線型模型에 의한 統制시스템의 設計 = A Multi-model Control System

다국어 초록 (Multilingual Abstract)

The purpose of this study is to provide a new method of setting standards for control system and is confined to the estimation of a cost function for the operations performed by several decentralized units in an organization. Especially a uniform cost control system is applicable to the operationally homogeneous units. In management literature, 'control' is defined as a process that glides activities toward an established gool or plan. The concept 'control' can be further explained as follows ;
a. The first phase is to set standards that represent the desired performance or to adopt an operation plan.
b. The second phase is to compare actual results with standards or the plan.
c. The third phase is to analyze the operating variance planning and corrective action.
The standards may be classified into (i) the attainable ideal standards representing the performance levels that are considered highly or most efficient under normal conditions, (ii) the attainable normal standards representing the average performance levels under normal operating conditions over a period of time, and (iii) the theoretical ideal standards represennting the best performance possible levels theoretically. From this point of view, we may regard the empirically estimated maximum output for a given combination of inputs in a production system as the attainable ideal standard. For the estimation of the attainable ideal standards, the concept of frontier cost function is to be reviewed first of all.
The modeling of production activities has long occupied a central role in econometric research, especially in searching for statistical estimators of the theoretical frame and in providing a stimulus for the development of new methods. Traditionally, accounting has been the major source of management information and managerial accounting, in particular, has provided data for the planning, control, and internal performance evaluation. The goal of this study is to contribute to cost accounting system in providing
more useful information in the planning and control process of management. The study is attempted to expand the models for analysis of cost behavior based on a multiple regression model and a linear programming model. It is formulated to estimate the frontier cost function which reflects the best operation possible in the system, which the multiple regression model yields the "average" cost function during a time period throughout the units. The frontier cost function can be used to yield the "target" cost which is accomplishable under the given conditions in the system.
The ANOCOVA model includes the cost efficiency parameter, Ai, to reflect the difference in the cost efficiency among the units. The ANOCOVA model improves the estimation of the average cost function by estimating the specification bias which stems from excluding the cost efficiency parameter the multipe regression model. Accarding to Johnston the ANOCOVA also utilizes "The best source of information for the short-run analysis" of cost behavior with the pooling data.
The cost function from the ANOCOV A model yields the "average" cost of operation in each unit under the normal operating conditions.
Once an operation results in an actual cost, we can analyze the performance of each unit by comparing ;
(1) The target cost. (by L. P. model)
(2) The average cost for the unit. (by ANOCOVA model).
(3) The actual cost.
This is another version of the traditional cost accounting variance analysis but the above method is unique in the sense that the target cost and the expected average cost for the unit are realistic estimates based on the analysis of the actual capabilities of the system under the given conditions.
In the empirical study, the data obtained from the 21 branches of the A bank in seoul is used to estimate the parameter of cost function by linear programming model for a firm as the attainable ideal standard. And using classical linear regression analysis, the attainable normal standard is estimated and compared with the frontier. In order to continue this study, we should find performance variables and influencing factors of firm performance efficiency and time performance efficiency.
If we are able to find the performance variables, from the analysis. we can achieve the good theory to find the efficiency variables.

목차 (Table of Contents)

• Abstract
• 第1章 序論
• 1. 硏究의 目的
• 2. 硏究의 方法
• 3. 硏究의 範圍 및 論文의 構成
• 第II章 統制시스템에 관한 理論的 考察
• 1. 差異統制의 性格과 意味
• 2. 管理會計에서의 統制制度
• 3. 統計的 管理制度
• 4. 費用函數를 利用한 計量分析方法
• 5. 線型模型에 의한 統制制度
• 第III章 線型模型에 의한 統制시스템의 設計
• 1. 重回歸分析에 의한 統制模型
• 2. 線型計劃法에 의한 統制模型
• 3. 共分散分析에 의한 統制模型
• 1) 業務能率과 費用
• 2) 共分散分析에 의한 費用函數의 推定
• 3) 共分散分析의 統計的 假定檢定
• 第IV章 線型模型에 의한 統制 시스템의 活用
• I. 重回歸分析을 活用한 統制 시스템
• 2. 線型計劃法을 活用한 統制 시스템
• 3. 共分散分析을 活用한 統制 시스템
• 4. 費用分散의 分析과 線型模型에 의한 統制
• 5. 豫測差異와 機會原價差異 및 線型模型에 의한 統制
• 6. 業績評價와 線型模型에 의한 統制
• 7. Demsky의 硏究
• 第V章 實證的 硏究
• 1. 資料의 構成 및 變數의 選定
• 2. 線型計劃模型에 의한 電費用函數의 推定結果 및 解析
• 1) 重回歸分析模型에 의한 費用函數의 推定話果
• 2) 線型計劃模型에 의한 費用函數의 推定話果
• 3) 共分散分析模型에 의한 費用兩數의 推定結果
• 3. 推定結果의 原價統制模型에의 活用
• 4. 각 模型의 推定値에 관한 graph 比較
• 5. 實證硏究의 要約
• 第VI章 結論
• 參考文獻