Abstract
The cost Malmquist productivity index (CMPI) has been proposed to capture the performance change of cost minimizing Decision Making Units (DMUs). Recently, two alternative uses of the CMPI have been suggested: (1) using the CMPI to compare groups of DMUs, and (2) using the CMPI to compare DMUs for each output separately. In this paper, we propose a new CMPI that combines both procedures. The resulting methodology provides group-specific indexes for each output separately, and therefore offers the option to identify the sources of cost performance change. We also define our index when input prices are not observed and establish, in that case, a duality with a new technical productivity index, which takes the form of a Malmquist productivity index. We illustrate our new methodology with a numerical example and an application to the US electricity plant districts.
Similar content being viewed by others
Notes
Initially the MPI was interpreted as a productivity change index. Recently, there has been a debate on the validity of the MPI to measure productivity change. Indeed, this is only true under quite restrictive conditions. We refer to O’Donnell (2012) and Peyrache (2014) who highlight this issue. Therefore, in the following, we interpret the MPI as a relative performance index (this is also the interpretation used by Grosskopf (2003))
See Cherchye et al. (2015) for a formal definition of those axioms in a similar context. Note that imposing monotonicity and convexity does not alter the cost evaluation (see, for example, Varian (1984) and Tulkens (1993) for discussion). We impose these extra technology axioms since they are required to establish the duality of our CMPI with the technical counterpart (i.e. the MPI). Finally, note that the CMPI could be defined under alternative returns-to-scale assumptions.
These output-specific prices have a similar interpretation as Lindahl prices for public goods. See Cherchye et al. (2008) for more details.
An alternative would be to recover the allocation of inputs to outputs. See our discussion in Section 3 (practical implementation).
For fuel, the EPA provides prices at the state level. For the nameplate capacity, we can proxy the price by transforming the electricity price (available at the state level too) since nameplate capacity is defined as the maximal electricity generated by the plants during one year.
PADD I (East Coast): Connecticut, Delaware, District of Columbia, Florida, Georgia, Maine, Massachusetts, Maryland, New Hampshire, New Jersey, New York, North Carolina, Pennsylvania, Rhode Island, South Carolina, Virginia, Vermont, and West Virginia. PADD II (Midwest): Illinois, Indiana, Iowa, Kansas, Kentucky, Michigan, Minnesota, Missouri, Nebraska, North Dakota, South Dakota, Ohio, Oklahoma, Tennessee, and Wisconsin. PADD III (Gulf Coast): Alabama, Arkansas, Louisiana, Mississippi, New Mexico, and Texas. PADD IV (Rocky Mountain): Colorado, Idaho, Montana, Utah, and Wyoming. PADD V (West Coast): Alaska, Arizona, California, Hawaii, Nevada, Oregon, and Washington.
Note that if data for the number of employees are available, we could allocate the employees for each electricity generation process.
References
Camanho AS, Dyson RG (2006) Data envelopment analysis and Malmquist indices for measuring group performance. J Product Anal 26:35–49
Caves DW, Christensen LR, Diewert WE (1982) The economic theory of index numbers and the measurement of input, output and productivity. Econometrica 50:1393–1414
Charnes A, Cooper WW (1962) Programming with linear fractional functionals. Nav Res Logist Q 2:429–444
Charnes A, Cooper WW (1978) Measuring the efficiency of decision making units. Eur J Oper Res 2:429–444
Chen Y (2003) A non-radial Malmquist productivity index with an illustrative application to Chinese major industries. Int J Prod Econ 83:27–35
Chen Y, Ali AI (2003) DEA Malmquist productivity measure: new insights with an application to computer industry. Eur J Oper Res 159:239–249
Cherchye L, De Rock B, Dierynck B, Roodhooft F, Sabbe J (2013) Opening the black box of efficiency measurement: input allocation in multi-output settings. Oper Res 61:1148–1165
Cherchye L, De Rock B, Vermeulen F (2008) Cost-efficient production behavior under economies of scope: a nonparametric methodology. Oper Res 56:204–221
Cherchye L, De Rock B, Walheer B (2015) Multi-output efficiency with good and bad outputs. Eur J Oper Res 240:872–881
Cherchye L, De Rock B, Walheer B (2016) Multi-output profit efficiency and directional distance functions. Omega (West) 61(C):100–109
Cook WD, Seiford LM (2009) Data Envelopment Analysis (DEA) - Thirty years on. Eur J Oper Res 192:1–17
Cooper WW, Seiford LM, Tone K (2007) Data envelopment analysis: a comprehensive text with models, applications, references and DEA-Solver Software, 2nd edn. Springer, New York
Cooper WW, Seiford LM, Zhu J (2004) Handbook on data envelopment analysis, 2nd edn. Springer, New York
Debreu G (1951) The coefficient of resource utilization. Econometrica 19(3):273–292
Despic O, Despic M, Paradi J (2007) DEA-R: ratio-based comparative efficiency model, its mathematical relation to DEA and its use in applications. J Product Anal 28:33–44
Färe R, Grosskopf S (2000) Network DEA. Socioecon Plann Sci 34:35–49
Färe R, Grosskopf S, Lovell CAK (1994a) Production frontier. Cambridge University Press, Cambridge
Färe R, Grosskopf S, Norris M (1994b) Productivity growth, technical progress and efficiency change in industrialized countries. Am Econ Rev 84:66–83
Färe R, Grosskopf S, Norris M (1997) Productivity growth, technical progress and efficiency change in industrialized countries: reply. Am Econ Rev 87:1040–1043
Färe R, Grosskopf S, Whittaker G (2007) Network DEA. In: Zhu J, Cook W (eds) Modeling data irregularities and structural complexities in data envelopment analysis, Springer, New York
Farrell M (1957) The measurement of productive efficiency. J R Stat Soc 120:253–281
Fried H, Lovell CAK, Schmidt S (2008) The measurement of productive efficiency and productivity change. Oxford University Press, Oxford
Grosskopf S (2003) Some remarks on productivity and its decompositions. J Product Anal 20(3):459–474
Huang M-Y, Juo J-C (2015) Metafrontier cost Malmquist productivity index: an application to Taiwanese and Chinese commercial banks. J Product Anal 44:321–335
Kao C (2010) Malmquist productivity index based on common-weights DEA: the case of Taiwan forests after reorganization. Omega 38:484–491
Kao C, Hwang S-N (2014) Multi-period efficiency and Malmquist productivity index in two-stage production systems. Eur J Oper Res 232:512–521
Malmquist S (1953) Index numbers and indifference surfaces. Trab De Estat 4:209–242
Maniadakis N, Thanassoulis E (2004) A cost Malmquist productivity index. Eur J Oper Res 154:396–409
Mayer A, Zelenyuk V (2014) Aggregation of Malmquist productivity indexes allowing for reallocation of resources. Eur J Oper Res 238:774–785
O’Donnell CJ (2012) An aggregate quantity framework for measuring and decomposing productivity change. J Product Anal 38:255–272
Oh DH, Lee J-D (2010) A metafrontier approach for measuring Malmquist productivity index. Empir Econ 38:47–64
Pastor JT, Asmild M, Lovell CAK (2011) The biennial Malmquist productivity change index. Socioecon Plann Sci 45:10–15
Pastor JT, Lovell CAK (2005) A global Malmquist productivity index. Econ Lett 88:266–271
Peyrache A (2014) Hicks-Moorsteen versus Malmquist: a connection by means of a radial productivity index. J Product Anal 41(3):435–442
Portela MCAS, Thanassoulis E (2010) Malmquist indices for measuring productivity in the presence of negative data: an application to bank branches. J Bank Financ 34:1472–1483
Ray S, Delsi E (1997) Productivity growth, technical progress and efficiency change in industrialized countries: Comment. Am Econ Rev 87:1033–1039
Salerian J, Chan C (2005) Restricting multiple-output multiple-input DEA models by disaggregating the output-input vector. J Product Anal 24:5–29
Sarkis J, Cordeiro JJ (2012) Ecological modernization in the electrical utility industry: an application of a bads-goods DEA model of ecological and technical efficiency. Eur J Oper Res 219:386–395
Shepard RW (1953) Cost and production functions. Princeton University Press, Princeton
Shepard RW (1970) Theory of cost and production functions. Princeton University Press, Princeton
Thanassoulis E, Shiraz RK, Maniadakis N (2015) A cost Malmquist productivity index capturing group performance. Eur J Oper Res 241:796–805
Tohidi G, Razavyan S, Tohidnia S (2012) A global cost Malmquist productivity index using data envelopment analysis. J Oper Res Soc 63:72–78
Tone K, Tsutsui M (2009) Network DEA: a slacks-based measure approach. Eur J Oper Res 197:243–252
Tone K, Tsutsui M (2011) Applying an efficiency measure of desirable and undesirable outputs in DEA to U.S. electric utilities approach. J Cent Cathedra 4:236–249
Tulkens H (1993) On FDH analysis: some methodological issues and applications to retail banking, courts and urban transit. J Product Anal 4:183–210
Varian HR (1984) The non-parametric approach to production analysis. Econometrica 52:579–598
Walheer B (2016a) A multi-sector nonparametric production-frontier analysis of the economic growth and the convergence of the European countries. Pac Econ Rev 21(4):498–524
Walheer B (2016b) Growth and Convergence of the OECD countries: a multi-sector production-frontier approach. Eur J Oper Res 252:665–675
Walheer B (2017) Disaggregation of the Cost Malmquist Productivity Index with joint and output-specific inputs. Omega, accepted 75:1–12
Wang Y-M, Lan Y-X (2011) Measuring Malmquist productivity index: a new approach based on double frontiers data envelopment analysis. Math Comput Model 54:2760–2771
Yang B, Youliang Z, Zhang H, Zhang R, Xu B (2016) Factor-specific Malmquist productivity index based on common weights DEA. Operation Res, forthcoming. 16(1):51–70
Yang YL, Huang CJ (2009) Estimating the Malmquist productivity index in the Taiwanese banking industry: a production and cost approach. Taiwan Econ Rev 37:353–378
Yu MM (2007) The capacity productivity change and the variable input productivity change: a new decomposition of the Malmquist productivity index. Appl Math Comput 185:375–381
Zelenyuk V (2006) Aggregation of Malmquist productivity indexes. Eur J Oper Res 174:1076–1086
Acknowledgements
We thank the Editor Victor Podinovski, the Associate Editor, and the two anonymous referees for their valuable comments that substantially improved the paper. We also thank participants of the 2017 CEPA International Workshop on Performance Analysis in Brisbane for useful discussion.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The author declares that he has no competing interests.
Appendices
Appendix 1
Table 5
In this Appendix, we show how the linear programs work for the illustrative example. In particular, we distinguish between two cases: when the input prices are observed and when they are not. Note that the output-specific input prices are not observed in this example.
1.1 Practical implementation when the input prices are observed
Step 1: Solve (LP-2) for each member of group A taking group A as the reference group for the technology.
For example for DMU 1 in group A, it is given as:
Note that the denominator is the actual cost of DMU 1 in group A: \({\mathbf{w}}_1^{A{\prime}}{\mathbf{x}}_1^A\) = 6*3 + 5*2 = 18 + 10 = 28.
Step 2: Solve (LP-2) for each member of group A taking group B as the reference group for the technology.
For example for DMU 1 in group A, it is given as:
Step 3: Solve(LP-2) for each member of group B taking group B as the reference group for the technology.
For example for DMU 1 in group B, it is given as:
Note that the denominator is the actual cost of DMU 1 in group B: \({\mathbf{w}}_1^{B{\prime}}{\mathbf{x}}_1^B\) = 4*4 + 3*4.5 = 16 + 13.5 = 29.5.
Step 4: Solve (LP-2) for each member of group B taking group A as the reference group for the technology.
For example for DMU 1 in group B, it is given as:
Step 5: Compute CMPI for the overall output and CMPI 1 and CMPI 2 for the two individual outputs by plugging-in the computed cost efficiency scores in (25), (26), and (27) for CMPI, and in (18), (19), and (20) for CMPI 1 and CMPI 2. As a final remark, note that tout output-specific minimal costs are also obtained when solving the linear programs, as explained in detail in Section 3.
1.2 Practical implementation when the input prices are not observed
Step 1: Solve (LP-4) for each member of group A taking group A as the reference group for the technology.
For example for DMU 1 in group A, it is given as:
Step 2: Solve (LP-4) for each member of group A taking group B as the reference group for the technology.
For example for DMU 1 in group A, it is given as:
Step 3: Solve (LP-4) for each members of group B taking group B as the reference group for the technology.
For example for DMU 1 in group B, it is given as:
Step 4: Solve (LP-4) for each member of group B taking group A as the reference group for the technology.
For example for DMU 1 in group B, it is given as:
Step 5: Compute \(\widehat {CMPI}\) for the overall output and \(\widehat {CMPI}^1\) and \(\widehat {CMPI}^2\) for the two individual outputs by plugging-in the computed cost efficiency scores in (32) and (33) or (34). Note that similar results could be obtained using (LP-3). We provide here the technical formulation in an illustrative purpose.
Appendix 2
Tables 6–
Rights and permissions
About this article
Cite this article
Walheer, B. Cost Malmquist productivity index: an output-specific approach for group comparison. J Prod Anal 49, 79–94 (2018). https://doi.org/10.1007/s11123-017-0523-5
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11123-017-0523-5