Benchmarking with DEA, SFA, and R

BAN423 Benchmarking with DEA, SFA, and R

Autumn 2018

  • Topics

     First, we will discuss why to benchmark, the concepts of key performance indicators (KPIs), and different measures of efficiency and productivity (e.g., radial vs. non-radial measures; technical efficiency vs. cost efficiency).

    Second, efficiency measurement with the non-parametric and deterministic data envelopment analysis (DEA) is covered. The measurement of technical efficiency, cost efficiency, allocative efficiency, and scale efficiency will be discussed. The underlying assumptions and strength and weaknesses of this method are outlined.

    Third, efficiency measurement with stochastic frontier analysis (SFA) is covered. Starting from simple regression approaches based on ordinary least squares (OLS), we introduce models that incorporate noise and inefficiency simultaneously when estimating technical efficiency and cost efficiency. Underlying assumptions regarding functional forms and distributions are discussed, statistical inference and strength and weaknesses of this method are outlined.is outlined.

  • Learning outcome

    Benchmarking is the systematic performance evaluation of decision making units (DMUs), such as, e.g., firms, organizations, industries, or governments. The course will provide students with the following knowledge and skills:

    Knowledge

    • Foundations of efficiency analysis: concepts from production theory, efficiency & productivity
    • Understanding how Data Envelopment Analysis (DEA) and Stochastic Frontier Analysis (SFA) work, how they are estimated, and which advantages and disadvantages they offer
    • How to setup meaningful benchmarking models with DEA and SFA, which data is required, and how can results be interpreted

    Skills

    • Ability to carry out scientific benchmarking DEA and SFA using the statistical software R
    • Good practices in efficiency analysis including, e.g., descriptive data analysis, implementation of outlier detection methods, and the meaningful visualization of results

  • Teaching

    This course combines lectures, exercises, and programming tutorial. Lectures focus on methodological issues. Pen and paper exercises help to deepen the knowledge obtained in the lectures with illustrative examples. In programming tutorials, the student will implement the learned methodologies with real data sets using the free software R.

  • Recommended prerequisites

    A basic understanding of microeconomic concepts such as production and cost functions is advised. First experience with linear programming and econometric estimation techniques is advantageous.

    Programming skills with R are helpful. However, if necessary, an additional introduction to programming with R will be offered.

  • Credit reduction due to overlap

    Course identical to BUS462.

  • Requirements for course approval

    A group project (3 to 4 group members), with a presentation during the course

    Course approval from BUS462 is valid for BAN423.

  • Assessment

    Group term paper, based on the group project

    This course is a continuation of BUS462 and the total number of attempts applies to the course (not the course code).

  • Grading Scale

    Pass/fail

  • Computer tools

    R (https://cran.r-project.org/) with the packages Benchmarking and frontier.

  • Literature

    Advised:

    Bogetoft, P. and Otto, L. (2011) Benchmarking with DEA, SFA, and R, Springer.

    Further reading:

    Coelli, T., Rao, D., O'Donnell, C. and Battese, G. (2005) An Introduction to Efficiency and Productivity Analysis, 2nd edition, Springer.

    Färe, R., Grosskopf, S. and Lovell, C.A.K. (1994) Production Frontiers, Cambridge University Press, Cambridge.

    Greene, W. H. (2007) The Econometric Approach to Efficiency Measurement. In: Fried, H., Lovell, C.A.K. and Schmidt, S., editors, The Measurement of Productive Efficiency. Oxford University Press, Oxford.

Overview

ECTS Credits
2.5
Teaching language
 English
Semester

Autumn. Offered Autumn 2018

Course responsible

 Dr. Stefan Seifert, University of Bonn