of AsiaUniversalBank that, technically, has this ratio at even a lower level of 4%. However, taking into account the strikingly specific composition of this bank's assets (see Asset quality above), using its performance indicators as benchmarks of any kind is deemed dubious. In Bakai and Ecobank the ratio of noninterest expense to assets is about 16%. But by the standards of more developed, not necessarily Western, countries even the 8% recorded for the Bank of Asia is too high. In the leading Kazakh banks, for instance, such as Kazkommertzbank or TuranAlem, noninterest income relative to assets is within 3.5-5.0%.
Deposits
Tolubai and Bakai are quite close to each other by the dependence on deposits for their financing, that is, by the ratio of deposits to assets, but are very different in terms of the composition of deposits. On the average for the last three years, the share of savings deposits in total deposits is 89% in Tolubai but only 58% in Bakai, with the sector average standing at 44%. Relative to assets, savings deposits are at a level in Tolubai that is nearly three times higher than in other banks, on average. Demand deposits account for 49% of total deposits on the average for all the banks, 35% in Bakai, 31% in Ecobank, but only 7% in Tolubai. From bank management standpoint, a share of demand deposits as low as it is in Tolubai is a serious drawback and a burden on interest expense. The ratio of interest expense to assets in Tolubai is more than two times higher than the sector average.
Yield on Deposits
Interest-bearing liabilities include all such funds for the use of which a bank has to pay something. In addition to time deposits usually accounting for the main portion of interest-bearing liabilities, banks may use such financing as, for example, interbank borrowings. But if a banks does not resort to such sources of financing, if its interest-bearing liabilities are identical to time loans, the ratio of interest expense to interest-bearing liabilities is, obviously, the weighted average interest rate the bank actually pays on its deposits. There are only three banks in Kyrgyzstan , namely: Tolubai, Bakai and the Bank of Asia, that have had 100% of their interest-bearing liabilities in deposits for the last three years, and above 90% on the average for the last five years. This allows to make a comparative analysis of deposit yields in these banks without special adjustments. Parenthetically, it can be added that smallest shares in total interest-bearing liabilities are made up by time deposits in such banks as Kazkommertzbank Kyrgyzstan (21%), AKB Kyrgyzstan (47%) and Kyrgyzpromstroybank (58%).
The sharp increase in inflation in 1999 rendered the real deposit yields in all the three banks negative. The first to recover from this shock was Tolubai which started to pay real interest on its deposits as early as 2000. With Bakai, this happened in 2001, with the Bank of Asia-only in 2002. Over all the last five years, the real deposit yield has remained considerably higher than in the other two banks, although in the last two years this gap started to narrow. In 2002, deposits yielded, in real terms, 11.5% in Tolubai, 3.4% in Bakai, and 6.4% in the Bank of Asia; in 2003-6.9%, 1.4% and 5.3%, respectively.
The real yield on deposits remained below the real yield on 6-month government securities up to 2001 in Tolubai and 2003 in the Bank of Asia; in Bakai it has remained in this condition till the end of 2003.
A.Khadjimouratov, Managing Director Granat Management Consulting Firm
(1) If not indicated otherwise, the computations herein are based on the banks' published financial statements.
(2) The Banker, April 2003.
(3) Herein, the average ratios of banks are calculated as truncated means based on an average of banks within the 20th and 80th percentile values for a given ratio. Thus calculated, and due to the changes made by some banks in their previous-year financial statements, some of the averages shown in this chapter may deviate somewhat from their values indicated in the first issue of this journal.
(4) By the time of this writing, Kairat has already been privatized.
(5) Due to the computational technique used (see footnote 3 in the subsection Asset quality), the sum of the averages may not add up to 100%.
(6) Strictly speaking, the cost of borrowed capital (for the borrower) is lower than the return it brings to the lender by the size of the income tax rate, since interest paid is a deductible expense.
Prepared by the Consulting Firm Granat
Supported by Eurasia Foundation, funded by USAID
"Management Horizons". Issue N 2, 2004
Back to Top