O»"?a Dewey Working Paper Alfred P. Sloan School of Management Identifying Inefficiencies in Multiple Output-multiple Input Organizations Identifying Inefficiencies in Multiple Output-multiple Input Organizations Organization Performance Mathematical Programming Accounting (auditing) Nonprofit and Public Sector Organizations

Data Envelopment Analysis (DEA) , a new methodology based on linear programming concepts, provides an approach to evaluate relative technical efficiency of nonprofit organizations 1) which have multiple outputs and inputs and 2) where the efficient production function is not specifiable with precision. This paper evaluates the reliability of DEA, compared with use of ratio analysis and basic regression analysis for efficiency measurement. DEA, ratio analysis, and regression analysis are applied to an artificial data, in which the efficient and inefficient units are known. Without knowledge of the technology, DEA accurately identifies certain inefficient Decision Making units (DMU's) when outputs and input are properly specified. In contrast, the ratio analysis and regression analysis techniques are found to be less reliable for identifying inefficient DMU's. The strengths and limitations of DEA are further elaborated to anticipate issues that may arise in subsequent field applications. Acknowledgement. The author most gratefully acknowledges the beneficial guidance and careful reading of early versions of this manuscript by his thesis advisor, W. W. Cooper. His thoughtful comments were most appreciated. In addition, the redirection of the original version suggested by Arie Lewin and the insights of the reviewers have been extremely helpful in developing the authors understanding of and insights about this research.