우리나라 은행산업의 효율성에 관한 실증분석 : DEA, Tobit, malmquist 기법을 중심으로 = A Panel Analysis on Banking Efficiency in Korea : Using the DEA, Tobit, Malmquist Index

다국어 초록 (Multilingual Abstract)

Korea financial institutions experienced substantial changes in the last 10 years. Fiercer competition among both banks and non-bank financial intermediaries, technological progress, reduced information costs and ongoing deregulation in the wake of the foreign currency crisis led to substantial changes in numerous financial systems.
Although the health of the Korea banking system has improved, its profitability remains weak. As net interest margins continue to decline and remain below those in other developed countries, enhancing core profitability remains an important challenge for banks.
There has been relatively abundant academic research undertaken on the profitability and cost efficiency of Korean bank, especially for the post-1997 period. Most studies have focused on the use of parametric and non-parametric techniques to analyse cost and overall technical efficiency of Korean banks. This paper utilizes data for the period 1998-2008 to offer a fresh perspective on the cost and overall efficiency of Korean banks. This study employed non-parametric approaches, which were the Data Envelopment Analysis (DEA) and Malmquist Production Index similar to most previous studies. This study also employs parametric approach, which are the Type Ⅰ and Type Ⅱ Tobit analysis in contrast to the previous studies.
The DEA is a linear programming technique for evaluating performance and benchmarking in a multivariate setting. The methodology uses information on the input-output combination of individual entities to construct an efficiency frontier enveloping the data. This frontier is then used to measure the efficiency of the individual entities relative to a benchmark entity, chosen by the model. The DEA produces efficiency estimates without a priori functional restrictions on the underlying production process. By duality, the DEA can be used to assess either cost or revenue efficiency, depending on the setup. Cost efficiency looks at how banks use their inputs to produce a given level of outputs. Revenue efficiency examines how much output banks can produce using the same inputs. This paper analyzes the efficiency and profitability of Korean banks from 1998-2008. This paper uses data envelopment analysis to analyze the cost and revenue efficiency of Korean banks.
One of the crucial issue to build a model for the assessment of banking efficiency is th identification of appropriate inputs and output. This issue is not straightforward and an extended and unresolved controversy remains in the literature. The two approaches most often used are the production and intermediation approaches, although other approaches such ad the asset, user-cost and valued added approaches are also well established in the banking literature.
The production approach emphasizes the operational activity and thus banks are primarily viewed as providers of services to customers. Under the intermediation approach financial institutions are views as primarily intermediating funds between savers and investors, Although these approaches can offer valuable guidelines regarding the definition of appropriate banking models to asses certain objectives, the empirical studies often emphasize particular issues of concern to the organizations, such that the choice of inputs and output of banking studies is very much influenced by the analysts' view of the banking activity, the issues under analysis, and the availability of data.
Here, the framework assumes that banks use two inputs to produce two outputs. The outputs consist of loans and deposits. The inputs include number of employees or number of branches and fixed assets. This paper obtained data from the balance sheet and profits and loss accounts of the FISIS (Financial Statistics Information System) that the Financial Supervisory Service reported.
When the number of emplyees (number of branches) and fixed assets are utilized input factors and loaned money and deposits as output factors, inefficiency of regional banks is large, and it was showed that the number of branches is the important question rather than the number of employees.
Such results are similar to those of the case when only loaned money is considered as output factor, but when only deposits are considered as output factor, there are not any significant differences in efficiencies between regional banks and commercial banks.
I employed Malmquist production index with the efficiency of the DEA models. The results showed that total productivity increased in all the factor combinations, but as technological efficiency decreased, banks decided frontier, indicating the expanded gap between leading banks and inefficient banks due to reorganization of banking markets.
This study decomposed Malmquist productivity index into two components, namely, efficiency change and technical change and found that most of growth of the total factor productivity was due to the technical change rather than the efficiency change. As technological efficiency decreased, however, the leading banks decided technological renovation for management improvement and then moved to efficient frontier. It has widened the efficiency gap between leading decision making units and inefficient banks.
The Tobit analysis including type Ⅰ and type Ⅱ models was also employed to regress the variable on the relevant explanatory variables. The censored regression technique that the dependent variable was the truncated technical efficiency based on the constant returns to scale over the 2003-2008 period showed that ROA had negative effects on efficiency. The sign and significance of ROA and BIS depended on which model was employed and which input-output combination was used. BIS had strong negative relations with efficiency when loans or deposits were used as the output factor, while asset was positively significant in the deposits only.
The efficiency analysis presented in this paper, of course, uncovered the existence of considerable inefficiencies may be the fact that cost efficiency is not the sole driver in the management of commercial activity of banks. Other targets, related to marketing success, quality of service, and financial gains from the intermediation activity may be given higher priority over the achievement of cost efficiency. This study also tried to cover all the presently available models, but the results were not so satisfactory as to explain and predict the competitiveness and efficiencies of Korean banking system accurately. This menas that there are many other aspects to be considered in treating DEA besides those attended to in this paper. A more extensive examination is left to a next paper.

목차 (Table of Contents)

• <목 차>
• Abstract
• 제1장 서론 1
• 제1절 연구배경 및 목적 1
• 제2절 연구방법 및 구성 2
• 제3절 선행연구검토 4
• 제2장 은행산업의 변화 13
• 제1절 은행산업의 구조적 변화 13
• 1. 외환위기 이전의 은행산업 13
• 2. 외환위기 이후의 은행산업 15
• 제2절 은행산업의 시장집중도 변화 19
• 1. 외환위기 이전의 시장집중도 19
• 2. 외환위기 이후의 시장집중도 20
• 제3장 은행산업의 효율성 분석방법 24
• 제1절 DEA에 의한 효율성 분석방법 24
• 1. DEA 분석모형 24
• 2. CCR모형과 BCC모형 27
• 제2절 Tobit 모형 30
• 제3절 DEA/Window 분석 32
• 제4절 Malmquist 생산성지수 33
• 제4장 우리나라 은행산업의 효율성에 관한 실증분석 39
• 제1절 변수선정 및 기술통계량 39
• 1. 변수선정 및 분석기간 39
• 가. 변수선정 39
• 나. 분석기간 42
• 2. 투입 · 산출요소의 기술통계량 43
• 제2절 DEA 분석 46
• 1. DEA - CCR모형 46
• 2. DEA - BCC모형 57
• 3. 규모의 효율성 60
• 4. 초효율성 분석 64
• 제3절 Tobit 분석 68
• 제4절 DEA/WINDOW 분석에 의한 효율성 분석 78
• 제5절 Malmquist 생산성지수 분석 80
• 1. 2003년 - 2005년의 생산성 변화 80
• 2. 2006년 - 2008년의 생산성 변화 83
• 제6절 분석의 결과 85
• 제5장 요약 및 결론 89
• 참 고 문 헌 92
• 부 록 99
• <표 목차>
• <표 1-1> 외국 은행을 대상으로 DEA 기법을 적용한 주요 문헌연구 11
• <표 1-2> 국내 은행을 대상으로 DEA 기법을 적용한 주요 문헌연구 12
• <표 2-1> 연도별 은행 증감 추이 16
• <표 2-2> 연도별 은행구조조정 현황 18
• <표 2-3> 외환위기 이전의 시장집중도 : 1995-1997 19
• <표 2-4> 외환위기 이후의 시장집중도 : 1998-2001 21
• <표 2-5> 은행의 대형화에 따른 시장집중도 : 2002-2008 22
• <표 2-6> 은행산업의 일인당 생산성과 비용-수익비율 23
• <표 3-1> 윈도우 수 32
• <표 3-2> DEA/Window 특성 33
• <표 4-1> 분석모형의 투입물 &#8228; 산출물 변수 42
• <표 4-2> 투입물과 산출물의 기술통계량 : 시중은행과 지방은행 44
• <표 4-3> 국내은행의 기술효율성(CCR 모형) 48
• <표 4-4> CRS모형에 의한 슬랙(2003년) 50
• <표 4-5> CRS모형에 의한 슬랙(2004년) 51
• <표 4-6> CRS모형에 의한 슬랙(2005년) 53
• <표 4-7> CRS모형에 의한 슬랙(2006년) 54
• <표 4-8> CRS모형에 의한 슬랙(2007년) 55
• <표 4-9> CRS모형에 의한 슬랙(2008년) 57
• <표 4-10> 국내은행의 순수기술효율성(BCC 모형) 59
• <표 4-11> 국내 은행의 규모효율성 63
• <표 4-12> CCR모형에 의한 초효율성 분석 : 1998년-2002년 66
• <표 4-13> CCR모형에 의한 초효율성 분석 : 2003년-2008년 67
• <표 4-14> Tobit 분석 : 고정자산, 예수금에 의한 효율성 68
• <표 4-15> Tobit 분석 : 고정자산, 대출금에 의한 효율성 69
• <표 4-16> Tobit 분석 : 고정자산, 대출금에 의한 효율성 70
• <표 4-17> Tobit 분석 : 고정자산, 대출금에 의한 효율성 71
• <표 4-18> Tobit 분석 : 직원수, 고정자산, 예수금, 유가증권, 대출금에 의한 효율성 72
• <표 4-19> Tobit 분석 : 직원수, 고정자산, 예수금, 유가증권, 대출금에 의한 효율성 73
• <표 4-20> Tobit 분석 : 직원수, 고정자산, 예수금, 유가증권, 대출금에 의한 효율성 74
• <표 4-21> Tobit분석 :고정자산, 직원수, 예수금, 대출금, 유가증권에 의한 효율성 75
• <표 4-22> Tobit분석 :고정자산, 직원수, 예수금, 대출금, 유가증권에 의한 효율성 76
• <표 4-23> Tobit분석 :고정자산, 직원수, 예수금, 대출금, 유가증권에 의한 효율성 77
• <표 4-24> CCR 모형의 창 분석 : 2003-2005년 79
• <표 4-25> CCR 모형의 창 분석 : 2006-2008년 80
• <표 4-26> Malmquist 생산성 지수모형을 적용한 생산성 변화 : 2003-2005 81
• <표 4-27> 기간별 생산성 변화 : 2003년-2005년 82
• <표 4-28> Malmquist 생산성 지수모형을 적용한 생산성 변화 : 2006-2008 83
• <표 4-29> 기간별 생산성 변화 : 2006년 - 2008년 84
• <그림 목차>
• <그림 3-1> TE, PTE, SE의 측정 26
• <그림 3-2> 총요소 생산성과 산출거리함수의 Malmquist 산출기준지수 35
• <그림 4-1> 투입물 평균값의 변화추이 45
• <그림 4-2> 산출물 평균값의 변화추이 45
• <그림 4-3> 국내은행의 효율성 평균값 추이 60
• <그림 4-4> 규모효율성 평균값 추이 64

참고문헌 (Reference)

1. Production Frontiers, Fare R, Lovell, Grosskopf, , 1994

2. DeterminantsofInnovation, Bhattacharya, Determinants of Innovation, , 2002

3. Data Envelopment Analysis, Stocker, Norman, , 1991

4. Fundamentals of Production Theory, F?re.R, , 1988

5. 중국 상업은행 효율성 분석, 오대원, 국제지역학회, , 2007

6. Theory of Cost and Production Function, Shephard, , 1970

7. 합병은행의 경영효율성 분석, 황진수, 한국산업경제학회, "『산업경제연구』, , 2005

8. The Measurement of Productive Efficiency, Ferrell, , 1957

9. Banking Efficiency in the Nordic Countries, Souminen, Hjalmarsson, Berg, , 1993

10. 은행구조조정 1년의 성과와 변화, 삼성경제연구소, 삼성경제연구소, , 1999

11. Quantifying Management Role in Bank Survival, Siems, , 1992

12. Measuring Efficiency of Decision Making Units, Charnes, W.W.Cooper, E.Rhodes, , 1978

13. Measure Cost Right : Making the Right Decision, Copper, Kaplan, , 1988

14. International Trade and Industrial Concentration, Kumar, , 1985

15. 외환위기 이후 은행산업의 생산성 변화, 최문경, , 2006

16. 외환위기 이후 은행 점포의 효율성 분석, 김정인, 이상규, 한국경영학회, "경영학 연구, , 2003

17. DEA를 이용한 신용협동조합의 효율성 평가, 홍봉영, 구정옥, 한국재무관리학회, DEA를 이용한 신용협동조합의 효율성 평가, , 2000

18. Measurement and Efficiency Issues in Commercial Banking, Humphrey, Beger, , 1992

19. 은행의 효율성과 생산성 변화의 결정요소, 모수원, 유진하, , 2008

20. Scale Economies in Banking : A Restructing and Reassessment, Benston, Humphrey, Hanweck, , 1982

21. An AnalysisofProductionsasan EfficientCombination ofActivities, Koopmans, , 1951

22. Data envelopment analysis for managerial control and diagnosis, Epstein, Data Envelopment Analysis for Managerial Control and Diagnosis, , 1989

23. Detecting Influential Observation in Data Envelopment Analysis, Wilson, , 1995

24. WTO가입 전후 중국상업은행의 효율성 비교 분석, 조대우, 제혜금, , 2007

25. A Procedure or Ranking Efficient Units in Data Envelopment Analysis, Anderson, , 1993

26. Bank Branch Operating Efficiency:Evaluation withDataEnvelopmentAnalysis, Sherman, Bank Branch Operating Efficiency: Evaluation with Data Envelopment Analysis, , 1985

27. OECD 국가들의 은행산업 효율성 변화 및 결정요인 분석, 최승빈, OECD 국가들의 은행산업 효율성변화 및 결정요인 분석, , 2003

28. Innovation, Market Structure and Fire Size: A Simultaneous Equations Mod, Koeller, "Innovation, , 1995

29. 은행영업점의 성과평가방법으로서의 DEA: 테스트와 비교, 안태식, 은행영업점의 성과평가방법으로서의 DEA:테스트와 비교, , 1991

30. 은행의 방카슈랑스영업 기술효율성 및 생산성변화 연구, 김재현, 정재욱, 조강필, , 2007

31. A study of Relative of Bank Branches:AnApplication Data EnvelopmentAnalysis, Vassiloglou, , 1990

32. Technical changes, Organizational Form, and the Strucure of Bank Production, Timme, Hunter, "Technical changes, , 1986

33. DEA모형을 이용한 은행지점의 효율성 평가절치 및 사례분석, 서우종, , 2001

34. An Envelopment-Analysis Approach to Measuring the Managerial Efficiency of Bank, Barr, Siems,

35. Productivity Developments in Swedish Hospitals: A Malmquist Output Index Approach, Roos, Grosskopf, Lindgren, F?re, , 1989

36. 우리나라 은행산업의 생산성변화 요인:Malmquist 방법론의 적용, 이상규, 권영준, , 1999

37. DEA and Commercial Bank Performance : A Primer with Applications to Missouri Banks, Yue, , 1992

38. Productivity Growth in the Hellenic Banking Industry, State versus Private Banking, Noulas, "Productivity Growth in the Hellenic Banking Industry, , 1997

39. A complete efficiency ranking of decision making units in data envelopment analysis, Mehrabian S, G.R.Jahanshahloo, M.R.Alirezaee, A Complete Efficiency Ranking of Decision Making Units in data Envelopment Analysis, , 1999

40. 허핀달지수는 CRk지수보다 우월한가: 가설 검증 및 정책적 함의, 이재우, 장영재, , 2000

41. 한국과 일본 조선산업의 효율성 및 생산성 비교분석에 관한 연구, 이기문, , 2007

42. Best Practice Analysis of Bank Branches An Application of DEA in a Large Canadian Banks, Paradi J.C, Schaffnit C, , 1997

43. 은행위기 전후의 5개 대형은행의 효율성 분석: DEA기법을 적용하여, 조병택 ( Byung Taik Cho ), 신동진 ( Dong Jin Shin ), 한독경상학회, , 2005

44. Productivity Changes in Swedish Pharmacies, 1980-1989: A Nonparametric Malmquist Approach, F?re, "Productivity Changes in Swedish Pharmacies, , 1992

45. Productivity Growth, Technical Progress and Efficiency Changes in Industrialized Countries, F?re.R, Z.Zhang, "Productivity Growth, , 1994

46. The economic theory of index numbers andthe measurement of input, output, and productivity, Erwin, "The Economic Theory of Index Numbers and the Measurement of Input, , 1982

47. Some Models for Estimating Techinical and Scale Inefficiencies in Data Envelopment Analysis, A.Chanes, W.W.Cooper, Banker, , 1984

48. DEA 모형을 이용한 전라북도지역 신용협동조합의 효율성 측정에 관한 연구, 고상순, , 2003

49. A Non-parametric Approach to Measurement of Efficiency and Technological Chang; The Case of Large U.S, Mehdian, Elyasiani, , 1990

50. DEA 기법, Malmquist 생산성 지수, Tobit 모형을 이용한 은행의 효율성에 관한 연구, 유진하, "DEA 기법, , 2009

51. IMF 체제 돌입 전후 시점의 국내 은행들의 경영효율성 평가:DEA 기법을 적용하여, 이용주, 한국생산성학회, , 2000

52. Using Data Envelopment Analysis to Measure the Efficiency of Nonprofit Organization : A Critical Evalution, Nunamaker, , 1985

53. Nonparametric Frontier Models for Assessing the Market and Cost Efficiency of Large-Scale Bank Branch Networks, Athananssopoulos, , 1997

54. Distinguishing Technical and Scale Efficiency on Non-Convex and Convex Technologies: Theoretical Analysis and Empirical Illustrations, Kerstens, , 1998

55. Efficiency of Swedish Banking Industry: An Application of Data Envelopment Analysis, Fifth European Workshop on Efficiency and Productivity Analysis, Mlima, "Efficiency of Swedish Banking Industry: An Application of Data Envelopment Analysis, , 1997