The impact of financial reform : the Turkish experience

I/(pi INDUSTRY AND ENERGY DEPARTMENT WORKING PAPER INDUSTRY SERIES PAPER No. 65 The Impact of Financial Reform: The Turkish Experience .~~~~f . .- December 1992 i, j: . .;* -~~ ~ ~~~~~~~~~~~ -- , I . ,' } -.1 - . . I ' - ... The World Bank Industry and Energy Department, OSP  THE IMPACT OF FINANCIAL REFORM: THE TURKISH EXPERIENCE by Izak Atlyas and Hasan Ers * December 1992 * Izak Atlyas is from the World Bank and Hasan Ersel is from the Central Bank of the Republic of Turkey. This paper was prepared under the research project 'The Impact of Financial Reform," RPO 678-13, and is adapted from Chapter 12 of the draft manuscript, Financial Reform: Theory and Experience. The authors are grateful to Gerard Caprio, Ashl Demirguc-Kunt, Ercan Kumcu and Emin Ozturk for helpful comments. The views expressed In this paper are those of the authors and should not be attributed to the Central Bank of the Republic of Turkey or to the World Bank, its Board of Directors, its management, or any of its member countries.  THE MPACT OF FINANCIAL REFORMs THE TUR H XPE NCE TABLE OF CONTENTS Pae No. EXECUMVE3 SUN04ARY ............................................................ i L A BREF HISORY OF FINANCIAL REFORM IN TURKEY ...... 1 Li Financdal Daelopment Dudng the Pre-Reform Period. From Etatsm to Planned Development .. 1 L2 lhe Suategy of Faaclal Refodm. 2 13 Main Institutional Elements and Implementation of Financial Reform .. 2 IL DEVELOPENTRS IN FINANCIAL MARK.ETS .S 11.1 Inancial Deepening ..... . .6 11.2 Forein ExhangeDposits.. 6 113 The dominane of Ithe Public Sector ... 6 1L4 Inteit Rates... 11S Developments In the Banking System ..9 DL6 Bank Behavior in TurkeEvidence fom Survy Results ........ .......... 19 ZIL DE1VELOPMENTS IDN CORPORATE FINANCE ............................. 21 1I.1 Trends In the Compostion of LJabflities ............ .. ................. 21 M.2 Varistions in Corporate Indebtedness and the Impact of Fnancial Refom ......................................... *24 IV. LESSONS OP THE TURKSH EDXPERIENCE ....... ........................ 28 REFERENCES .......... 30 ANNEX TABLES .......... 32  THE IWACr OF FINANCLAL REFORM: THE TURKISH EXPERIENCE EXECUTIVE SUMMARY lbis paper examines the Impact of financial reform on the behavior of financial sector agents as well as on corporate sector. Financial liberalization along with trade reform, lay at the core of the comprehensive economic reform program launched in 1980 in Turkey. Initial liberalization attempts were interrupted by a crisis, in 1982 After the crisis Turkey adopted a more cautious approach to financial reform. Efforts were made to develop a legal and institutional famework for the functioning of financial markets and for the supervision of financial institutions. The Capital Markets Board became operational in 1983 to promote and regulate securities markets. A new banking law was enacted in 1985, providing the legal basis for prudential regulation and supervision. A Savings Deposit Insurance Fund was introduced in 1983. Foreign currency denominated deposits were allowed in 1984, and the determination of the exchange rate was progressively liberalized. Banking industty expanded until 1987; after 1987 bank assets/GNP ratio started to decline, mawily due to more serious prudential regulation and supetvision, resurgence of inflation and loss of credibility in macroeconomic policies. Concentration in the volume of deposits declined. Foreign exchange deposits expanded rapidly. Small and medium sized private bans exhibited high profitability and gowth Nevertheless, there were serious shortcomings in the Turkish experience as weLL Lessons that can be derived from the Turkish experience are the following. The first lesson is that any potential adverse effect of interest rate liberalization is smaller when the corporate sector is financially strong The new literature on the link between finance and real activity emphasizes the importance of borrowers' net worth in detmining the availabfilty of external finance. In their exposition of the recent theory, Qertler and Rose (1991) argue that a decline in borrowers' net worth is likely to reduce the extent of financial intermediaton. More generally, the financial condition of the borrowing sectors is seen to be a critical determinant of the degree of efficiency in the allocation of loanable funds. The experience in Turkey, especially the response of the financial and corporate sectors to interest shocks confirms this view. In 1982, when the corporate sector was experiencing a decline in gross margins, liberalization of interest rates created a vicious qcyle whereby distress bornwing by an illiquid corporate sector and accommodation of especially troubled banks resulted in unsustainably high interest rates and over- borrowing. By contrast, in response to the interest rate shock of 1988, the corporate sector could rely on internally generated funds to reduce their stock of short term loans and adjust their balance sheets. The absence of distress borrowing is closely linked to the absence of a separate gross earnings shock. The second lesson is that fiscal policies may limit the benefits of financial reform. The major shortcomig of the Turkish financial reform was maceconomic uncertainty and the burden that public sector borrowing placed on financial markets. Concern with macroewnomic uncertainty is evident in bank managers' responses to survey questions, where uncrainty is identified as a major deterrence to increased intermediation  I l - towards firms that are peroeived as riky. In addition, the Government's recourse to the financial system to fnance the budget deficit crowded out financial fos to the private sector. Hence, the potential positive Impact of financial reform was oDnsiderably hidered by a lack of discpline in government flnances. In fact, it can be said that the Government has been the main bewefiday of developments in fnancial markets. A related lesson Is that liberlization alone may be insufficient to increase and diversify aternl sources of funds in the corporate sector. The analysis of firm-level data reveals that firms' reliance on internally generated funds did not change over the reform period. The importance of size in accessing long term bank loans is also not diminished. Mhese findings are consistent with the limited impact of financial reform on the financing of the private sector. Nevertheless, there Is also some evidence that the Importance of agency problems and collateral requirements were reduced over the reform period, at least for the finms in the sample. The Turkish strategy of financial reorm emphasized Institutional development in so far as authorites tried to establish a reguatory environment for the development of foreign exchange and interbank money markets. However, overwhelming focus on the banking sytem led to the neglet of the development of non-bank financal institutions and instruments. Even though company data reveal the slow emergence of finamce bills as a new instrument of corporate finance, development of alternative sources of funds, including the stock market, has lagged behind significanty. This may be contrasted with Korea, where direct financing became a major component of sources of funds in the 198Xs Fialy, it also seems that a more efficient judical system may play an important positive role in expanding the benefits of financial reform The problem is not onl one of refmig the legislation that governs financal tLansactions, but perhaps more Important, one of incrasing the processing capacity of courts Strengthening the confidence of financial intermediares in the judical system may enhance their willingness to ecpand into new types of clients or areas of business, and develop more complex financial contacts.  THE IMACT OF FINANCIAL RE MOI TUIE TIRK[SU EXPERIENGE FInancial liberalization, along with tade reorm, lay at the core of the comprehensive economic reform program launched In 19 In Turkey. 7l1 purpose of this paper is to examne the impact of fancial liberalization on the bebavior of financial secor agents as well as on corporate sector. In the first section, the histoxy of the Turkish financial liberalization program is briefly summarized. The second section addres the changes in the structure of the financidal system and examines the behavior of banks. The corporate sector is examined in the fourth section, while the fifth section is devoted to the conclusions and the lessons that can be derived from the Turkish experien. L A BRIEF HISTORY OF FINANCIAL REFORM IN TURKEY L. Fnaial Development Durn the PrReform Petiod From Etatsn to Planed Deveopment As demonstrated in Akyuz (1984), the Turkish financial system during the 1970s was relatively underdeveloped even when compared to countries at simlar levels of industrializatiom Historical faciors such as the rather centralized hacter of the goverment apparatus in Turkey, as well as the etatist growth and industrialization strategy adopted In the 1930s in response to economic cisis considerably enhanced the role of government in economic lif Economic policies in the 19SQs which Iitilly attempted to increase the role of the market mechanism were unsuccesfl and led to a maore drastic, abeit uncoordinated, increase in the level and extent of government intervention. The search for coordination led the post 1960 regime to adopt a planned development strategy, which seemed to fit the countufs etatlst traditions It was hoped that such a planned approach to development will prevent inconsistencies that arise from short sightedness In economic decision-making. Me planned development model adopted in the early 196 placed a heavy emphasis on the real sector. The planning strategy in Turkey can be interpreted as assuming either the existence of a centralized mechanism to allocate financial resources or a perfectly accommodating financal system. However, in Turkey there was no central mechanism to allocrZ financial resouroes, and whether the financial system was accommodating to the imperatives of development plans was also highly questionable. Akyuz (1984) examined the performance of the Turkish economy for the 1971-1981 period and concluded that the wvorkings of the financial system was an important impediment in allocating savings Into investments according to the requirements of development plansY Until the first oil criss, Turkey succeeded in sustaining a relatvely high rate of growth under a stable economic environment. However, staing from 1974, and paricularly after 1977, the need for reforming the existing growth strategy as well as the allocation mechaism becme evident. After two unsuccesful attempts in 1977 and 1978, a comprehensive reform program was launched In 1980. 'Me main features of the program was a switch from import substitution to export promotion, trade reform, and the Liberalization of the financial sector. 1/ Fry (1972) reaches a similat conclusion for the First Five Year Plan period, 1963-1967.  -2- .2 The Strat of Financial Reform The Turkish refotm strategy was to promote financial market development through deregulation and induciag competition by easing entiy Into the banking sector. Opening up the banking system to foreign competition was seen as an Important element of enhancing competition. The reformers paid much less attention to promote non-bank financial institutions. In this sense, the Turkish reform strategy was overconfident in Its reliance on competition among bank in developing the financial markets. This strategy left the considerable dominance of banks in the fdnancial sector unchallenged, and in fact, and surengthened it by allowing them to perform in the newly emerging financial markets Ihe reforms broadened the spectrum of the activities of banks, consolidated and strengthened the existing unversal anking system. Within this system, ban were allowed to collect deposits, extend loans, underwTite securities, trade In securities, establish and manage mutual funds, manage their own as well as their customers' securities portfolios, participate in corporations and engage in foreign exchange transactions. The reasons for relying exclusivel on the banldng system and completely neglecting problems of conflict of interest is mainly attributable to, first, the popular confidence in banldng institutions, which increased especially after the so- caled bankerse crisis of 1982, and second, the political strength of these institutions.1 Being the only type of financial institution available, banks were sucoessful in exploiting opportunities offered by the liberalization program and diversed their activities into the newly emerging financial markets, particularly securities markets, which enabled them to broaden the domain of their dominance. Ihe existence of an inherited universal banldng system and complete absence of 'Chinese Walls or 'Fire WallsW among various financial activities even In the post-liberalizaton regulations, enabled banks to impede the emergence of non-bank competitors in the financial markets. 1.3. Main Intitutona Elements and ImplementatIon ofFinancia Reform Prior to the reform initiative, the Turkish financial system was characterized by features typical of financial repression: There were ceilings on interest rates for deposits and credits. Real interest rates were negative, liquidity and required reserve ratios were high, preferential credits existed and were subject to subsidies. The public sector deficit was financed, to a great extent, by monetization, that is, by direct advances from the Central Bank. Entry into the banking system was restricted. Foreign exchange operations were constrained. The liberalization program was launched in January 1980 by the then ruling Justice Party government. However, the program became effective almost one year later, under the military rule. yW Some attempts were undertaken to create competition through diversity in the early 1980s, by promoting non-bank financial institutions, or 'bankers" as they were called in Turkey. These institutions were much less regulated than banks, and played an active role In precipitating the financial crisis of 1982. See Atiyas (1990).  -3 - Deregulation in the financial markets began wth the abolishment of Interest rate ceilings on loam and deposits and the introduction of certificates of deposits (CDs) In July 1980Y Simultaneously, in order to curb inflation, a tight monetary policy was followed. The reduction in aggregate demand caused a deep delne in corporate earnags. Distressed borrowing by financially fagile companies further deteriorated their balance sheets and put an upwards pressure on interest rates Meanwhile, espedally smaller and financially weake banks engaged In a fierce competition for deposits to finance non-performing loans Accompanied by a persistent reduction in the rate of inflation, real interest rates on deposits reached 20 percent in 1982 Neither the policy makers, nor the financial intermediaes and corporations were ready to deal with the dynamics of competition in an unstable environment. In particular, the regulatory bodies of the Government, especialy the Cental Bank, were not capable to monitor closely the behavior of banksy The developments in the financial markets led to a major crisis in 1982, and the liberization process was partly reversed. The government intervened In and closed down five small private banks The majority of their assets and liabilities were taken over by state-owned bank. In 1983 the Central Bank reregulated the deposit rates. The 1980-1982 period can be considered the 'infancy phase' of the Turkish liberalization program. In this phase, the main dynamic element behind the liberalization efforts was deregulation. The authorities, naively, hoped that the competitive eoonomic environment created through deregulation would, in a relatively short time, increase the efficiency of the financial system in alocating resources. The 1982 crisis ended these hopes and the reformers shifted their emphasis to the creation of the Institutional framework necessary for the working of a market system. Consequently, the second phase of the liberalization program (1983-1987) was more consistent in the sense that policy makers recognized the need for laying the institu- tional foundations of the financial system. In this framework, the Capital Market Board (CMB) was established In 1982 and became operational in 1983 to promote and develop securities markets. The Capital Market Law of 1981 empowered the CMB to regulate primary markets for equities and bonds. The Idw envisaged a *merit system" rather than a "disclosure system' for the issuance of securities; that is, It authorized the CMB to reject an issue whenever its analysis concluded that the financial soundness of the issuer was unsatisfactory. In 1983, secondary market operations were regulated by a decree, and within this framework the Istanbul Stock exchange was reopened in 1985 and became operational in 1986. There are two distinguishing features of the securities market regulation in Turkey, the first is the role of the banks. The regulations did not attempt to restrict banks from engaging in any type of ac- tivities either in the primary or in the secondary markets. On the contrary, in some instances, such as in the case of establishing and managing mutual funds, banks were granted a monopoly position. The second feature is that it created a central authority (CMB) to control and monitor effectively developments in the securities markets. It is evident that this choice was not in line with the general attitude of the reformers since it N For more details in this period, see Akyuz (1990) and Atiyas (1990). 41 Besides raising deposit rates, banks attracted funds by issuing CDs through non-bank financial institutions. The Central Bank had no means of keeping track of the volume CDs issued.  -4 - grted the CMB discaetdonazy powems. Thi diergence ca be explained, panly, by the shocking eerience of the 1982 crisis, which caused a deteroration In popular confidence for non-bank financial institutions. The second important development was the enactment of a new Banking law In 1985. The new banldng law was aimed at improving the struca weakness of the bankiag system which abruptly manifested itself during the 1982 acisIL The new law Introduced a provision for a mnimum capital base for banks In addition, the law established a capital adequacy ratio (5 pecent in 1989, to reach 8 percent in 1992). The rato is calculated under the BIS guidelines for determining prinmay and secondaty capital and risk weights. Credit extended to a single customer was limited to 10 pert of bank equity capitaL Investment in partidpations are limited to 100 perent of capitaL The law also obliges banks to use a uniform chart of acoounts In 1983 a Savings Deposit Insurance Fund was introduced to prevent the reemergence of the liquidity problams faoed by the banks in the 1982 cdsis However, the coverage of the insurance system is limited, (currently 100 percent of the fist TL2S mion and 60% of the next 25 mIllon). Deposits over TL50 milon are not covered. The insurance fund is not authorized to assist the liquidation and rehabilitation of fnancialy weak banks Such authority lies with the Treasury. Ihe cwrent banking law does not aliow the Government to take over the ban inject new capital when necessary, or buy the non-performing assets of the bank. The law does authorize the Government to change the management of problem banks, and to implement measures to improve their liquidity.Y Another novelty of the law was the introduction of the definition of non-performing loan. The law forced banks to report non-perfo1rming loam separately and required them to cover the defaulted loans through pvsions. Finaly the law introduced a standard accounting system and obliged external auditing of the banks The Banking Law also authorizes Swom Bank Auditors (SBA) associated with the Treasury to examine banks' legal compliance and financial standing. SBA carry out on-site examinations once every two years, though the frequency is higher for problem bank. In addition, the Banking Department within the Treasury monitors the financial performance of banks through quanerly financial statementsY The Central Bank also has a role in supervision. The Bank Supervision unit In the Central Bank, established In 1986, carries out off-site supervision by following about 50 quarterly and monthly reports The Central Bank's analysis concentrates on capital adequacy, asset quality, profitability and liquidityY Finally, banks are re- A draft Banking Law envisages to replace the Savings Deposit Insurance Fund with a Deposit Insurance Corporation, and endows it with wider powers to deal with problem banks. As of 1990, there were 36 active SBA and 9 assistant SBA. II As of 1991, there are S inspectors and 20 examiners at the Central Bank, excluding data processors and clerical staff.  - 5.- qWred to have an external audit evey yeai by independent auditing flri authorized by the Treasury and the Central BankllY The Central Bank established an interbank money market in 1986. Unti 1987, the Central Bank acted as a blind broker. Later, howeve, the Cental Bank started to get Involved il transactions as a dealer, in order to solve short term liquidity problemLsY In May 1985, govranent securities bepn to be auctioned on a weekly basis by the Cental Bank In 1987, the Central Bank started open market operados In the area of foreigp exchange, residents were allowed to hold foreigl cuency denominated deposts in 1964. Banks were allowed to keep foreign curency abroad. In 1984, banks were allowed to set their exchange rates within a margin around the Central Bank rate. Further steps to liberalize exchange rate deternation waere Implemented during this period. Te third phase of the liberalization episode, 1988.90, was chancterized by a "reform fatigue. Ihere were some major developments during this period, especilly in the area of foreign exchange and the capital account. The Central Bank established a Foreig Ehange and Banknotes Market in 1988. Exchange rates are determined in the market with the parUcipation of banks and other relevant fdnancal institutions.3 In 1989, foreign exchange operations and international capital movements wer liberazed entirely and the TL became convertible. In 1990, banks were left completely free to determine their exchange rates. However, these interventions were mostly realizations of previous decisions. In fact, during this period there was a visible decline in the appetites of tle, reformers in consolidating the achievements of the previous attempts, which manifsted itself in the postponement of the much needed regulatory changes such as in the capital market law, baning law, corporate law and (except for minimal amendments) banlkruptcy law. IL DEVELOPMNTS IN FINANCIUL MARKI'S Although the financial liberalization measures were announced as early as 1980, the impact on the structwure of the finandal system became effective after 1986, ie. after the establishment of the fudamental institutional framework necesary for the operation of fitancial markets. In this section, we review seral important features of developments in the financial system during the reform period. It Bans are usually invited every year to the Central Bank to review the findings of examiners and external auditors In case financial weaknesses are identified, the Central Bank sends follow-up letters to banks; whenever necessay, the Treasury is also informed with views on necessary measures to be taken. 21 Even though maturities available in the market include overnight, 14 weeks and 1.3 months, most transactions undertaken are overight. The market grew rapidly and the average volume of daily transacions increased from TL6.4 billion in 1986 and TL200 billion in 1987 to TL22 trilion in 1990. 1O/ Particpants in the market are commercial banks, speca finance houses and institutions authorized by the Treasury. The Central Bank both acts as a blind broker and trades in the market.  -6-^ ILl Flnancj3 Deepening Dwing the 1980&1990 period the Turkish financial system grew considerably. In order to describe the characteristics of the development of finandal markets, a single summary measure Is not sufficient. The traditional measures of finandal deepening such as MI (currency in circulation + sight deposits) or M2 (MI + time deposits) underestimate the growth of the financl system, since they do not take into account the deveopments in the securities markets and the emergence of foreign currency denominated deposits. On the other hand, the sum of all financial assets (curmacy in drculation + total deposits + total securities) exaggerate the development In the financial system since a bulk of public sector securities are held by banks. Tables A1 ana A.2 present data relevant for the two measures A closer inspection of Table A.2 reveals that the share of total securities (private and public) in total financial assets (i.e. sum of currency in circulation, total deposits and total securities) increased fist at a mild rate from 22% In 1982 to 25% In 1986, then climbed sharply to 43% in 1990. IL2 lI.Foreign exchanae deposits The second important feature of the liberalization policies was the role played by the foreign exchange and foreign exchange denominated financial assets. In 1984 Turkish citizens were allowed to hold foreign exchange deposits in banks, and these became an important financial savng Instrument after 1986. This is reflected in Table A.l as a considerable divergence between two monetary aggregates M2 and M2Y (M2 + Foreign Excbange Deposits) ly The goverrment's policies concerning foreign exchange denominated fnancial instruments were somewhat mixed. On the one hand, at various times, the policy makers supported such instruments by allowing them (as in the case of foreign exchange deposits) or by introducing them (as in the case of foreign exchange denominated public sector securities); on the other hand, authorities tried to prevent the widespread use of such instruments for fear in institutonalizing the indexation and therefore currency substitution. 11.3 The dominance of the public sector The course of development of financial markets in Turkey was shaped not only by the policies aimed at promoting these markets but even more by the financing needs of the public sector itselU One of the main targets of the reform program was to reduce the public sector deficits. It was also thought that in order to minimize the inflationary impact of the deficits as well as to introduce an enduring fiscal discipline to the public sector, the deficits should be financed from the financial markets at the competitive rates. fli The increased popularity of the foreign exchange deposits can be explained by currency substitution phenomenon. In Turkey, traditionally gold and, for the last decades at an increasing rate, foreign exchange are held as hedges against inflation. TMe introduction of foreign exchange deposits no doubt had a promotional effect on currency substitution, since it legalized foreign exchange holdings. However, its main consequence was in the manifestation of the phenomenon, rather than a radical change. For empirical evidence on high elasticity of substitution between domestic and foreign currency deposits, see Kumcu (1989) and Iskenderoglu (1989).  -7 - Economic polides followed during the 1980-1990 petiod were not successfud It curbing the public sector borrowing requiement (PSBR). Indeed, as can be seen from Table IL, the average PSBR/GNP ratio was higher (7.196o), In the second half of the 19809, from its aleady high level (6.4%) in the fkrst halt TabeILI PUBLIC SECTOR BORROWINO REQUIURM Year PSBR/GNP (%) 19. 10.5 198 4.9 1962 43 1983 ~~~~~~~~~~~~~6.0 1964 6.5 196S . 4.6 1986 4.7 198 7.8 1968 63 1989 7.1 1990 9.4 Souiue: Sae Paiang O on and th Undecey otau and Foreig ahde Tlhe persitence of relatively high publc sector borrowing requirements, when coupled with the stated objective of the authorities In reling more on domestic finandal markets for their flnancing, induced the introduction of a series of measures to enhance the marketabity of public sector securities. The fkst type of measures were aimed at forcing banks to allocate a portion of their portfolio to public sector securities. In order to achieve this aim, banks were obliged to hold public sector securities against their liquidity requirements.y Therefore, a demand for such securities, closely related to the expansion of depos- Its, was created. The second type of measures were implemented to enhance the attractiveness of these secaities. Their yields became market determined after the introduction of the auction system in 1985. Fnlly, the fact that returns from these securities were free from Income taxes (they were still subject to a 10% withholding tax) contributed to the dramatic shift in the demand. IN( In 1983, the previously complicated structure of reserve and liquidity requirements were simplified by unifying rates at 10 percent and 25 percent respectively, for all deposits. In 1985, the reporting frequency was chahged from one month to one week and the compliance lag was shortened from six to two weeks. In 1986, legal requirements were extended to foreign exchange deposits. The liquidity requirement ratio was divided into two components according to which banks were required to keep 5 percent of their liabilities as free reserves at the Central Bank and vault cash and 12 percent as government securities Tne ratios increased gradually over time, reaching a total of 35 percent (with 30 percent in the form of government securities). For more details, see Akkurt et. al. (1991) and Bayazitoglu, Ersel and Ozturk (1991).  -8- As a result of these measures, the public sector enjoyed the advantages offered by the devlopments In the securities markets. It can be seen from Tables A3 and AS that the publlc sector dominated the primaly and secadary markets. Although the share of the public sector started to decline In both markets towards the end of the decade, it is stUI above three quarters Finally, as can be seen from the fgures in Table A.4, the yields of public sector securities were Indeed competitive with domestic private sector paper, also, rates on savings deposits are higher than the contemporaneous rate of inflation until 1988. After 1988, yields on government securities lagged behind those on savings deposits and the rate of inflation. Starting from 1986, the Treasuty became the major supplier of securities. However, the persistence of large public sector deflcits and the thinness of financial markets situated the Treasury in a disadvantageous position despite its monopoly status and kept interest rates high. The introduction of the auction mechanism for government securides institutionalized this phenomenon. Even though, at least in 1987 and 198, It was common to blame buyers' collusion for high interest rates, in his analysis of auctions between 1987-89, Alkan (1990) has argued that banks may actually enjoy high interest rates without attempting to collude, once they are sophisticated enough to follow the trends in the auction market and aware of the desperate need of the Treasury to sell public securities. Persistent high public sector deficits exerted pressures on the already thin financial markets. lhe first consequence of this phenomenon was the crowding out of the private sector. In spite of efforts to promote capital markets and Induce the corporate sector to change its financing pattern by relying more on secuities markets, particularly through equitis, the results were far from satisfactory The second effect of the high public sector deficits and their financing pattern was on the banldng system. Banks considered placng their funds on public sector securities a safe investment. Since the borrowing needs of the public sector seem to be countercyclical, the permatnent existence of the Treasury in these markets also cushioned the banks against fluctuations in earnings from their loan portfolio. IIA Interest rates When the Government intervened following the crisis of 1982, the Central Bank was authorized to fix interest rates on deposits, which were lowered from 50 to 45 percent in January 1983. Regulation of interest rates continued until 1988 and adjustments were made according to movements in the rate of inflation so as to keep them positive in real terms fLi According to survey results reported In Alkan (1990), in 1989 about one-half of banks' purchases of Government paper was to meet legal liquidity requirements. Most of the rest were resold either directly or through repurchase agreements. The share of purchases as an asset alternative to credit was only about 5%. Results of another survey, reported in Atiyas, Ersel and Ozturk (1992) and summarized in more detail below, indicate that in general bank managers perceived the non-obligatoty portion of government securities as instruments of short-term liquidity management Only under exceptional circumstances were these instruments perceived as substitutes to credit. Finally, Ersel (1992) also provides evidence that the non-bligatory holding of government securities by banks and the non-bank public was more than 50%.  -9 - lfation started to pick up towards the end of 198XO In the fall of 1988, infation fgures wer higher than the expectations of the Government and a run out of the Turkish Lra into foreign exchange ensued; parallel market exchange rates shot up. While intervening into both the parallel market and the foreign exchange market by seling foreign exchange, the Central Bank also announced the liberalation of deposit interest rates. Banks responded frantically and announced one year deposit rates varying between 70 and 88 peroent, up from a utdform rate of 65 percent. Compared to the seeminy similar episode In 1982, the banks' response was possibly driven by the uncertain and speculative economic environment rather than widpread financil distress Nevertheless, banks offering lower rates were besieged by depositots trying to switch their fbnds to those offering higher rates. There was also a severe competition in the interbank market. The Central Bank re-intervened by advising a temporary informal but uniform cefling of 85 percent on one year deposit interest rates. Monetary authorities also required banks to report desired changes in interest rates in advance Both the foreign exchange market and the market for the Turkish Lira were stabilized as a result of the intervention, despite several inconsistencies and reversals in policies. However, except for a few exceptional periods in December 1988 and the first few months of 1989, real Interest rates on deposits remained negative (Table A.4).iJ Even if temporary, high deposit Interest rates tesulted in a large inflow of deposits into the banking system. In the folowing period, especialy in 1989, the banking system was characterized by an excess supply of funds and faced substantial difficulties in placing them. Again, compared to the 1980-82 period, there was no comparable increase in demand for credit; as discussed in the next section, the increase in interest rates did not generate a vicious cycle of distress borrowing in the corporate sector. The banking system suffered a temporary reduction in profitability in 1989 (see Table III1). asl5 Developments in the baniing svstem The Turkish bankng system's response to the implementation of the financial liberaization program was generally considered as quite accommodating. Banks, which immediately adapted themselves to the new conditions, started to modernize themselves by switching from manual methods to computerized systems and by recruiting qualified personnel to extend their activities beyond their traditional markets. However, a doser examination of the developments in the banking sector reveals that its growth was nevertheless neither smooth nor completely in line with the expectations of the reform program. Under the liberalization program the Turkish banking system demonstrated a rather peculiar development pattern. The characteristics of this development pattern may be summarized as follows: (i) Until 1987, the Turkish banking industry expanded. On the one hand the new banks entered into the market, on the other, existing banks followed an expansionary policy to secure their market shares. As can be seen from the Table 11.2 bank assets/GNP ratio climbed from 28 percent in 1980 to 54 percent in 1987. However, starting from 1987 this trend was reversed. After reaching its peak level in that year, the rate IN Data on bank loan rates are not available. However, there is ample anecdotal evidence that banks apply a wide range of lending rates, depending on the perceived quality of borrowers. See below.  - 10- of grwth of the banking system slowed down and the bank assets/GNP ratio declined steadily to 43 percent In 1990. TablW RATIO OF BANK ASSES TO GNP Year (%) 1980 28S l 98 34.4 1983 41.5 1984 413 1985 409 1986 4&6 1987 54.0 1988 X-S51.0 1989 463 1990 43.2 Source: SouroeTotal ass figures am obtained fm Babjn varkou haue, .te Bank's Assation of Turkey. he ON? figures armf the Stae Insttute of StatlatL The enactment of the new bankng law in 1985, whbch put the baking system under more serious prudential regulation and supervision, the incease in uncertainty due to the resurgence of inflation, and loss of credibility in the economic policies followed, were the major factors behind the reversal in the growth trend. (ii) A second dimension of structural change in the banking sector manifested itself as a decline in concentration, measured by the share of three largest banks in total assets (Table IL3).' jL/ Data collected by the Banking Department of the Central Bank of the Republic of Turkey. The figures in Table 11L3 are computed only for national banks. As demonstrated below, the share of foreign banks in total assets remained negligible throughout the period.  - 11 - TsbIe liJ THREE BANK CONCENTRATION RATIO (Mesuedas % of toual assets accounted by thre lqauDs banks in each categoly) ___~ Co_mea Bank Year Prafte Dauks AD Nationa Bank 1980 70 51 1981 67 48 182 69 49 1983 72 56 1984 74 57 1985 73 53 1986 68 SO 1987 65 48 988 62 45 1989 S4 46 1990 52 44 Sour= he 1980 pm are reproduced fom Pdeba (1991)Table &11. The data are from Bnk In Turbey, vaious Issues, The Banks Assodation of Turkey. The 1990 figures are own calculatins based on the 1990 issue of the me publication.  ^12- These fis clearly show a decline in concentration in Turkdsh baningP Note that concetraidon ratios for private ban decline faster than then for the commercial banking system as a whole. This can be attributed to two factors The nst is that the growth rates of Turkish priate commercial bans varied considerably, and dynamic medium or small size banks grew faster thln large banks Ihe former found opportunity to expand relatively rapidly, whereas fot the latter type of bank, restructuring implied slower growth. The average rate of growth of total assets of various categories of commerial bans giv in Table 11.4 for the 1986-1990 period supports this view. Table IL4: AV1RAGE RATE OF GROwni OF COMMERCIAL BANK ASSETS 1986-1990 (%) No. of Brache State-Owned PriVate Feign r ToWal 1-30 I 633 51.1 59.2 31-100 50.6 7. 26.7 522 101+ 47.0 424 44.8 Souroe Data colected by the Banking Depa 8m of fte Cental Bak of the Republc of Turkey. The second factor is the continuation of the dominance of the large state held banks in the banidng system, which explains the relatively slow decLine in the three bank concentration ratio for all W~I As can be seen from the following table, a similar picture emerges if one computes the concentration ratio on the basis of total deposits (i.e., the sum of TL and FX denominated sight and time deposits). Note, however, that the ratios decrease at a slower rate in this case. TREEI AND FOUR BANK CONCENTRATON RATIOS (Meaued as % of total deposita accounted by the three and four brgest bafnk in each cay) Prvte aAll natinal banks b'ree Bank Four Bank Three Bank Four Bank 1980 71.7 78.1 53.2 64.3 1981 69.4 77.2 50.6 63.1 1982 72.5 79.6 55.3 65.2 1983 75.8 825 56.5 65.9 1984 74.6 80.2 57.3 66.0 1985 74.8 80.8 54.5 64.8 1986 71.0 77.5 53.0 62.8 1987 58S 62.5 49.6 59.3 1988 665 755 47.4 56.9 1989 59.8 69.5 453 53.2 1990 57.0 66.6 43.3 513 Source-Data colleded by the Banking Department of the Central Bank of the Republic of Turkey.  -13 - commetcial banks In fact, the share of total assets held by state owned commercial banks was almost constant in the 1986-90 period (Table IL5). Table 115f SHARE OF STATE-OWNED COMMERCIAL BANKS IN TOTAL ASS (%) 1986 47.0 1987 47.0 1988 462 1989 419 1990 482 Source: Data colleted ty the Banking Dqaliet of the Central Bank of the Republic of Turkey. (qil) Another major change was the increased share of securities in banks' portfolio. The share of securities In the total assets sharply increased in the second half of the 1980-1990 period gable 11.6). Table 1: SHARE OF SECIUTIES iN TOTAL ASSMS (%) 1980 17 1983 3A 1987 1in 1990 11.2 Source: Central Bank of Turkey. This sharp Inaease In the securities portfolio can wholly be attributed to the Increase in holdings of government securities. Banks' obligations to hold government securities against high liquidity requirements created a rather large and expanding market for these papers. In addition, the high yields of public sector securites, especially after taking the tax advantages into account, madc them attractive for banks and to a lesser degree for corporations to hold them in their portfolios. This change in the asset structure of the banking system was also reflected in its earnings structure Table 11.7 shows that the share of credit related income (interest and commissions) in total earnings declined over the 1986-90 period whereas that of income from securities market operations increased. While this trend is apparent in banks of all size categories, medium-sized banks seem to be less involved In securities market operations.  - 14 - (iv) During the 1980.1990 period, banks also increased their capitalization. The ratio of capital (defined as the sum of paid up capital reserves, revaluation fund and profits) to total liabilities considerably increased from an average of 7.6% In 1980.1984 to 9.0% in 1985-1990 (Table 11.8). TlW1.7: SHARE OF INCOME FROM CREDIT AND SECURmTEs MARKE ORIENTATIONS IN BANK EARNINGS () Number of Branches 1-30 31-100 101+ Total 1986 48 56 62 61 _________________ hi 13 5 16 15 1987 41 39 60 57 _________________ hi 17 9 1S 14 1988 a' 37 36 54 52 _________________ hi 16 12 18 17 1989 a' 50 40 50 49 _______ ____ __ h 20 17 34 23 1990 4 43 55 54 hi w 21 12 22 21 * TToal earnings indudes net rather than gross income from foreign acchange opentions. a/{ Share of income from credit opertn hi Share of income from securties market operations& Table II: CAPrrAL STRUCIURE IN THE COMMEiRCIAL BANKING SYSlTEM Year Share of Equity Capital in Total ilabiitie (f) _ _1980 2 5 0 1981 63 1982 6.9 1983 ~ 7.7 1984 9.1 1985 8.0 1986 8.9 1987 8.2 1988 9.1 1989 9.4 1990 10.1 Source- Banks in Turkey. Various Issues, lhe Banks' Association of Turkey.  - 15 - (v) The opening up of the Turldsh economy also affected the structure of the Turkish banldng system. As can be seen from Table 11.9, the share of claims on and llabilities to non-residents of the commercial banks increased sharply after 1986 Table 1L* COMMCLAL BANK' EXPOSURE TO NON-RESI- DENTS (% of balance sheet total) Year Claim on LJablitkes to non-residents non-residents 1981 0.1 0.1 1982 0A 0.3 1983 0.9 0.8 1984 1.8 2.0 198S 0.2 0.2 1986 6.4 3.8 1987 6.6 3.9 1988 9.8 4.9 1989 &.0 4CS Source: skcnoglu. Ourk and Teael (1991). (vi) Duing the 1980-1990 period, total domestic credits as a percentage of GNP declined (Table ILIO). The relatively high level of credits in the beginning of the decade was mainly the result of the existence direct medium and long term credits of the Central Bank. The Central Bank was acting in a dual capacity, both as an offical semi-development bank and as a central bank. In the 19BOs, in line with the policy of curbing directed credits, the Central Bank's credit lines were firslt reduced and then phased out, thereby ending the peculiar dual Identity of the Bank, and creating an opportunity to transform itself into an institution solely responsible from central banking.1y However, the rest of the banking system was not able to fill the gap created by the elimination of directed credit. After 1987, decline in the ratios of commercial bank credits also contributed to the reduction of the credit-GNP ratio. JLI In 1989, the Central Bank's practice of extending medium and long term credits in the form of advances against bonds was terminated. It was decided that rediscount credits would be extended only against short term company paper and their use as tools of selective credit policy came to an end.  -16, Table MI CREDrS AS PERCENrAGE OF T1H GNP Year Centad Bank Comral Bank Other Total credits cardits re,dits craeits 1960 7 1S 3 25 1981 6 16 3 26 0 198y2 _ _ ___ _ 17 3 25 1983 4 17 3 24 1984 2 14 2 19 1985 3 i5 2 20 1986 2 19 2 24 1987 3 21 2 26 1988 2 17 2 21 1989 2 16 2 19 1990 17 2 20 Soure Qartrly Bulletin. Various sues, Centul Bank of the Repubc of Turkey. vii) The profitability of the Turkish banking system increased considerably In the second half of the 1980. Data in Table IIll, tabulated for the commercial bank only, reveal a significant increase in bank profitability, despite accompanied inaease in operating costs T8bklell.- BANK PROFITABIiT (as % of total asses) Year GEM OC NEM PBT 1983 S.0 3.6 13 1.0 1984 5.7 3.3 2.4 1.9 1985 3.8 2.9 0.9 0.6 1986 . S.4 2.8 2.6 1.9 1987 73 2.8 4.5 3.0 1988 85 33 5.2 3.5 1989 7.4 3.8 3.6 2.4 1990 12.3 5.1 7.2 3,6 GEM: Gross Economic Margin - Intereat Recehied - Interest Paid + Other Income (Nct) NEM: Net Economic Martin: GEM - OC PBT: Profits Before Taes - NEM . Other Expenses (Net). 0C Operating Costs. Total AMet Arithmetic averages of the year.end values. Souroe: lskendepiglu Ozturk and Temd (1991 calculated from data compiled by the Banking Department of the Central Banl  - 17- The increase in operating costs in the 19884990 period is noteworthy. A closer examination of income statements reveals that personnel expenses remained relatively stable during the period, ewept for an Increase In 1990 due to lbor agrements sined by big priate baks and a ww wage setg system Introduced by public banks whih also led to sharp pay raises. By contrast, there was a sharp increase In other administrative expenses As dicussed In Akkurt, et al (1991), these trends may be attributed to banks' tendency to prefer investing in computerization over employing more staff This is especially true for larger banks which have established ATMs and on-line connecdons among branches. Nevertheless, the incrase in bank profltability was not uniform among various categories of banks Small domestic and foreign commercal banks are in general more profitable than larger banks and state-owned banks in Turkey (Table I1.12). The profitability of the foreign banks declined sharply after 1988, and, in contrast to national banks, did not recover In 1990. The worst performing c*tegoty of banks was medium sized state owned banks, which declared losses in 1987 and 1989. Table PROFrT BEFORE TAXESIAVEBAOE TTAL ASSETS FOR COMMER- CIAL BANKS IN TURBYEY 1987 1988 1989 1990 Prkate domesic banks 1-30 Bmachas 4.1 7.1 5.1 53 31-100 Brandis 2.7 35 2.8 4.1 101 + Bmwhes 3.3 3.7 29 3.9 1-30 Banches - 31-100 Branches -2.1 3.9 -4.1 02 101 + Branaches 2.5 4 1.6 23 Freip banks 1.30 Branches 5.8 81 S.0 43 30-100 Branches 4A 8.6 4.7 4.4(e) 101 + Branches . (L) One bank only . Sowce: Data oolected by the Banking Department of the Central Bank of the Republic of Turky.  - 18 - The nee in bank profitability is espedally noteworthy in view of the decline In the degree of concentration In the banking systemlW 'he differenoes in the profitability and patter of growth of the various categories of banks indicate that small domestic commercial bank weoe the most dynamic elements in the bankAng system. Vii) CODntary to the expectations of poliy makers, the tole of foreigp banks in enhancing competition In the Turkish banking system remained negligible. As shown in Table IL13 foreign bank are relatively small and their share in the total assets of the banking system is not sigificant. A closer Inspection of their activities also reveal that foreign banks never had the intention of becoming major players In the Turkish banking system. They concentrated their activities on foreign trade related areas where they had, espedally in the first half of the period considered, absolute advantage due to their knowhow in these areas and their international connections and, in general, did not attempt to compete with domestic banks in traditional banking activities. Pehlhvan and Kirkpatrick (1991) argue that this resulted from restrictions placed on foreign banks. Nevertheless, some foreign banks did play an Important role in trainng a new generation of middle level bank managers, who were subsequently employed in domestic banks Table MIl. OWNESRSHIP COWMMOiON OF THE BANKINO SYSTMM 1980 196 199 Prbatadometk bauab *-- (a) I 26 26 29 45.7 467 443 State-owned banks (a) 13 12 11 (b) 51.5 50.1 52.0 Foreign banks (a) 4 18 26 (b) 2 8 ?.7 3S (a) Number of banks. (b) Share in total asseu of banking system. Source: Banks in Turkey, Various Issues, The Banks Association of Turey. 1/ Denizer (1991) finds eonometric evidence that bank profits are positively related to concentration. Hence it seems that profitability increased despite a reduction in the degree of monopoly power in the banking system.  -19- 1.6 Jkbehavior iJ Turkev: Evidence from survW results 'Me liberalization measures had a dual effect on bank. On the one hand, by broadening the spectrum of fnancial markets and opening them to banks, the liberalization progm enabled banks to diversify their actvities On the other hand, the uncetaides Inherent in the liberalization programs aggra- vated by the inconsistencies In he implementation of economic policies induced banks to foilow rather conserative strategies In this section we attempt to gain some additional insights into bank behavior in Turkey and report results of a questionnaire distributed to the bank managers and the accompanying intervews made in the Summer of 1991. A complete set of results are reported in Atlyas, Ersel and Ozturk (1992). The main Idea behind the survey was to understand relationship between banks and their clientele in loan markets and the mode of competition. The questions were inspired by the recent theoretical literature on Imperfections in financial markets that result from problems of information and costly contract enforcement. In particular, based on the theoretical exposition in Calomiris and Hubbard (1990)12 and prors about the structure of credit markets in Turkey, the relationship between banks and their clientele, and the nature of competition, the questionnaire hypothesized a pool of potential borrowers segmented into a group for which banks have more complete information (*blue chip companies") and others for which banks have little or no information. The survey found that banks' customer patterns displayed significant variation. The majority of small and medium bank concentrated their activities on predominantly "blue chip' companies in large cities. Banb with an extensive network of branches, on the other hand, had a much wider customer base In general, an overwhelming majority of bank managers confirmed the suggestion that whereas there is a fierce oompetition among banks for blue chip companies, in the non-blue chip sector competition is among firms requesting those services. Moreover, banks were found to be reluctant to expand their customer base. It is well known that bank loan rates exduhbit wide variation in Turkey. It is also generally believed that, contrary to practices in economies with developed financial markets, many banks in Turkey prefer to announce a Umaximume lending rate rather than a prime rate. Survey results suggest that all large and multi-branch banks prefer announcing a maximum rate and then negotiate it downwards during bargaining for a loan package. T'e basic reason for this preference is that with borrowers that are perceived to be risky, negotiating lending rates downwards is easier than asking for premiums over a prime rate. Small banks, on the other hand, are more likely to work on the basis of a prime rate,V but are very reluctant to announce 12/ Calomiris and Hubbard (1990) develop a model where the credit market is divided into a 'Walrasian' segment, where borrowers' relevant characteristics are costlessly identified, and an "Information intenslve segment, where borrowers' characteristics are private information unavailable to lenders. ZQ/ This is not surprising since most small banks were found to concentrate their business on blue chip companies.  -20 - It to the general public& A pomsibl explanation fr banks' aversion to anucing a prime rate is that It trflecls their reluctac to expand thdr customer base. Only a small number of blue chip companies, and companies that belong to the same conglmoerate groups as the bank are able to obtain credit without colateral In general, roughly 80.90 pement of all bank lending is made agaist colateraL Bank were asked about their pdmary concan in lending to a fim that has not yet established a reputation. 'ey wer asked whether their pdmary concerm would be not being able to dish firms' rbk and return chaactrstic or not being able to monitor borows once loans are made Bank' respose depended very much on tbeir organizational struc Small bak dealing ith blue chip companies Indicated that neither problem was a major concetn; that response reflected to a large extent that they would not attempt to include new borrowers into their clientele. Large banks' responses differed according to their Institutional capabilities. Tbose that had established a wel hactioning information gathering network indicated monitoring as a major problem Some banks indicated that they had units that undertake the monitoring function and therefore are more concerned with ea-ante Information gathering. Another question inquired about what would make banks more wilig to lend to borrowers currently not in their customer base: 'Te options were reduction In maceconomic instability and uncertaintis, establishment of an efficient legal system, establishment of a well funcdoning intormation gathering system or credit subsidies. Only one bank indicated that subsidies would help. Reduction of macroeconomic uncertainties and establshment of an information gathering system to coUect necessary data to make loan deciions were the options ch by the majority of banks.W Rearding establishment of an efficient legal system, it tuned out that bank had almost complete mistrust of the egal system ad based their lending strategies so as to keep recourse to the legal system at a minimum Improvements were seen as either irrelevant or improbable. Almost aU bank managers favored making packge deals with their customers. This can be interpreted as an indication for banks' desire to operate as muld-product firms, i.e., diverging their activities along various financial services. DlJ An alternative explanation could be that announcing a prime rate would increase the degree of competition in the bankdng system. ~ZI The emphasis on macroeconomic uncertainty as a determinant of bank's exposure to risky bonrowers is consistent with predictions in the literature. See Caprio (1991, 1992) for the importance of the degree of uncertainty in bank's choice between risky and riskiess assets. 23 The legal system in Turkey is notoriously slow and overburdened. Courts in Istanbul are sometimes loaded with 70-80 cases per day. There is also anecdotal evidence of an increase In economic crimes," especially bounced checks (see, for example, the weekly Nokta, August 4, 1991). Perhaps unsurprisingly, in the 1980s a "check mafiae emerged in large cities, which specialized in debt collection through intimidation and use of force. The mafia functioned mainly in settling debts among smaller businesses.  -21 - All bank managers distinguIshed credits fom otfier bank sevices and stressed that their securities, monqe and foregdp exchange market activities cannot be conidered as substitutes for credit. Although the manags seem to draw such a compmetarity betw credit services and bans' operations for liquidity and foreig exchange magement, most of them also added the qualfcation that *under exceptional condltions, such as those that prevailed in 1989 when high cost of loanable funds created dfficulties In phcing loans, thewe acities may certaWy become substitutes for loan. Therefore, It can be conluded that thers is an asymmetric relationship between these acities, which is conditional upon the state of the credit markeL mL DEVELOPMEN'I IN CORPORATE FINANCE fls section focusses on patterns of financing in the corporate sector and reviews trends during the finandal reform period. The main source of data Is the balance sheets and income statements of a sample of 81 firms registered at the Capital Markets Board. The data set Is not representative of the corporate sector in Turkey. It consists of large, financially relatively healthy firms which possibly have better access to extemal financing than an average Turkish firm.l IIL1 Trends in the composition of liabilities Some indicators for the composition of corporate debt is provided in Table U1I. In order to assess whether access to exnal finace vies across goups of fims differentiated by size, the sample has been split into two groups of equal number of firms on the basis of the value of total assets in constant prices. Specifically, frms where the value of total assets (in constant 1985 prices) were below 78 billion Turkish Lras in 1984 were classified as smalL TableILi reprs the mean value for each variable for the group of small (5) and large (L) firms, along wdth Indicators of whether the differce in the means is statistically signifficant or not. The main components of debt (excluding liabilities in the fonn of taxes and insurance payments) are short and long term bank lans, direct finaning (fnance bills), and other liabilites. Other short term liabilies includes borowings from owners, but mainly cOnSiSts of trade credits. The following patterns emerge: As a percentage of total assets, large firms are generaly more indebted than small firms, but the difference is often insignificant and is completely eliminated by the end of the sample period. This is mainly due to a steady reduction in the degree of indebtedness of large firms. In general, the composition of debt differs across the two sue categories; however, the differences decrease over time. A S( For more details on the data set, see Ersl and Ozturk (1990b). The analysis that follows is subject to two qualifications. First, financial liabilities are expressed in terms of stocks rather than flows. Data on sources and uses of funds, which would help provide a more complete picture of financing patterns, are unforunately unavailable. Second, the anaysis of patterns of external financing does not take into acoount etenal equity, due to lack of firm level data on issues of new shares to the market for the sample period. However, only a smal portion of increases in equity (on average 14 to 15 percent in 1990-91 according to CMB data). The CMB currently requires firms to realize at least 15 percent of capital increases through market issues.  -22- larger proportion of total debt consists of short term debt in the case of small fkms, hower In 1989 the difference is no longer statisticall significant Regarding the breakdown of sources of debt, large firms use more bank loas whereas small firms rely more heavily on other liabilities. However, by 1989, the share of bank loans in the two sub-samples Is no longer signficantly different. A doser inspection reveals that this is primarily due to an impressive reduction in the share of short term bank loans In the liabilities of large firms. By contrast, large firms continue to be dominant in the use of long term bank loans thrughout the period. However, the share of long-term bank loans decrease over time for both large and small firms, suggesting that financal eform faied to increase the availability of long term loanable funds. The share of other short term liabilities is significantly higher in small firms' liabilities, most likely reflecting the importance of trade credit. Ffially, the share of finance biUs is very small, and close to zero in the case of smaull fims. Hence, direct borowing plays an insignifcant role in corporate finance. Nevertheless, their importance seem to show an increasing trend in 1987-89 in the case of large firms. With the exception of long term bank loans and trade credit, the hability structure of small and large firms exhibit a pattern of covergence during the 1980s. There may be two reasons for this increasing similarity. The first is structural changes in financial markets that generate more uniform treatment of large and smal firms An alternative reason may be that firms in the data set grow over time; if after a certain threshold the impact of size on liability structure diminishes, then firm growth would explain the increased similarity. In order to see whether this was the case, the data set was split on the basis of observations rather than firms and a lower cutoff point for size was established. Accordingly, observations where the value (in constant 1985 prices) of total assets were below 50 billion Turkish liras were classified as small In this classification, the number of observations in the S category change between one third (in the earlier period) and one fourth (later period) of the data set The results are displayed in Table =11 The main difference between tables IILI and IIL2 lies in the fact that in the latter the share of short term debt remains significantly higher In observations that correspond to small firms, and mainly due to a higher share of short term other liabilities and lower share of long term bank loans. This underscores the importance of trade credit for small firms and suggests a pattern of development where firms switch away from trade credit as they grow and find long term bank loans more accessible. The dedine in the use of short term bank loans by large firms in the 1988-89 period deserves further comment. As discussed in Ersel and Ozturk (1990a, b), especially large and healthy firms started reducing their use of short term bank loans in 1989. The emergence of this trend can be traced back to 1988, when the cost of bank credit increased sharply, following the complete deregulation of deposit interest rates. At the same time, the increased risk of default made banks more cautious in selecting their customers. The coupling of self-restraint by prime borrowers and cautious behavior of banks resulted in a surplus of funds. This reduction in intermediation in response to the interest rate shock contrasts sharply with the experience of the early 198(s. Then, the interest rate shock was accompanied with a rapid deterioration of corporate earnings. Firms, finding themselves Illiquid, resorted to distress borrowing, further deteriorating their balance sheets positions. The crucial difference between the two periods is that in 1989 fims could rely on relatively high earnings to reduce their exposure to short term loans.  -23- Ib ILI: COMPOSITMON OF CORPORATE DEBT I Sime 1982 16 1984 1986 1986 1987 1988 1989 DebAau Ratio S 0.63 0.32 0.4600 04S" 0.48 0.48 -0.0 049. L 0.38 056 0.S3 0.33 0.4 0.52 0.| 0 0.49 Shot Tem Det S 0.76" 0.74"0 07630 0.78' 083" Oil" 0.78 .72 L 0.6S 0.64 0.66 0.69 0.68 0.70 0.9 0.6S Long Term Debt S 0.24* 0.26* 0.24" 0.22 0.17" 0.190 0322 0.28 L 035 036 0.34 031 032 030 031 0.3S Bank L S 0.40"0 0.40" 037 039"6 ' 0.37" 03** 0.37* 0 L 04 054 0.56 056 058 0.45 0.49 0.40 Short Term Bank Lan 5 0.24 0.21* 0.19" 0.22 0.23 0.19 0.22 02Z L 028 0.27 0.29 0.28 033 0.24 0.25 020 Long Term Bank Lom- S 0.17* 0.19' 0.190 0.17* 0.12"* 0.140 0.15* 0.14 L 0.26 0.7 0.27 028 0.27 0.22 024 0.20 Fian Bils S 000 0.00 0.00 0.00 0.00 0.00 0.00' 0.00" L 000 0OO 0.00 0.00 0.00 0.01 0.01 0.03 Other Uabilitia S 0.57* 057"* 0.60" 059"0 0.60"* 0.63" 0.61" 0.0 L 0.40 0.42 0.41 0.42 037 0.47 0.47 0.5 Short Tem Other Liablities S 02" 054 0.7"* O.6"* O8"* 0.61" 0.6"0 0.50' L 036 0.37 0.37 0.40 036 0.45 0.43 0.41 Long Term Other Liabilities S 004 0.04 oa03 0. 0.03 0.01 OOS 0.10 L 0.04 0.05 0.04 0.01 0.02 0.02 0.04 0.14 jg Shar in total debt. Tat and Insurance premium lIbilities are acluded. Stats Indicate signifcance lvels at which the null hypothesis of equality of means is rmected. A single star denotes 10% and a double star denotes 5% lerel of signiflcance, respectively. S: Firms with total assets smaller than 78 biffion TL in 1984. L Farms with total as larger than 78 bMion 11 in 1984.  -24- Tableii COMPOSMTON OF CORPORATE DEBT g/ - - - - - -- - 1 Sbe 16 1983 1964 195 198 "8 98 i1989m DebWAaedt Rtio S 6S 0.S2 0.44" 0.420 0.53 050 0.50 0.4 L 0.57 0.56 0S 0.54 0.50 0.50 0.SO 0.47 Short Tesm Deb S 076 0.76" 0.77" 0.79' 0.90" 0.01 0.79 0.82" L 0.67 6S 067 0.70 0.71 0.73 071 01 6 Lon Term Debt s 024' 0.24" 023 " 021' 0.10"' 0.19 021 0.18*- L 033 03S 033 030 0.29 02 0.29 036 Bak Lans S 0.41' 039"* 0.3700 036"0 037"0 OAO 03S 034 L 0.1 052 0.53 0.3 1 039 0.46 039 Short Term Bank LOS S 0.22 021 0.19' 0.21 0.30 023 0.21 0.26 L 0.29 06 0.27 027 029 0.21 0.24 0.19 Long Term Bank L S Q9 0.18' Q8 04" 0.07"* 0.17 0.14 007"* L 0.23 0.26 026 026 022 0.18 0.22 020 Finance Bils S 0.00 0.00 0.00 0C00 0.00 0o. 0.00 0.00' .___________________ L 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.02 Other bitlies S 0OS. ' 0.58 0.60"W 0.60" OA2 0.58 0.62' 0.64 L 0.44 0.44 OA4 OAS 0.4 0.4 0.50 0.56 Short Tenn Otier Usbes S 0.54 0.5** 0.58"O 0.8 0.0 0.58 0.- 0.55" L 038 039 0.40 0.43 0.42 0.2 Q46 042 Long Tm Other sLabiiles S O.- 0.03 0.01 003 0.02 Q0 0.04 0.8 ____ ____ ____ ____ ____ ____ ____ ____0 2 "* L 0.06 .O05 0.04 0.02 0.02 .02 0.05 0.13 A Shares in total debt. Tax and Insurance premium Uabilies are cluded. Stan indiate sgficance levels at which the nu11 hypothes of equality of means is rejected. Single str denotes 10% and double star denotes 5% vael of sgnificace, respectivel. S: Observatos with total assets smaler than 50 billion TLh L: Observatios with total asset larger than 50 btlion 1T 11L2 Variations in corporate indebtedness and the impat of finandal reform We now focus in more detail on variations in corporate indebtedness with a view to examine the impact of financial reform on firm financial behavior. Follwing the literature on the determinants of capital structure, we attempt first, to explain cross-section variations in corporate indebtedness by flrm specific variables. Second, to gauge the impact of the process of financial reform, we test whether the importance of  -25- these varbles in explang aoss section varIations in indebtedness have changed over time. In partdcular, we would ike to eamine whether there have been anly chaWes In () the importance of size in accessing external finance, (I) firms' relance on Internally generated funds, (iII) the role of collateral and (lv) the role of firm Investment opportunities that are difflcult to monitor by creditors. Because we expect that the Impact of these variables should vary across different types of financial instruments, several indicators of corporate indebtedness are used There are three qpes of short term debt. Short term bank loans (aluding long term bank loam maturing in the currt year), other short term debt (most of which consists of trade credit) and total short term funded debt (that Is, bank loans, other short term Liabilities and finance bils, exluding debt in the form of taxes and socal security premiums). The measures of long term debt are long term bank loans (induding those that are due In the current period ) and total long term debt (cootsisting of bank loans and bonds, excluding other long term Liabiities). The variables used in the analysis are the following. SIZE measured as the logarithm of total assets in constant prices, reflects existence of transaction costs in the proviion of extemal finance If fixed trasactions costs are important, then size should be positiely related to indebtedness. Profitability, measured as operatmg income per unit assets (OlTA), captures the importance of internal finance. It has been argued that bemause of imperfections in fnancial markets, external finance is more expensive than internaly generated funds; in other words, firms pay a premium for external finance. If that is the case, then, everyng else equal, profitability should have a negative impact on indebtedness since increased profitability implies higher availability of retainable internally generated funds. When firms produce unique products, their bankruptcy and eventual liquidation imposes costs on suppliers, customers and workers Since higher debt ratios imply higher probability of bankruptcy, these costs are relevant to capital structure decisions. Hence uniqueness is apected to be negatively related to indebtedness. Polowing Titman and Wessels (1988), we measure uniqueness by the ratio of selling expenses to sales (SES). Equity oontrolled firms have a tendency to invest suboptimally to expropriate wealth from the firm' creditors (Myers, 1977). Rational creditors are aware of this possibility and request an 'agency premiumt that increases the cost of debt. Hene, firLt with high growth opportunities are likely to face higher cost of debt This is especially true with long term debt. In fact, it has been suggested that short term debt may be relatively free of this agency problem, since creditors have the option of terminating their commitment whenever the investment behavior of fims jeopardize the value of debt claims. Growth opportunities are measured by the percentage change In total assets in constant prices (OTA). Assets that can be used as colUateral reduce the cost of debt and therefore are expected to be positively related to indebtedness. Collateralizable assets are measured by the ratio of inventories and net fixed assets to total assets for (INPTA) long term debt, and by the ratio of accounts receivable to total assets (ARTA) in the case of short term debt. The degree of indebtedness is also expected to be related to the existence of non-debt tax shields, although the dirction of the relation is in principle ambiguous (Dammon and Senbet, 1988). Non debt tax shields are measured by the ratio of the change in accumulated depreciation to total assets (DTA).0 Finally, volatility of firm's earnings, measured by the standard deviation of the ratio of operating income to total assets (SOITA), is expected to be negatively related to indebtedness 2.5 Annual flows of depreciation expenses were not available from the income statements; hence the first difference of accumulated depreciation was used to approximate the flow.  -26- As in most of the literatue, the variables are avaged over time. Howeve, In order to test for the stability of the coefficients, the sample peiod was divided Into two: 1984-86 and 198789. A dummy variable (C) which Is equal to 1 for the 197-89 period and mero otherwise was used both by itself and interactively with the explanatoty variables. Te results are reported in Table IL3. Short term bank loans are positvely affected by collateral assets and negatively affeted by profltability and growth opportunities The coefflient on size Is positie but not significant. The effect of growth opportunities diminish in the second period. Other liabilities, on the other hand, is not affected by size, growth opportunities or coLateral assets but negatvely related only to profiltabiliq. Moreover, no significant stmctural change is detected in the second period. These results uncover an interesting contrast between short term bank loans and trade credit, namely that at least in the 198486 period, trade credit was less subject to agency problems than bank loas This may be because trade creditors have better Iormation than banks about firms that they do business with and they are better able to monitor borrowers' behavior in the course of business. However, firms still prefer to use internally generated funds to trade credit, as reflected in the significant coeffcient on the profitability variable. Te coefficients in the equation for total short term funded debt display a pattern similar to those in the equation for short term bank loans except that non-debt tax shields also appear significant. Contrary to components of short term debt, size does play a significant role in explaining variations in long term bank loans. Moreover, the importance of size is not diminished in the second sample period. As expected, assets that can be used as collateral are also positively related to long term debt. Surprisingly, the agency cost variable has a significant but positive coefficient; fast growing firms carry a higher share of long term bank loansP Fmally, profitability is negatively related to long term bank loans. Another interesting result Is the negative sig on TOINPA collateral assets become insignifcant in the second period. One of the survey results reported above Indicated that blue-chip companies were able to borrow without collateraL Given that the data set consists mainly of large firms, the significance of both INPTA and TINPTA suggests that the ability to borrow without collateral developed in the late 1980X aI What may be captured here is the special institutional features of long term bank loans in Turkey. Most of these funds are loans from development banks or, at least until late 1980s, preferential credits from the Central Bank. As reported in Ersel and Ozturk (1990a), access to such loans requires a time consuming process of project evaluation that is costly for borrowers. The positive relation between long term bank loans and GTA may therefore reflect a self selection process whereby only firms with large investment projects apply for these loans. Hence, OTA may be a bad proxy for agency problems in the case of long term loans. Alternatively, in the case of loans from development banks, it may also reflect development banks' ability to monitor borrowers after a loan is made.  -27- Tatle3: DETBRMINTn OF CORPORATZ INDEITDNESS Short-term Othe abort- Tota ahort- Long-rmn Toa log- bank tan tem debt tm de bank lan term debt eep 4000 OA46 0384 4323 4233 (400N) (2431) (208) -2M (-1.453) SIZE 0.012 4013 0006 0.027 ____ _ (4OA93) (0 ) 28784) 49 OrTA 4214 4347 4054 418s 4.23 l ~~~~~~~(-1.791) (2413 (-M (4-143 (-Z SES 0.133 4078 032 0.174 0313 (0587) ( 0) (0.46C0 (0910) (1-1) DTA 021 4.m 4.466 -079 410 (-1.424) 43)-1 (-2.625) (462) 4E26 ARTA m50B 064 Q67S (5.255) (1426)_ (S_ _ JNPTA 0200 0.13 GTA 40202 4.010 4236 0.1V9 176 t-2792 (40.120) (-2,25 (3M A SOrTA '0250 0.264 Q039 4.019 4.1S7 _______________ (0812) (0.774) (0.110) (4.081) (4V9 T 0.298 0.040 OA75 0329 0361 (1.312) (0.145) (L678) (1504) (n S93) T5s=z 4.024 40.006 40.038 4.020 4.023 (-I-le) (402M7 O(-1X79) (-130 -443) T*OrTA 0.036 4.171 4.tSS 0OA 0.0 (022Zi (41 96) ( Q784) (04OMs) rSES 40237 4124 4.A84 '0280 4B3O (40.730) {4.319 (4-21 (-1A 0-13" TADTA 4Q141 0.156 4.054 0.181 0.108 (4.766) (0.712) (.0.234) (1.190) (0.681) TARTA 4.129 0l119 A.0M2 (4.992) (0.767) (4.143) T*UPTA 4.207 40.209 (-11820) (-1.778 TGTA 0.241 0.189 0.426 4.187 40.101 (1.90S) (1.243) (2.710) (-1.766) (40.925) TSOrTA 4.001 0307 0.157 4.009 0.088 (4.003) (0.716) (0355) (4.029) (0.290) Adj.R2 0368 0.220 0.553 0.085 0.072  -28 - We can now summarle the main results of this sectionL First, it seems that financial reform did little to reduce firms' preference for internally gerated funds over extenal finance. This finding is valid for al types of fiaial instruments. Second, the importance of size In obtaining long term loans is also not d_mbnsed In the 1980L The fact that the data set is biased and cossts of largeflns makes interpretadon difflcult. In paticular, it is not poble from these results to make Infnces about the Impact of financial reform on enterpries which are smaller than those in the data set. Howeer, coupled with the slugish lncrease in financal deepening and banls' reluance to expand their customer base, both reported earlier in this paper, one is tempted to concude that fiancial reform did not yet result in increased avaiability of long term Investment fuids to smaler rms On the other hand, empiical results also suggest that collateral requirements and the adverse impact of agency problems generated by growth opportunities on acquiring short term bank loan decrased in the late 198 s may reflect an increase in banks' ability to monitor borrowers as a result of financial reform. Alternathiel, it may also reflect that as a rult of long term relations with banks, firms in the sample became better able to rely on their reputation to gaint access to bank o0111S1M TV. LESSONS OF THE TURKSH EXPE:EUENCE After the crisis of 1982, Turkey adopted a more cautious approach to fiancial reform. Efforts were made to develop a legal and insdtutional framework for the functioning of financial markets and for the supervision of fnancdat institutions. Any lessons that can be dedved from the Turkish experience need to be qualified; financial reform process is likely to require more time before Its Impact can be fully evident. With these caveats, ths section attempts to draw some prelmnary lessons from the Turkdsh experience. The new literature on the link between finance and real activity emphasizes the importnce of borrowers' net worth in determinng the availability of extenal finance. In their exposition of the recent theory, Gertler and Rose (1991) argue that a decline in borrowers' net worth is likel to reduce the extent of fnancial intermediation. More generally, the fiancal condition of the borrowiAg sectots is seen to be a citical determinant of the degree of efficiency in the allocation of loanable funds. The experience in Turkey, especially the response of the financial and corporate sectors to interest shocks confirms this view. In 1982, when the corporate sector was experiencing a decline in gross margins, liberalization of interest rates created a vicious cycle whereby distress borrowing by an illiquid corporate sector and accommodation of especially troubled banks resulted in unsustainably high intetest rates and over-borrowingff By contrast, in response to the interest rate shock of 1988, the corporate sector could rely on internally generated funds to reduce their stock of short term loans and adjust their balance sheets. The absence of distress borrowing is closely linked to the absence of a separate gross earnings shock. The lesson is that any potential adverse effect of interest rate liberalization is smaller when the corporate sector is financially strong. 2.7I In a model where firms build reputation over time, Diamond (1989) predicts that older firms will find it optimal to choose safer projects because they have incentives to maintain their reputation capitaL 2& The fact that the crisis involved ewxessive (and quite likely inefficient) intermediation rather than disintermediation is possibly explained by expectations of a de facto deposit insurance.  -29- The major shortcoming of the Turklsh financial refoirm was macroeconomic unertainty and the burden that public sector borrowing placed on financial markes Concem vith maaoewonomic luncetainty Is evident In bank ma_e' responses to survey questions, where uncertainty is Identified as a major detenuce to increas IntermdiatIon towards fim that are perceived as rlsky.A In addition, the Govemment's recoue to the fiancial system to finance the budget deficit crowded out financial floWs to the priate sector. Hence, the potential positive Impact of financial reform was considerably hndered by a lack of disdpline In goverwment flnane. In fact, It can be said that the Government has beea the main beneficiary of deveopments in fianca makets 'The lesson is that fiscal pocies may limit the benefits of financial refonL The analysis of firm-level data reveals that firms' reliance on inteunaly generated funds did not ciange over the refbrm period The importance of sie in acesing long term bank loans is also not diminibhed. lhese findings are consistent with the limited impact of financial reform on the financing of the privte sector. Nevertheless, there Is also some evidence that the Importance of agency problems and collateral requirements were reduced over the reform period, at least for the fims in the sample. The Turkish strategy of financial reform emphasized institutional development in so far as authorities tied to establish a regulatory envionment for the devel ment of foreign exchange and interbank money markets However, overwhelming focus on the banking system led to the neglect of the development of non- bank financial lnstitutions and instruments. Even though compangy data rveal the slow emergence of fnance bills as a new Instrument of corporate fiance, development of alternative souroes of funds, including the stock market, has lagged behind significantly.I Fially, it also seems that a more efficient judicial system may play an important positive role in expanding the benefits of financial reform. The problem is not only one of reforming the legislation that gover financial transactions, but perhaps more important, one of increasing the processing capacity of courts Strengthening the confidence of financal intermediares in the judicial sysem may enhance their wiagness to expand into new types of clients or areas of business, and develop more complex financal contracs ;2 Ersel and Ozturk (1990b) provide further econometric evidence on the impact of macroeconomic uncertainty, in their case identified with variability in inflation rates Their analysis of firm level data shows that greater uncertainty increases the share of short term debt and especaly short term other debt (mainly trade credits) and reduces the share of bank loans. 30 The most striking contrast would be with Korea, where direct financing became a major component of sources of funds In the 1980s. See Cho and Cole (1992).  -30- REFERENCES Ada, Z (1991), "lIterest Policy and Its Impact on Turkish Banldng System,^ Mimeographed, State Planning Organiztion, Anara (In Turkish) Akkurt, A., Haldoglu, D., Karayalcin, A., Ko4. N., Ozwet, C., Sene, A., Usta, N. and D. Varol (1991), -Developments In the Turkish Banking Sector: 1980-1990," mimeo, Central Bank of the Republic of Turkey. Akyuz, A. (1984), Financial Structure and Relations in The Ibrkish Economv. Industdal Development Bank of Turkey Publications IstanbuL ______ (1990), 'nancial System and Policies in the 1980s, In T. Aricanli, and D. Rodrik (eds.), The Political Economy in the 1980s, MacMi1an Press. Alkan, A. (1990). "Auctions for Treasury's Domestic Borrowin&" Research Report prepared for the Central Bank of the Republic of Turkey, mimeographed. (In Turkish) Atiyas, L (1990), "he Private Sector's Response to Financial Liberalztion in Turkey. 1980-1982," in T. Aricanli, and D. Rodrik (eds.), The Political Economy of Turkev in the 1980s. Macmillan Press. Atiyas, 1, H. Ersel and E. Ozturk (1992), "Customer Patterns and Credit Rationing in Turkish Banks," forthcoming, Central Bank of the Republic of Turkey, Research Department. Bayazitoglu, Y., IL Ersel and E. Ozturk (1991), FInancial Market Reforms In Turkey Between 1980-1990," mimeo, Central Bank of the Republic of Turkey. Calomiris, C W. and R. 0. Hubbard (1990), Firm Heterogeneity, Intemal Finance and 'Credit Rationing'," Economic Journal voL 100, pp. 90-104. Caprio, Gerard (1991), "Policy Uncertainty, Infotmation Asymmetries and Financial Intermediation," World Bank, WPS. No.853. ________ (1992), 'Banking on Financial Reform? A Case of Sensitive Dependence on Initial Conditions," Chapter 3 in Caprio, Atiyas and Hanson (1992). Caprio, Gerard, Izak Atiyas and James Hanson, (1992), Fmancial Reform: Theory and Experience draft manuscript, The World Bank. Cho, Y. J. and D. C Cole (1992), 'The Role of the Financial Sector in Korea's Structural Adjustment," in V. Corbo and S. Sun (Eds.) Structural Ad)ustment in a Nely Industrialized Countrv Lessons From Korea, forthcoming.  -31 - Dammon R. M. and L W. Senbet, (1988), he Effect of Tam and Depreciation on Corporate Investment and Financial Leverage, Journal of Fina June, pp. 357-373. Denizer, C (1991), Market Concentration and Bank Profitability In Turkey, mimeo. Diamond, D. (1989), "Reputation Acquisition In Debt Markets, Journal of Political Econom. VoL 97, pp. 928-862 Ersel, H. (1992), 'The Potential for Secondary Market for Government Domestic Debt Instruments in Turkey,' Central Bank of the Republic of Turkey, Discussion Paper, No. 9201/A, Ankara. Ersel, H., and E. Ozturk (1990a), 'he Credit Delivery System in Turkey,' Central Bank of the Republic of Turkey, Research Department Discussion Paper 9003, Ankara (1990b), 'Liberalization Attempts and the Financial Sture of Turkish Corporadons,' Central Bank of the Republic of Turkey, mimeo. Fry, M. J. (1972), FSnance and Development Planning h urkev. J. Bl1, Leidea. Gertler, M. and A Rose (1991), inance, Grwth and Public Policy,' World Bank, Polcy Research Worldng Papers, WPS. 814, and Chapter 2 in Caprio, Atiyas and Hanson (1992). skenderoglu, L (1989), Money Demand and CurMency Substtution in Turke: A Dynamic Model Approac1 Unpublished ALSc. Ihesis, Middle East Technial Unfesity. Iskenderoglu, L, E. Ozturk T. Temel (1991), 'he Turkish Banking System: Income Statement (1983-1989) and Balance Sheet (1981-1989),' Central Bank of the Republic of Turkey, Research Department Working Paper, February 18, Ankar Kumcu, R (1989), 'Some Evidence on Curren Substitution in Turkey,' Unpublished Manuscrpt, Central Bank of the Republic of Turkey. Myers, S. (1977), 'The Determinants of Corporate Boroowing,' Journal of Financdal Economics pp. 147-175. Pehlivan, Hatice (1991), An Analsis of Financial Liberalization and Finandal Development in Turkev 1980- 1989. Unpublished Ph.D. Thesis, University of Bradford. Pehlivan, H. and C Kirkpatrick (1991), 'The Impact of Transnational Banks on Developing Countries' Banking Sector An Analysis of the Turkish Experience, 1980-1989,' paper presented at the British Society for Middle Eastern Studies Annual Conference, School of Oriental and African Studies, University of London, July 10-12, 1991. Titman, S. and R. Wessels (1988), 'The Determinants of Capital Structure Choice,' Journal of Finance. VoL 43, pp. 1-19.  -32- ANNEX TAXI Table A.1: FINANCIAL DEEPENING MIIOM P M|NP M2Y/GNP 1982 11.6 22.8 193 126 23.9 1984 9.6 22.4 23.1 198S 89 24.1 25.U 1986 9.0 24.S 27.S 1987 9.6 23.0 27.4 198 8. 19.7 25.1 1989 7.8 20.3 25.3 1990 8.1 203 253 b2Y * M2 + Fore4n Exchange Deposit Souce: Stat itute of Statstics, CeuM Bankof te Repubc of Tuatq.  -33 - kIL&Z STOCKS OF FINANCEAL ASmS (9, OF a ) 1982 1983 1984 195 19s6" 1987 1968 1989 1990 L Cm4q In duulatoo 4.2 39 16 14 4 3.6 &1 32 3.6 IL Toaideposits 186 20.0 195 22.4 24.1 28 22.0 22.1 26.7 SW deg todsu 7?4 8.7 CO ISS 5$ 6O 5.1 4. 5.6 Thedeposts 11. 11.3 12.9 152 15 I1A 11.5 125 12.4 Fogdpw1mng 0.7 1.7 31 4.4 5.4 5.0 5.0 mL Than asuortles 6s 6.1 7.0 &1 9.4 128 12.4 13.6 232 Ll Totalp te seres 6 Zs 2 2.1 4.0 44 9.4 Phiate d nta- 06 0.5 A 0.4 0S 0.9 09 0 1.0 Sares 2.0 0 1.8 2.0 2 3.1 39 84 .2 Toal pubc eu1 3.5 4.8 6.0 6.9 9.2 84 9.0 133 Government bons 2.1 3.1 29 3.7 38 4.1 49 6.3 10.4 Thasmy BIDs 1.7 0.5 1.9 1.8 2.1 33 2.5 2.1 3.1 Reenue sbar 0.1 05 09 1.8 1.0 C 0.3 cent2e _ Total 29.3 30.0 30. 33.9 36.9 40 37.5 38S9 53.5  '34, ra3: Prn hsa |(Sak8 Volume, Billon rL) 1982 1963 1984 1965 198S 1987 1988 1989 1990 Bonds 10.7 14.8 12.1 33a3 75.0 338.0 177.s 545.5 793.8 oDMMercIal Paper - =5.8 271.0 4453 215.1 Bank BS 48. 6l1 182.0 173.9 Shasux 9.10 34.7 63.8 96.4 102.0 187.2 364.3 97.5 4106 Mutual Fund Partki- - _ _ 45.0 53.0 161.5 855.0 pation Ceuicates TO_ P_vate 19.7 49.5 75.9 129.7 225.6 692.1 1048.0 2317.7 5970.5 Securities GvOenmIenit Bonds S92 199.9 199.4 673.0 1269.4 2045.4 3816.2 9061.0 124584 Treasury Bib | 2S7.6 782 495.1 1217.0 1787.9 3954.5 5114.9 7643.3 8442.7 Revenue Sharnft - 10.0 140.0 220.0 6600 400O 600.0 Certificates Toal Public Sector 316.8 278.1 103;L1 2030.6 32773 6659. 8931.1 171043 21SO1.1 Securities I Total Secudties 336.5 WA 1108.0 21603 3509 n7 0 99791 _1942. 274716 Mema Item.: lbe Share Public 942 84.9 932 94.0 93.6 90.6 89.5 81 78.3 Sector Securites In Total (%) -r:atMkBa _ R _0= 42-43 so-4* Source: Capita Market Board, Annual Report 1990, Tables 42 43 and 45.  .35- 1T"e A.4: YIELDS OF FINANCaAL ASSETS (Des %) _ _ _ __ = = S . _- . - _ - -= 1_980 196 1962 1983 1964 1965 1966 1987 1968 1989 1990 Savgw depots d# 2. S.O 50 425 45.0 50.0 51.8 48 1 63 6.1 _73 Privte secto seurvift IV Bo"d 25.6 36N0 37.4 386 472 49.9 49.9 483 67.1 6S.4 553 Commei pae -. . 50.0 54.9 648 533 shares 120 30.4 79.9 1102 -52 468 865 295. -373 11.2 SS5 Mutual fund paripa- . .683 67 462 tion certfilctes_ Public se"tr securities W Governmet bonds 29.0 34.0 34.0 31.8 43.0 50.6 S1.0 47.0 62.4 S22 48.2 Tram"ybil 28.0 30 - - 51.0 50.7 49.6 45. 57.9 47.5 43.8 !noome sharig . - _ 43.0 50.1 59.1 95.6 952 61A ForpEgn ecbange Sf Nom!Ial appreciatio of 1S5.0 44.0 35.6 47.0 53.4 27.1 29.1 327 75.4 25.7 25.8 U.S. dollar Nominal appciation of 1233 2S.4 30.2 262 34.7 36.2 63.1 621 57.7 32.2 419 Dutschemark fx'deposits - - 40.4 41.5 44.S 89.2 S1. 34.7 US. dollar F deposits . 46.5 79 73 653 502 483 _Deutseadc_k _ Memo item: Inflation rate 110.2 36.6 29.9 31A 48.4 45.0 34.6 38.9 75.4 69.6 63.6 Source: State Institute of Statistics, Capital Markets Board, Central Bank of the Republic of Turkey. W Quarterly average of the highest after tax rate on saving deposits when compounded. AKIWRT eL aL (1991, Table 4). _y 1980-1985 figures are obtained from ADA (1991, Table 2.5), 19861990 fgurs are fm Capital Market Board Annual Report 1990, Table 39. S' 1980-1985 are obtained from ADA (1991, Table 2.5 1986-1990 figures are from Capital Market Board Annual Reoprt 1990, Table 40. _/ Inflation rate is measured by consumer price index (Yearly average)  -36 - Table SECONDARY MARKET VOLUME (Bilion L) 1986 1987 1988 1989 1990 (1) Bonds 104. 394.9 1032.2 1463.3 3006.0 (2) Cammera paper - 51.7 175.7 856.9 670.8 (3) Bank bills 364 95. 1703 188 196.7 (4) Private debt instrument (1+2+3) 140.9 542.2 1378.2 2508.5 3 (5) Shres &7 105A 1484 1735.9 113.0 (6) Private scor securties (4+S) 149.6 647.6 1526.6 424.4 1918. (7) Govemet bonds 546.3 15201 26309 108280 61742.9 (8) Treasuy bilis 1411.9 3219.8 73207 18762 32023.3 (9) Income shaig cetifctes 289.2 329.5 3946 1098.1 10783 (10) FX-Indesed bMlls - 86.4 13.7 1558,4 1631.9 (11) Publ sewo secudties (7+8+9+10) 2247.4 5185.8 10359.9 322473 96476.4 (12) Total secuitIes (6+11) 2397.0 5834 11886.5 36491.7 115662.9 Memo item: Share of public sector 93.8 88.9 872 88.4 834 securities t.ded in the seod. market ) Sor:= Capital Market Board, Annual Report 1990.  - 37 - INDUSRY SERM PAPERS No. 1 Japanese Direct Foreign Investment: Patterns and Implications for Developing Countries, Februaly 1989. No. 2 Emerging Patterns of International Competition in Selected Industrial Product Groups, February 1989. No. 3 Changing Firm Boundaries: Analysis of Technology-Sharing Alliancs, February 1989. No. 4 Technological Advance and Organizational Innovation in the Engineering Industy, Match 1989. No. S Export Catalyst in Low-Income Countries, November 1989. No. 6 Overview of Japanese Industrial Technology Development, March 1989. No. 7 Reform of Ownership and Control Mechanisms in Hungaty and China, April 1989. No. 8 The Computer Industry in Industialzed Eoonomies: Lssons for the Newly Indizin Februay 1989. No. 9 Institutions and Dynamic Comparative Advantage Electrnics Industry in South Korea and Ta}wan, June 1989. No. 10 New Environments for Intellectual Property, June 1989. No. 11 Managing Entry Into International Markets: Lessons from the East Asian Experiene, June 1989. No. 12 Impact of Technological Change on Industrial Prospects for the LDCs, June 1989. No. 13 The Protection of Intellectual Property Rights and Industrial Technology Development in Brazil, September 1989. No. 14 Regional Integration and Economic Development, November 1989. No. 15 Specalization, Technical Change and Competitiveness in the Brazin Electronics Indu"ty, November 1989. No. 16 Small Trading Companies and a Successful Export Response: Lemons from Hong Kong, December 1989.  - 38 - NDUSTRY SERIES PAPERS cont'd No. 17 Flowers: Global Subsector Study, December 1989. No. 18 The Shrimp Industry: Global Subsector Study, December 1989. No. 19 Garments: Global Subsector Study, December 1989. No. 20 World Bank Lending for Small and Medium Enterprises: Fifteen Years of Experience, December 1989 (out of prin). No. 21 Reputation In Manufactured Goods Trade, December 1989. No. 22 Foreign Direct Investment From the Newly Industrized Economies, December 1989. No. 23 Buyer-Seller Lnkb for Export Development, March 1990. No. 24 Technology Strategy & Policy for Industrial Compeddveness: A Case Study of Thailand, February, 1990. No. 25 Ihnestment, Productivity and Comparative Advantage, April 1990. No. 26 Cost Reduction, Product Development and the Real Echange Rate, April 1990. No. 27 Overcoming Policy Endogenety: Strategic Role for Domestic Competition in Indusial Policy Reform, April 1990. No. 28 Conditionat in Adjustment Lending FY8O-89. lhe ALCID Database, May 1990. No. 29 Internationl Competitiveness: Deerminants and Indicators, March 1990. No. 30 FY89 Sector Review Industry, Trade and Finance, November 1989. No. 31 The Design of Adjustment Lending for Industuy Review of Current Practioe, June 1990. No. 32 National Systems Supporting Technical Advance in Industry: The Brazilian Experlence, June 1990. No. 33 Ghanas Small Enterprise Sectr: Survey of Adjustment Response and Constaints, June 1990. No. 34 Footwear: Global Subsector Study, June 1990.  - 39 - INDUSTRY SERIES PAPERS cont'd No. 35 Tightening the Soft Budget Constraint in Reforming Socst Economies. May 1990. No. 36 Free Trade Zones in Export Strategies, December 1990. No. 37 Electronics Dedopment Strategr. The Role of Government, June 1990. No. 38 Export Finance in the Philippines: Opponunities and Constraints for Developing Countly Supplies, June 1990 (not yet published). No. 39 The U.S. Automotive Aftermarket Opportunities and Constraints for Developing Country Suppliers, June 1990. No. 40 Investment As A Determinant of Industrial Competitiveness and Comparative Advatage Evidence from Six Countries, August 1990 (not yet published). No. 41 Adjustment and Constrained Response: Mawi at the Threshold of Sustained Growth. October 1990. No. 42 Export Finance - Issues and Directions Cse Study of the Philippines, December 1990. No. 43 The Basics of Antitrust Policy: A Review of Ten Nations and the EEC, February 199L No.44 Tehnology Strategy In the Economy of Taiwan: Exploiting Foreig Linkges and Inesting In Local CapabilIty, January 1991 (not yet publid. No. 45 The Impact of Adjustment Lending on Industry in Afican Countries, June 1991. No. 46 Banking Automation and Productivity Change: The Brazilian ExprIence, July 1991. No. 47 Global Tr