MARKET CONCENTRATION OF BANKS AND UNEMPLOYMENT: EVIDENCE FROM THE BANK MARKETS IN BOSNIA AND HERZEGOVINA AND SERBIA

U ovom radu fokusirali smo se na istraživanje odnosa između koncentrisanosti bankarskog sistema Bosne i Hercegovine i Srbije, te uticaj na zaposlenost/nezaposlenosti u obe posmatrane zemlje. Osnovni zadaci i ciljevi istraživanja se odnose na determinisanje uticaja internih faktora banaka na zaposlenost/nezaposlenost, te prepoznavanje efekata unutrašnjih faktora na stopu nezaposlenosti. Kao zavisnu varijablu u ovom istraživanju koristili smo stopu nezaposlenosti, dok su kao nezavisne varijable poslužile sledeće: HHI indeks koncentracije, stopa rasta ukupne bankarske aktive, stopa rasta likvidne aktive prema ukupnoj aktivi. U cilju proučavanja kointegracije varijabli korišćena je metoda automatske regresijske raspodele, kao i metoda korelacije i regresije. Period istraživanja obuhvata period od 2008q1 do 2018q4. Rezultati istraživanja su pokazali da najsignifikantniji značaj na smanjenje nezaposlenosti mogu imati sledeće nezavisne varijable: stopa rasta ukupnih kredita, stopa rasta bankarske aktive, koncentracija kredita i depozita (posebno slučaj banaka u BiH). S druge strane, preveliki obim likvidnih sredstava prema ukupnim sredstvima iznad zakonskog minimum može uticati negativno na privredni rast i zaposlenost.


Introduction
Commercial banks have a significant impact on the development of other economic sectors in a society. Opening of banks intended for financing fast-growing sectors in the economy affect both the increase in investments and the increase in employment. Market concentration is one of the most important measures of competitiveness of the economy. Regarding the banking sector, competition is very important because, as in any other market, it affects the efficiency and quality of the offered services, and also affects other sectors of the economy. Moreover, the concentration of banks affects better efficiency in the context of fundraising, but on the other hand, it reduces the distribution of loans. In this regard, if clients are more oriented towards better banks, the approval of loans is more rational and is reduced, allowing for debtors with special conditions to get a loan.
The financial model in Bosnia and Herzegovina is bank-centric, characterized by a very high level of competition. On the other hand, in addition to the competition in the banking system of BH, there is also a prominent concentration. For the purpose of illustration, in the banking system of the Republic of Srpska, there are four large and four small banks which are dominant, with the banks holding almost 80% of the banking market of the Republic of Srpska. In the Federation of BH, the five largest banks hold almost 74% of the loan and deposit market. The financial sector of the Republic of Serbia is also modelled from commercial banks and is a key factor in maintaining financial stability. At the end of 2017 in the total sum of the financial system, the banking sector accounted for a structural share of about 90.7%.
In Bosnia and Herzegovina, with a view to strengthening financial stability, one of the prerequisites that were set up in 2002 was the restructuring and privatization of the banking sector. In the same year, the total number of privately owned banks amounted to 34 (mostly domestic private ownership), with six state-owned banks, only in the Federation of BH. The total number of employees as of 31 December 2002 was 7,519. After fifteen years, in 2017, the total number of privately owned banks amounted to 20, while the number of state-owned banks was 3. The total number of employees amounted to 9,624. Thus, there were tendencies to reduce the number of stateowned banks, as well as the total number of banks, accompanied with a slight increase in the number of employees. At the end of 2004, there were 43 commercial banks in Serbia with 23,463 employees. Out of this total, 11 banks had the majority ownership by foreign shareholders, 18 banks had the majority ownership by domestic individuals and legal entities, and 14 banks were majority owned by the Republic of Serbia. By comparison, at the end of September 2017, a total of 29 banks operated with a total of 23,055 employees. In other words, there was a decrease in the number of banks by 14 commercial banks, as well as a slight decrease in the number of employees.
The average unemployment rate in the EU 28 in 2017 was the lowest since the onset of the global economic crisis, amounting to around 7%. Bosnia and Herzegovina, as well as the neighboring countries, according to the surveyed unemployment rate, recorded a downward trend in the unemployment rate. Despite the lowered unemployment rate, Bosnia and Herzegovina and Macedonia recorded the highest unemployment rates. The unemployment rate in Bosnia and Herzegovina at the end of 2017 amounted to about 20.5%, while in the Republic of Serbia it was around 13.5%.
The main goal of this paper is to investigate how much the privatization of banks, along with the influence and entry of foreign banks in Serbia and Bosnia and Herzegovina, had an impact on the reduction of the overall unemployment rate. The paper consists of five parts. The first part refers to the introductory considerations in terms of the structure of the banking system, ownership, privatization, unemployment. The second part deals with the reference literature in the segment of positive/ negative implications of the concentration of banks on economic growth. The third part provides a short theoretical review of the basic concentration measure (HHI index) and its analysis across all three balance sheet categories, as well as a short time analysis of unemployment. The fourth part refers to the data necessary for the analysis. The fifth part Niži nivo konkurencije utiče na smanjenje tržišne efikasnosti kroz povećanje monopolskih profita (Berger, 1995). Hipoteza efikasne strukture (ES) navodi da je koncentracija rezultat promenljivih uslova u kontekstu troškova, a ne izmeni visine ulaznih barijera. Dakle, preduzeća koja imaju efikasniju proizvodnju bolje posluju, te time postaju i veća, čime im se povećava tržišni udeo, pa tržište postaje koncentrisanije (Demsetz, 1973

Review of Reference Literature
The structure of the banking market in recent years has attracted considerable attention from researchers and economic policy makers, and several studies have pointed to the role of the banking market structure in the context of bank efficiency, financial stability, monetary policy transmission, economic growth, and access to credit. According to the classic economic theory, banking concentration is detrimental to economic growth, as incomplete competition is associated with inefficiency, which makes it difficult for companies to access credit. According to Bain (1951), a high concentration of market power leads to higher prices and higher profit margins.
A lower level of competition affects the reduction of market efficiency through the increased monopoly profits (Berger, 1995). The Effective Structure Hypothesis (ES) states that concentration is the result of the changing conditions in the context of costs rather than altering the height of the input barriers. Hence, the companies that have a more efficient production operate better and thus become larger, thereby increasing their market share, and making the market more concentrated (Demsetz, 1973). Claessens and Laeven (2004) state that the increased competition in the banking sector is associated with the faster growth of other sectors of the economy that rely heavily on external financing.
In terms of concentration fragility, larger banks in a concentrated market weaken the stability through three channels: a) larger banks are considered large enough institutions to fail, where they receive guarantees from the government, which increases the moral risk, in consequence, b) larger banks are primarily oriented towards charging higher interest rates because of their market power, leading to risky placements taken by borrowers in order to compensate for such high interest rates, resulting in a high risk of borrowing, and c) managerial efficiency such as risk diversification in assets and liabilities may decline which affects the high level of operational risk (

Measures of Market Concentration of Banks and Unemployment Rate
The Herfindahl Hirschman Index (HHI) is a widely accepted measure of market concentration where the index value is obtained when the market shares of each institution (in our case, banks) are individually squared. The HHI index is expressed as follows: The table below illustrates the HHI index for the placements, deposits and bank assets of the Federation of Bosnia and Herzegovina for the period: 2008q4 -2018q4.    Na tržištu rada u zemljama EU registrovane su niže stope nezaposlenosti u 2018. godini, gde From the above table it can be noted that the HHI index in terms of loans recorded the highest value in the fourth quarter of 2008 (1,806), the lowest in the last quarter of 2012 (1,427) and the average value of 1,609. These values point to the conclusion that the banking market in the FB&H is moderately concentrated with a dominant share of 3 to 5 larger banks in the system, which have an oligopoly position. HHI index in terms of deposits reached the highest value in the last quarter of 2017 and 2018 (1,818), the lowest in the last quarter of 2013 (1,532) and the average value of 1,672. If we observe the average value, then we also come to the conclusion that this is a moderate concentration, but in terms of the maximum value in two years, the market could be concentrated. In fact, as one of the main reasons for high savings in up to three banks, there is a high degree of trust on the one hand, while on the other hand, the possible reasons could be: no alternative savings institutions, insufficient development of the financial system, high level of risk, lack of quality projects, etc. For illustration purposes, the overall trend of savings deposits as a base for the financial potential of banks in the FB&H at the end of 2017 amounted to about 4.22 billion euros. Nearly 55% of savings are in the two largest banks in the banking system in the FB&H, while the remaining five banks have an individual share of less than 2.0%.
The banking sector of the Republic of Srpska recorded slightly higher values in terms of the HHI index for all three categories of the balance sheet banking positions. The highest value of the HHI index in terms of loans was recorded in the fourth quarter of 2008 (2,148), while on the other hand the lowest value was recorded in the fourth quarter of 2013 (1,479). The average value of the HHI index in the loans category amounted to 1,790, which indicates the presence of moderate concentration in the system. When it comes to the movement of the index in terms of deposits, the highest value was recorded in the fourth quarter of 2017 (2,125), the lowest value was recorded in the last quarter of 2012 (1,558), with the average value of 1,778. Therefore, the movement of the HHI index in terms of deposits indicates the presence of a moderate concentration. Generally speaking, savings in banks in the BH have been recording positive growth every year, which on the one hand reflects an increase in confidence in the banking sector largely because of the deposit insurance system, while on the other hand, advantage is given to savings in relation to consumption as a result of uncertainty in the context of future economic opportunities in the country. 1 Report on the situation in the banking system of the Republic of Srpska as of 31 December 2018 is not available.   Prema Anketi o radnoj snazi u Republici Srbiji, ukupna stopa zaposlenosti je iznosila u proseku oko 46,7% i bila je za oko 1,5% veća u odnosu na prethodnu godinu. Zaposlenost je imala najveće učešće prema uslužnim sektorima privrede, koji u najvećoj meri daju doprinos rastu bruto domaćeg proizvoda, dok je zaposlenost u poljoprivredi smanjena. Najniže prosečne stope nezaposlenosti ostvarene su u Rumuniji od oko 6,48% i Bugarskoj od oko 9,48%.
Iz prethodnog grafika se vidi da je prosečan broj zaposlenih u jednoj banci u Srbiji imao prosečno volatilan trend, za razliku od Bosne i Hercegovine gde je u poslednjem kvartalu In contrast to the balance sheet categories of commercial banks (loans, deposits, and assets) in BH where moderate concentration was recorded, there is no concentration in commercial banks in Serbia for all three observed balance sheet categories. One of the reasons could be a higher number of domesticowned banks (at the end of the fourth quarter of 2010, there were 10 domestic-owned banks, while at the end of the fourth quarter of 2017 this number was reduced to 8 banks).
The labour market in the EU countries registered lower unemployment rates in 2018, with a rate of 6.8% in the EU28, while in the EU27 the rate was also 6.8%, which is lower than in 2017 by about 10%. It is evident that the highest average unemployment rate after Macedonia (28.93%) was recorded in Bosnia and Herzegovina, standing at about 25.89%. In the observed period, BH had the lowest unemployment rate of 23.4% in 2008, and the highest of 28.0% in 2012. The rising unemployment was not reflected during the global economic crisis, but two years after the crisis, with BH recording a rising trend in the unemployment rate up until 2012, due to the weakened economic activity of the real sector, the new labour force and the long-term status of the unemployed persons. A slight drop in the unemployment rate in BH was achieved only in 2016, and the unemployment rate in 2017 stood at 20.5%. The third place in terms of the average unemployment rate belongs to Serbia with a rate of around 18.44%. The lowest unemployment rate in Serbia was recorded in 2017 (13.5%), while on the other hand the highest unemployment rate was registered in 2012 (24.0%). Until 2012, a growing unemployment trend was evident in Serbia, which has been replaced by a declining trend from 2013 to 2017. According to the Labour Force Survey in the Republic of Serbia, the total employment rate was on average about 46.7%, which was by about 1.5% higher than in the previous year. Employment had the largest share in the service sectors of the economy, which largely contributed to the growth of the gross domestic product, while employment in agriculture was reduced. The lowest average unemployment rates were recorded in Romania, about 6.48% and Bulgaria, about 9.48%. The figure below analyses the tendency of movements in the average number of employees in one bank in Serbia and Bosnia and Herzegovina for the period 2008q4 -2017q4.
Stopa rasta likvidne aktive prema ukupnoj aktivi -proračunava se tako što se uzmu u razmatranje najlikvidniji delovi aktive banaka, te podele ukupnom aktivom i isti iznos pomoži sa 100. The figure above shows that the average number of employees in one bank in Serbia had an average volatile trend, unlike Bosnia and Herzegovina which in the last quarter of 2017 recorded a higher number of employees in one bank than in the initial observed period. Therefore, this is not a realistic situation given that since 2016 there have been changes in terms of bank mergers, resulting in the concentration of the number of employees in banks with higher market shares. For illustration purposes, the total number of banks in BH in the fourth quarter of 2008 was 30, while the total number of employees amounted to 11.054. In the last quarter of 2017, the total number of commercial banks in BH was 23, while the total number of employees amounted to 9.623. Therefore, there has been a decrease in both the number of banks and employees due to various factors such as the impact of the global financial crisis, the rationalization of operations, the strengthening of market influence through status changes of mergers, loss of operating licenses, etc. During the global financial crisis and in the post-crisis period, foreign commercial banks in Serbia began to tend to partially reduce the number of employees, while domestic banks in the same period received fewer new employees. As

Data
This analysis focuses on the following variables: the dependent variable will be the rate of unemployment in BH and Serbia (GRUN), while the growth rate of total banking assets (GRTBA), the total growth rate (GRTL), the growth rate of liquid assets to total assets GRLATA), the banking concentration of loans, deposits and assets measured through the HHI index (BCloans, BCdeposits, and BCassets) will be used as independent variables. The selection of the observed variables and analysis is the high unemployment rate in Bosnia and Herzegovina on the one hand and the contribution of the banking sector to reducing the unemployment rate on the other. The survey covers the period from the first quarter of 2008 to the fourth quarter of 2018. The analysis will be based on quarterly data because the annual level data are an unreliable sample for econometric evaluation. The data have been collected from the official pages of the Central Bank of Bosnia and Herzegovina, the Banking Agency of the Federation of Bosnia and Herzegovina, the Banking Agency of the Republic of Srpska, Agency for Statistics of BH, National Bank of Serbia and Statistical Office of the Republic of Serbia. The measurement and expected effects of dependent and independent variables are given below. Stopa rasta ukupne bankarske aktiveproračunata je kao razlika bankarske aktive tekućeg kvartala u odnosu na prethodni kvartal, te dobijeni iznos podeljen sa prethodnim kvartalom i pomožen sa 100.
The growth rate of liquid assets to total assets (GRLATA) -It is calculated by considering the most liquid parts of bank assets, dividing them by total assets, and multiplying this amount by 100.
The growth rate of total banking assets (GRTBA) -It is calculated as the difference between the banking assets of the current quarter and the previous quarter, divided by the amount from the previous quarter and multiplied by 100.
The growth rate of total loans (GRTL) -It is calculated as the difference between total loans of the current quarter and the previous quarter, divided with the amount from the previous year and multiplied by 100.
Bank concentration measured through Herfindahl-Hirschman index -HHI can take different values in an interval from 0 to 10,000. The concentration index does not grow linearly, which means that, for example, an amount of 3,000 does not mean that the concentration in the system is 30%. If the concentration index ranges from 0 to 1,000, the market is considered to be unconcerned, that is, a high level of competition is present.
If the index ranges from 1,000 to 1,800, then it can be said that there is a moderate concentration of the market. And if the index ranges from 1,800 to 10,000, the market is concentrated, i.e., there is a monopoly (Banking Agency of the Federation of Bosnia and Herzegovina, 2017).

Methodology
In order to test the stationing of the variables, we will use the ARDL methodology as well as the method of correlation and regression. In order to investigate the impact of banking concentration, internal and external factors of the banking sector in relation to employment/unemployment in Bosnia and Herzegovina and Serbia, the following model was used for short-and long-term analysis:

+ β 4 HHIasset t + β 5 HHIcredit t + β 6 HHIdeposits t + ε t (2)
Where: GRUN t -unemployment rate in time t GRLATA t -the growth rate of liquid assets to total assets at time t GRTBA t -growth rate of total assets of commercial banks in time t GRTL t -the growth rate of total bank loans at time t HHIasset t -bank concentration of assets measured by the HHI index at time t HHIcredit t -bank concentration of loans measured by the HHI index at time t HHIdeposits t -bank concentration of deposits measured by the HHI index at time t ε t -regression residual at time t.

Bosnia and Herzegovina
The highest value of the arithmetic mean was achieved at the unemployment rate of about 516%, followed by the growth rate of liquid assets to total assets of around 27%, the growth rates of banking assets of 23% and the growth rate of loans of about 16%. The greatest volatility in terms of the first-order statistical indicators, i.e. standard deviations also recorded an unemployment rate of about 32%, which is followed by the same trend as the arithmetic mean. The second place in terms of standard deviation was taken by the growth rate of banking assets (3.03%), and the third by the rate of growth of liquid assets (2.48%), and the growth rate of total loans of about 1.78%.
The share of liquid assets to the total assets of banks in Bosnia and Herzegovina reached its maximum value of about 41% in 2007, while in 2013, liquid assets to total assets amounted to around 19% and in 2018 to around 29%. It is evident that commercial banks in BH are more than liquid, where, due to the lack of quality projects and investments, they rationalize their placements on the one hand, and hold inactive liquid assets on the other hand. This is one of the main reasons of the slow economic growth and development, as well as employment. One of the possible solutions for the activation of banking placements is the establishment of a domestic institution that would award ratings to the most successful companies and monitor their effectiveness and efficiency. Thus, the excess liquidity of banks would be relieved and the banks' confidence in one part of the real sector of Bosnia and Herzegovina increased. One of the more precise measurements of volatility, i.e. skewness, can be defined as the ratio of the third point around the middle and the standard deviation to the third potency. In a normal distribution, the skewness is 0, i.e., the likelihood of a negative or positive yield is equal. In our case, all independent variables had a slightly higher value than zero and a positive value, except for the unemployment rate, which recorded a slightly higher value than zero of the negative sign. The sign (-) signifies a sharper tail and a short tail distribution on the right-hand side and alleviates the slope and prolonged tail distribution on the left side. Unlike skewness, kurtosis is defined as the fourth-point ratio around the middle and the standard deviation to the fourth potency. Fluctuations in finance are used to describe extreme events, that is, those events that fall within the distribution. In the case of normal distribution, kurtosis equals 3. The value above 3 has only been recorded by the growth rate of liquid assets relative to total assets.

48
The table below shows the results of the correlation matrix between dependent and independent variables.
From the previous correlation matrix, it can be seen that the strongest correlation (in terms of positive strength) with the dependent variable, i.e. the growth rate of unemployment was achieved with the following independent variables: the index of concentration of commercial bank assets in BH -HHI (0.656), and the growth rate of liquid assets according to total assets -GRLATA (0.332). This result is completely justified and logical, as with the increase in their liquid assets the banks rationalize in the domain of granting new loans to different sectors in the economy, which leads to managerial inefficiency. This is, in turn, justified by the fact that the process of selecting debtors in the BH economy is hampered by bad debtors, which further partially affects the retention of the existing staff, i.e. the lack of employing new managers in the banking sector.
In order to solve the excessive liquidity of the banking sector in Bosnia and Herzegovina, the first step is to form a domestic institution that will give credit ratings to the best companies on the basis of the analysis of all business indicators and business transparency. On the other hand, the strongest negative causality was recorded in the following independent variables: the HHI in terms of deposits (-0.607), the HHI in terms of credit (-0.527), and the growth rate of the commercial bank assets in Bosnia and Herzegovina (-0.502). All three independent variables are interdependent, as the increase in assets and deposits is the basis for greater credit expansion, which consequently leads to an increase in the number of employees in banks and partly to a reduction in the unemployment rate and indirectly to an increase in the employment rate of individual companies from the real sector through increased distribution lines. It should be noted that the distribution of loans according to the branches of activity is disproportionate. For the sake of illustration, the concentration of loans by business sector in the Republic of Srpska in the third quarter of 2017 indicates the continuity of the loan concentration in the manufacturing sector at about 15%, in trade at about 13%, in the sector of the Government and government institutions at about 12%, in the construction sector at only 4%, catering only 1%, loans for general consumption 31%, housing loans 11% and the remaining 13% (accounted for by agricultural loans with 2%, financial intermediation, real estate, rental, business services, etc.). The most significant number of employees at the end of 2017 related to the following branches of activity: the field of industrial production (especially the manufacturing industry), trade and hotel  Uglavnom, veći deo tih sredstava je u obliku deviznih depozita koji se drže u inostranstvu na računima stranih banaka beskamatno ili uz vrlo nisku kamatu.

Srbija
Kretanje vrednosti aritmetičke sredine je fluktuirajućeg karaktera u kontekstu posmatranih varijabli. Najveća vrednost aritmetičke sredine je zabeležena kod stope rasta bankarske aktive, zatim kod stope rasta The results of the stationary test showed that most of the variables are in the model of a stationary type, except for the series of the following symbols: GRLATA, HHIASSET, and GRTL. Table 9 illustrates the results of the econometric analysis between the unemployment rate in Bosnia and Herzegovina as dependent variables and the following independent variables: growth rates of liquid assets relative to total assets, growth rates of total banking assets, growth rates of total loans, concentration index -HHI for loans, deposits and assets. The total number of observed observations in the model is 48. The correlation coefficient between dependent and independent variables is 0.73, while the adjusted correlation coefficient is 0.70. Therefore, this is a medium strong correlation.
The most significant variables have shown the following: GRTBA and HHICREDIT. Both independent variables have a negative sign, which means that the increase in the growth rate of total banking assets, as well as the increase in the concentration of loans measured by the HHI index with other unchanged conditions, are followed by a decrease in the unemployment rate for certain branches of business.
Therefore, with the increase in the rate of growth of banking assets for one unit, with other factors unchanged, there is a consequent decrease in the unemployment rate by about 10.83 units. A stronger impact on the reduction of the unemployment rate based on the basic econometric model was achieved by the HHI concentration index in terms of loan placement. Therefore, with the increase in HHI for loans for one unit, with other factors unchanged, the unemployment rate is reduced by about 282 units. From the first quarter of 2008 to the fourth quarter of 2018, the assets of commercial banks in BH increased by about 58%. Also, capital participation recorded a trend of growth given the fact that 2002 and the following years were the years of more intensive capital investment in banking institutions, as well as the year of privatization and formation of new banks. If we observe loans, the relative growth of total loans in the last quarter of 2018 was by about 57% higher than in the first quarter of 2008, and only by 5.7% higher than in the last quarter of 2017. On the other hand, the least significant variables in the model turned out to be the growth rate of liquid assets to total assets, as well as the HHI index in terms of assets. With 2 Stationarity refers to the constancy of a series of data over time, primarily through statistical measures: mean values and variances. ukupnih kredita, stope rasta nezaposlenosti, te stope rasta likvidne aktive prema ukupnoj aktivi. Najveću volatilnost u pogledu statističkih merila prvog reda, tj. standardne devijacije je isto zabeležila stopa rasta bankarske aktive, što prati isti trend kao i aritmetička sredina. Drugo mesto u kontekstu standardne devijacije pripada stopi rasta ukupnih kredita, zatim stopi rasta nezaposlenosti, te stopi rasta likvidne aktive prema ukupnoj aktivi (tabela br. 10). Od posmatranih statističkih merila drugog reda mera asimetrija je za sve posmatrane varijable imala negativni predznak, gde predznak (-) ispred varijable označava oštriji nagib i kratak rep distribucije na desnoj strani, i blaži nagib i produženi rep na levoj strani distribucije. Najveće vrednosti spljoštenosti ostvarile su stopa rasta nezaposlenosti od 4,22 i stopa rasta likvidne aktive prema ukupnoj aktivi od 3,42. U slučaju normalne distribucije spljoštenost iznosi 3. Tabela 11. prikazuje dobijene rezultate Dickey-Fuller testa.

Serbia
The movement of the arithmetic mean is of a fluctuating character in the context of the observed variables. The highest value of the arithmetic mean was recorded in respect of the rate of growth of banking assets, then the growth rate of total loans, the rate of unemployment growth, and the growth rate of liquid assets to total assets. The highest volatility in terms of first-order statistical indicators, i.e., the standard deviation also recorded the growth rate of banking assets, which follows the same trend as the arithmetic mean. The second place in the context of standard deviation was taken by the growth rates of total loans, then the rate of unemployment and the growth rate of liquid assets to total assets (Table 10).
Among the second-order statistical criteria, skewness had a negative sign for all observed variables, with the sign (-) in front of the variable signifying a sharper tilt and a short distribution tail on the right side, along with the slighter slope and extended tail on the left side of the distribution. The highest values of kurtosis were reached by the unemployment rate of 4.22 and the growth rate of liquid assets to total assets of 3.42. In the case of the normal distribution, kurtosis is 3. Table 11 shows the results of the Dickey-Fuller test.
The results of the stationary test showed that most of the variables are in the model of a stationary type, except the HHI concentration index in terms of assets. The table below shows the correlation between the dependent and independent variables in the model. According to the results of the correlation analysis, the independent variables that had the strongest impact on the reduction of the unemployment rate were the following: the growth rate of total
The largest structural distribution of loans in 2018 was related to the area of transport and communications, manufacturing and real estate business. An increase in the growth rate of liquid assets to total assets for one unit, with other factors unchanged, leads to an increase in the unemployment rate of around 4.92 units. For the analyzed period, the largest liquid assets rate according to the total assets of banks in Serbia was achieved in the first quarter of 2008 (36.93%) and in the third quarter of 2008 (36.01%). On the other hand, the lowest values of the growth rate of liquid assets to total assets were realized in the second quarter of 2017 (13.11%) and in the first quarter of the same year, about 13.31%. The banking sector of Serbia is extremely liquid 3 , as is the banking sector in BH. However, the following question arises: Can the excessively high liquidity, higher even than the legal minimum, be a limiting factor for further economic growth, structural changes and steady growth of unemployed persons? The business banks in Bosnia and Herzegovina are looking for an answer to the lack of quality investment projects, as well as the poor paying capacity of debtors. Other characteristics of the banking business performance in Serbia are related to the variable maturity of deposits, poor collection of loans and bad assets, limited capital inflows from abroad, the lack of adequate regulation and the underdevelopment of financial markets.

Conclusion
According to the theory of the concentrating influence on developed financial markets, increased concentration can affect the creation of a monopoly position where banks set highinterest rates, and their efficiency decreases with regard to the distribution of placements to different branches of the economy. Also, the increased concentration of commercial banks has an impact on the distribution of resources among the various economic sectors, so the unemployment rate is also different among the economic sectors. The concentration of commercial banks in BH is largely moderated by the tendency of growth, the dominant influence belonging to the 3 to 5 largest banks. The banking sector of the Republic of Serbia is characterized by a satisfying degree of competition and a low concentration of activity.
The research results show that most of the observed variables are based on the stationary type ADF model. Correlation analysis and basic model results show that the strongest correlation with the growth rate of unemployment as dependent variables was achieved through the following independent variables: the HHI index of credit and deposit concentration with unbalanced distribution across all branches of the economy, then the rate of growth of loans and growth rates of banking assets. The unreasonable distribution of placements is primarily due to the presence of moderate concentration in some years (in the case of Bosnia and Herzegovina) and the uncertainty of investment projects. On the other hand, the weakest correlation was recorded with the following independent variables: the growth rate of liquid assets to total assets and the HHI concentration of assets. The banking sectors of Serbia and Bosnia and Herzegovina are more than liquid, their capital adequacy rates being above the legally prescribed minimum. Also, the results of the basic model for the case of Serbia show that the strongest influence on the reduction of the unemployment rate was achieved by the HHI index in terms of credit, deposits and credit growth rates, while the correlation occurred with the following independent variables: the growth rate of liquid assets to total assets and the HHI index in terms of assets.
Commercial banks in Serbia and Bosnia and Herzegovina had a modest impact on increasing employment either directly through employment in the financial sector, or indirectly by lending to the real sector to start up new businesses. However, in the forthcoming period, banks, above all in BH, should focus on lending to the growing branches of activity through the release of high liquidity, all in 3 Business banks in Serbia are obliged to maintain a lower liquidity indicator at a level not lower than 80% by the end of 2017 and after 2017 at a level not lower than 100%.
Alihodžić A. Market Concentration of Banks and Unemployment: Evidence from the Bank Markets in Bosnia and Herzegovina and Serbia order to increase economic growth and reduce unemployment. Thus, the tactical reduction of liquidity in favor of a large increase in placements towards growing private sectors can result in greater effectiveness and efficiency, which will ultimately manifest as a reduction in the unemployment rate.
The successful response to a turbulent environment is forecasting the performance of business banks, and defining all the elements and instruments that performance can accomplish. For further research, it would be necessary to take a broader range of internal and external variables as well as an even longer data set.