DEA模型을 利用한 單一意思決定單位의 多時點間 相對的 效率性 測定에 관한 硏究 = Multi- period efficiency measurement for a DMU using the DEA model

다국어 초록 (Multilingual Abstract)

The purposes of this thesis are first to examine the applicability of DEA models for the multi-period efficiency measurement for the single DMU, and second, to propose the procedure of the method.
Static efficiency measurement can be performed by the comparison among the efficiencies of multiple DMUs for a specific period of time. Most previous DEA models for efficiency measurement are static methods. Dynamic efficiency measurement is done by the comparison among the efficiencies of the single or multiple DMUs for multiple periods of time. The Mamlquist-type Productivity Index(MPI) model is a typical example of the dynamic efficiency measurement methods.
The DEA models for static efficiency measurement are assumed the homogeneity of DMUs involved. If the DMUs are not homogeneous, the comparison of efficiencies of multiple DMUs is meaningless. Hence, if the single DMU has no homogeneous DMUs for comparison, the DEA model still can not be applied to static efficiency measurement.
The productivity index models with MPI for dynamic efficiency measurement are the methods of productivity changes measurement for two periods rather than the methods of efficiency measurement for multiple periods or point of time. Hence, If we are going to measure the dynamic efficiency over three periods using the MPI model, the multiple periods are divided into two periods. These are not the dynamic efficiency comparison truly, because multiple periods are not compared all together at the same time.
In this paper we are going to propose a DEA model for dynamic efficiency measurement of the single DMU, which does no have any other homogeneous DMUs to compare, as based on the follwing.
The first, the efficiency measurement can be performed for the multiple periods. Because the several productivity index models with MPI are the methods to compare the productivities of two periods.
The second, the comparative efficiency comparison can be applied for the single DMU. Because the input/output data of multiple periods for the single DMU can be treated as that of �different DMU's by the window analysis in Charnes et al(1985).
Therefore, we can define it as following.
(ⅰ) If the relation of input/output for multiple periods are homogeneous, the DEA models can be applied to compare for efficiency mesurement among the multiple periods.
(ⅱ) By above definition (ⅰ), DEA model can be applied for the single DMU above.
(ⅲ) By above definition (ⅰ) and (ⅱ), the single DMU can be found the new improvements of efficiencies for the previous static DEA models.
We try to propose that the new method of DEA model as following, by above definitions.
(Ⅰ) If the single DMU has no homogeneous comparing set, the DEA model is applied to the input/output data of the multiple periods for the single DMU.
(Ⅱ) The DMUs which were measured to efficiencies 1.0 by previous static DEA method are followed up the procedure (Ⅰ).
We can apply the dynamic DEA model for the single DMU, and can find the new improvements for the DMUs were measured to static efficiencies 1.0 by those methods and procedures.
The next, we tried to the empirical analysis using the above procedure (Ⅰ) and (Ⅱ). The target industries were all of the bio-medical/pharmacy and the security corporations in the domestic stock market. In result, the two and three DMUs in each of the two industries were measured to static efficiencies are 1.0, and one of the two DMUs and one of the three DMUs were found the improvements using the dynamic DEA model. We also implemented to compare the result of the dynamic efficiency measurements with the trend of two industries, and we were able to see the similarity by above comparison.
In spite of those tries and results, the problems of previous static DEA models itself were could not be solved. The first problem was the selection of input/output factors, the second was the selection guideline/scale decision of comparing DMUs set, the third was the time lag of input/output timing. This paper could not overcome those problems, too.
However, in conclusion, the contribution of this paper is to extend the application range of the DEA model from multiple DMUs to a single DMU, and from a specific period to multiple periods. Moreover, we revealed to usefulness of the DEA model, which is can be applied to comparative efficiency measurement for static and dynamic in the same way.

목차 (Table of Contents)

• 제 1 장 서론 = 1
• 제 1 절 문제의 제기 = 1
• 제 2 절 연구의 목적 = 2
• 제 3 절 연구 방법 및 구성 = 3
• 제 2 장 효율성 이론과 측정 방법 = 6
• 제 1 절 효율성의 개념 및 이론 = 6
• 1. 효율성의 개념 = 6
• 2. 효율성에 관한 이론 = 8
• 제 2 절 효율성 측정을 위한 제 방법 = 12
• 1. 정태적 방법론 = 12
• 가. 비율분석법 = 12
• 나. 회귀분석법 = 13
• 다. DEA모형 = 15
• 2. 동태적 방법론 = 34
• 가. 지수체계 모형 = 34
• 나. MPI모형 = 36
• 제 3 절 선행연구 고찰 = 38
• 1. 효율성에 관한 선행연구 = 38
• 2. DEA모형에 관한 선행연구 = 41
• 3. 동태적 효율성에 관한 선행연구 = 43
• 제 3 장 연구의 틀 = 45
• 제 1 절 DEA모형의 다시점간 효율성 비교 가능성 제시 = 45
• 1. 단일DMU에 대한 다시점간 효율성 비교의 의의 = 45
• 2. 다시점간 비교 가능성 = 48
• 가. 지수체계 방법론에 의거한 가능성 = 48
• 나. Sengupta의 방법론에 의거한 가능성 = 49
• 다. Wei, Sun과 Xiao의 방법론에 의거한 가능성 = 50
• 3. DEA모형 적용 가능성 = 51
• 가. Charnes의 window분석법에 의거한 가능성 = 51
• 나. 효율성 비교 관점의 전환에 의거한 가능성 = 53
• 4. DEA모형 적용을 위한 요건 및 가정 = 55
• 가. 표준속성과 상충관계 = 55
• 나. 투입·산출관계의 적시성 = 56
• 제 2 절 적용 단계 및 절차 = 57
• 1. 측정방법의 통합 및 적용 = 57
• 2. 통합적용 방법의 결과 및 해석 = 61
• 제 4 장 수치예 적용 및 결과 비교 = 63
• 제 1 절 절차 설계 = 63
• 1. 목적 = 63
• 2. 분석자료의 설정 = 63
• 3. 변수의 설정 = 65
• 4. 비교 방법의 결정 = 66
• 5. 분석 도구의 결정 = 66
• 제 2 절 분석 및 결과 해석 = 68
• 제 3 절 기존 방법과의 결과 비교 = 71
• 1. MPI모형의 결과 및 비교 = 71
• 2. 서열상관계수 비교 = 73
• 3. 비교 결과의 정리 = 75
• 제 5 장 실례 적용 = 76
• 제 1 절 분석자료 수집 = 76
• 제 2 절 투입·산출 변수 설정 = 78
• 제 3 절 DEA모형의 수립 = 80
• 제 4 절 분석결과 및 해석 = 85
• 1. 한시점·다수DMU간 분석결과 = 85
• 2. 단일DMU·다시점간 분석결과 = 86
• 3. 다시점간 효율성 개선치 제시 = 89
• 제 5 절 결과 정리 = 92
• 1. 의약품제조업에 대한 분석결과 정리 = 92
• 2. 증권업에 대한 분석결과 정리 = 93
• 3. 다시점간 효율성 비교에 대한 결과 정리 = 94
• 제 6 장 결론 및 시사점 = 96
• 참고문헌 = 98
• Abstract = 107
• 부록 = 110