It is often said that no improvements can be achieved without evaluation of present performance. In the case of an organisation, ensuring competitiveness necessitates a built-in mechanism, whereby systematic performance evaluation is fed back into the management process for the improvement of management efficiency or productivity.
Evaluations are conducted based on the results of measurement of activities and performance. Organisational productivity is usually assessed by measuring performance results. For this purpose, input and output must be expressed in quantitative terms. It is still not an easy task to estimate the ideal output for a given input. When it comes to "public" organisations, it is very difficult to quantify the output. It is even more difficult to calculate the cost of each product when an organisation makes several kinds of products concurrently.
Many models have been developed to meet the growing need for the measurement of efficiency. These include the performance indicator, comparative performance analysis and measurement of the relative efficiency of units performing similar tasks.
This paper discusses the usefulness, especially in the context of Korean banks, of the Data Envelopment Analysis (DEA) method which was developed as a model for the productivity evaluation of not-for-profit organisations.
The DEA method involves application of linear programming techniques to measure the management efficiency or productivity of organisations. The DEA uses empirical data to form a data envelopment, the edge of which expresses an efficiency frontier. By comparing efficiency frontiers, more efficient organisations can be distinguished from less efficient ones. The DEA then identifies and analyses factors which contribute to the inefficiency of organisations.
One of the strengths of DEA, which was developed by Professors A. Charnes and W.W.Cooper at the University of Texas at Austin in 1978, is that the amounts of various outputs and their inputs need not necessarily be expressed in monetary terms. Instead, the amounts, as expressed in the units used in the original data, can be inserted into the model without any further conversion. This has several advantages. One is that measurement errors, which may take place when the traditional cost function approach presupposes a hypothetical specific function model, are avoided. The other is that the DEA can ensure a reliable and comprehensive evaluation. This is not easy when the ratio analysis approach or the productivity index approach is used. Both these approaches have drawbacks as they use many ratios or indices, some of which often lack consistency.
The DEA has certain additional features. The overall performance of government invested organisations is usually evaluated in Korea by summing up the weighted scores for each of the performance areas. Problems arise when the weightages are assigned in a subjective fashion. The overall performance may vary significantly depending upon the weightage given for each of the performance areas. This may also be avoided by using the DEA. Further, as the DEA deals with inputs and outputs without time lag, possible differences in measurements which may arise out of the substitution of input factors, can be avoided.
With all the strengths of the DEA method, attention must also be drawn to its limitations. Since the DEA is a method of evaluation in relative terms, a number of organisations performing the same or at least similar work need to be identified for the purpose of comparison. In other words, there must be a relatively large number of organisations doing the same or similar job: county offices, banks, schools, hospitals, courts of law, military camps, etc. Second, the DEA presupposes the existence of reliable empirical data. The reality, however, is that the Management Information System (MIS) is rarely seen to function well in most cases. In the absence of such information, financial statements serve as the primary source for performance evaluation. Moreover, as the DEA requires physical data for analysis, problems will arise unless a reliable physical data is stored in a database.
The productivity of 20 domestic banks was measured in a comparative perspective by the author in 1992. Initially, the author chose some county offices, hospitals and government-supported research institutions as research subject. Around that time, however, domestic financial industries remained sheltered in the government umbrella, despite increased foreign pressures for liberalisation and increased competition. It was felt that the banking industry needed management innovation most urgently to compete with its foreign counterpart. This situation made the author change the research subjects to banks.
Korean banks had long been expected to contribute to the public interest, not to mention maximising profitability. They were also characterised by the production of multi-item services. It was, therefore, felt that their management efficiency or productivity could not be appropriately expressed by just citing business profits or net profits. If profitability were the only business objective, the Citizens National Bank and other commercial banks would avoid too many small-amount transactions. In practice, however, the banks cannot, because of their public nature, turn down small amount transactions with individuals or small-scale business concerns. Under these circumstances, it would not be justifiable if short-term finance companies were considered more efficient than' commercial banks merely in terms of the per-capita deposit amount or per-capita business profits.
The Citizens National Bank would not have received a high performance rating if profitability had been the major factor for consideration. The Bank has, however, developed a good EDP system a long time ago and had very successfully managed a very heavy workload coming from a myriad of small-amount transactions. With this perspective in mind, the Bank would be labelled productive.
For a balanced evaluation of banks, contribution to the public interest had also to be considered in addition to profitability as depicted in the financial statements. In the research conducted by the author in 1992, the production mechanism and input-output variables were identified at the outset. The number of employees, the building space, capital and annual expenditure were used as the major input variables. The major output variables included the volume of deposits, loans, investment in securities, foreign exchange fees, and number of transactions. The number of transactions was included as it was considered to represent technical productivity. If a bank handled a large number of transactions efficiently, even if they were small in amount, it would be considered productive in terms of goal-orientation or effectiveness.
Data for five years was collected for the aforesaid input and output variables for 20 banks, and a DEA was conducted on the computer. As a result, seven banks were considered efficient while thirteen inefficient. It should however be remembered that efficiency is a relative concept. In other words, efficient banks can be so termed only when the other banks are less efficient in comparison with them. For example, four banks, named A, B, C and D for convenience, had a similar production structure and produced the same amount of output. If Bank A used 414 employees, 16.099m2 of building space and US$ 13,375,000 of capital more than the others for the same output, this meant that Bank A used the three input factors excessively, as cited before, or unproductively or inefficiently in other words, and that management improvement would probably cut down on these excessive input factors.
One of the other research findings worth mentioning the fact is that the 20 banks were categorised into two groups: national commercial banks and local commercial banks and that the former turned out to be more efficient or productive than the latter. The reliability of this finding was tested by referring to the relevant performance evaluations by the Office of Bank Supervision (the national banking regulatory agency) and the results of a survey of the banks' middle managers. The test results showed that the finding was quite reliable.
What should be mentioned here with special emphasis is that the result of a DEA may vary with the choice of input-output variables. A good set of input-output variables will ensure higher reliability of the analysis.
The DEA is a simple method for practical application. It can serve as an easy tool, especially at the initial stages of audit, for sampling of audit subjects or for finding possible weaknesses in the audited projects. When there are many entities doing almost the same work such as public hospitals, schools, military camps, bank branch offices, and county offices, to cite just a few, a field audit cannot be carried out for all the entities falling under the same category, and the need for sampling arises. Using the DEA, auditors can easily identify relatively inefficient or unproductive entities for a field audit, and will identify possible weaknesses or defective work areas of the chosen entities. Needless to say, auditors should supplement the findings using the DEA, with other audit techniques.