Abstract
In this paper we explain the discrepancy between the efficiency of the average production unit and average efficiency by means of a covariance term relating size and technical efficiency and the extent of reallocation inefficiency. We show that, regardless of the degree of reallocation efficiency, the efficiency of the average production unit is lower than average efficiency when either size and efficiency are uncorrelated or larger units are more efficient than smaller ones. If however larger units are less efficient than smaller ones, the efficiency of the average production unit is higher than average efficiency as long as reallocation efficiency prevails. In addition, reallocation efficiency and a zero covariance term are sufficient conditions for the efficiency of the average production unit to be equal to average efficiency.
Similar content being viewed by others
Notes
In this it is implicitly assumed that the reallocation process involves changes in the internal margin only (i.e., input movements across and mergers/splits of constituent firms), excluding entry and exit of firms. Another approach for aggregating across firms that follows points (a) and (b) is Johansen (1972) vintage approach; for more see Färe and Grosskopf (2004, pp. 109–115).
We use Li and Cheng (2007) terminology for structural and aggregate efficiency, which follows closer Farrell’s (1957) spirit. His notion of structural efficiency clearly reflects reallocation of inputs available to the industry as well as restructuring of firms (see p. 262) while aggregate efficiency, which he called it ‘efficiency of an industry’, takes industry structure as given (see p. 261). Some authors (i.e., Bjurek et al. 1992; Ylvinger 2000; Nesterenko and Zelenyuk 2007) used the term structural efficiency to mean what we refer to as aggregate efficiency. One should keep that in mind when reading Ylvinger’s (2000) criticism on using the average unit to measure industry efficiency. On the other hand, Färe et al. (1992) estimated structural and aggregate efficiency by using what they referred respectively to as the industry and the firm linear programming models.
In a multi-output setting, allocation inefficiency does not matter in the aggregation of technical inefficiency scores as long as each firm is as allocatively inefficient as the industry (Färe and Grosskopf 2004, p. 116). Even in this case though we cannot avoid using price dependent weights.
The same is true in the case of input orientation as long as production technology employs a single input and thus can be represented by an input requirement function.
Throughout the paper when a quotation is used, words or phrases in brackets are our own additions.
Here we follow Färe et al. (1985) tradition and we define output-oriented technical efficiency to be equal or greater than one. Notice however that Farrell (1957) defined all his efficiency measures to be between zero and one, i.e., the inverse of (1). Our main results hold for Farrell (1957) efficiency measures as well after the appropriate adjustments to aggregation weights in (5) are made by means of the denominator rule, which requires (see Färe and Karagiannis 2013) the weighting shares to be defined in terms of denominator variable of the relevant index in order to ensure consistent aggregation of ratio-type performance measures.
Färe et al. (1992) used a single output, constant-returns-to-scale model and thus can aggregate firms’ output just by inserting the total amount of factors. This cannot be done without the assumption of constant returns to scale as mentioned by a referee. No such an assumption is made in (2) below or in Eq. (4) of Li and Ng (1995) that provides the efficiency measurement counterpart of Eq. (3) in (Färe et al. 1992).
Notice that Farrell (1957) considered a single-output technology, an input-oriented measure of efficiency and implicit in his definition are the assumptions of identical firm technology and constant returns to scale. Färe and Karagiannis (2014) show that his suggestion for the aggregation weights could have been correct if he had referred to potential (instead of observed) output shares. In that case the aggregate efficiency measure would reflect potential input savings.
In the case of multi-output technologies and allocative inefficiency, price dependent weights (i.e., revenue shares) should be used for consistent aggregation of technical efficiency scores.
Since the following aggregation scheme uses observed output shares as weights, which is the denominator of the \( F_{O}^{k} \left( {x^{k} ,y^{k} } \right) \) index, it does not violate. monotonicity and moreover, it is consistent with an intuitive interpretation of industry output as the one that can be produced by making all firms work with average technical efficiency (Färe and Karagiannis 2013). For similar concerns on productivity measures see van Biesebroeck (2008).
Bjurek et al. (1992), using the inverse of \( F_{O}^{k} \left( {x^{k} ,y^{k} } \right) \) as their efficiency index, proposed the ratio of industry total observed (actual) to potential output as a measure of aggregate technical efficiency.
Li and Cheng (2007, proposition 2) shown that aggregate technical efficiency for a group or industry with constant-returns-to-scale technology is equal to Ylvinger’s (2000) “structural efficiency” measure with restricted (i.e., common to all firms) weights in aggregating potential and observed (actual) outputs.
ten Raa (2011) referred to reallocation efficiency as industrial organization efficiency.
Nesterenko and Zelenyuk (2007, proposition 2) decomposed reallocation efficiency into a technical and an allocative component, as has been done with other radial economic efficiency measures (i.e., revenue or cost). However their assertion that the allocative component can take values greater or less than one is incompatible with the output-oriented nature of reallocation efficiency.
References
Aigner DJ, Chu SF (1968) On estimating the industry production function. Am Econ Rev 58:826–839
Arocena P, Waddams Price C (2002) Generating efficiency: economic and environmental regulation of public and private electricity generators in spain. Int J Ind Organ 20:41–69
Bjurek H, Kjulin U, Gustavsson B (1992) Efficiency, productivity and determinants of inefficiency at public day care centers in Sweden. Scand J Econ 94(supplement):173–187
Bogetoft P, Otto L (2011) Benchmarking with DEA, SFA, and R. Springer, Berlin
Briec W, Dervaux B, Leleu H (2003) Aggregation of directional distance functions and industrial efficiency. J Econ 79:237–261
Charnes A, Cooper WW, Rhodes E (1981) Evaluating program and managerial efficiency: an application of data envelopment analysis to program follow through. Manage Sci 27:668–697
Du J, Liang L, Chen Y, Bi GB (2010) DEA-based production planning. Omega 38:105–112
Fang L (2013) A generalized DEA model of centralized resource allocation. Eur J Oper Res 228:405–412
Fang L, Li H (2013) A comment on solving the puzzles of structural efficiency. Eur J Oper Res 230:444–446
Färe R, Grosskopf S (2004) New directions: efficiency and productivity. Kluwer Academic Publishers, Dordrecht
Färe R, Karagiannis G (2013) The denominator rule for share-weighting aggregation, unpublished manuscript
Färe R, Karagiannis G (2014) A postscript on aggregate Farrell efficiencies. Eur J Oper Res 233:784–786
Färe R, Zelenyuk V (2003) On aggregate Farrell efficiencies. Eur J Oper Res 146:615–620
Färe R, Grosskopf S, Lovell CAK (1985) The measurement of efficiency of production. Kluwer Academic Press, Dordrecht
Färe R, Grosskopf S, Li SK (1992) Linear programming models for firm and industry performance. Scand J Econ 94:599–608
Farrell MJ (1957) The measurement of productive efficiency. J R Stat S Ser A 120:253–290
Førsund FR, Hjalmarsson L (1974) On the measurement of productive efficiency. Swed J Econ 76:141–154
Førsund FR, Hjalmarsson L (1979) Generalised Farrell measures of efficiency: an application to milk processing in Swedish dairy plants. Econ J 89:294–315
Førsund FR, Kittelsen SAC (1998) Productivity development of Norwegian electricity distribution utilities. Resour Energy Econ 20:207–224
Halsteinli V, Kittelsen SAC, Magnussen J (2001) Scale, Efficiency and organization of norwegian psychiatric outpatient clinics for children. J Ment Health Policy Econ 4:79–90
Hjalmarsson L, Veiderpass A (1992) Efficiency and ownership in Swedish electricity retail distribution. J Prod Anal 3:7–23
Johansen L (1972) Production functions, North-Holland
Kittelsen SAC, Førsund FR (1992) Efficiency analysis of Norwegian district courts. J Prod Anal 3:277–306
Li SK, Cheng YS (2007) Solving the puzzles of structural efficiency. Eur J Oper J 180:713–722
Li SK, Ng YC (1995) Measuring the productive efficiency of a group of firms. Int Adv Econ Res 1:377–390
Lozano S, Villa G (2004) Centralized resource allocation using data envelopment analysis. J Prod Anal 22:143–161
Mar-Molinero C, Prior D, Segovia MM, Portillo F (2014) On centralized resource utilization and its reallocation using DEA. Ann Oper Res 221:273–283
Nesterenko V, Zelenyuk V (2007) Measuring potential gains form reallocation of resources. J Prod Anal 28:107–116
Ng YC, Li SK (2000) Measuring the research performance of Chinese Higher Education Institutions: an application of data envelopment analysis. Educ Econ 8:139–156
Odeck J, Alkadi A (2004) The perfromance of subsidized urban and rural public bus operators: empirical evidence from Norway. Ann Reg Sci 38:413–431
Olley GS, Pakes A (1996) The dynamics of productivity in the telecommunication equipment industry. Econometrica 64:1263–1297
Pachkova EV (2009) Restricted reallocation of resources. Eur J Oper Res 196:1049–1057
Peyrache A (2013) Industry structural inefficiency and potential gains from mergers and break-ups: a comprehensive approach. Eur J Oper Res 230:422–430
ten Raa T (2011) Benchmarking and industry performance. J Prod Anal 36:285–292
Torgersen AM, Førsund FR, Kittelsen SAC (1996) Slack-adjusted efficiency measures and ranking of efficient units. J Prod Anal 7:379–398
Van Biesebroeck J (2008) Aggregating and decomposing productivity. Rev Bus Econ 2:122–146
Ylvinger S (2000) Industry performance and structural efficiency measures: solutions to problems in firm models. Eur J Oper Res 121:164–174
Acknowledgments
This paper was presented at the Conference in Memory of Professor Lennart Hjalmarsson, hold at Gothenburg, Sweden, Dec. 7–8, 2012. I would like to thank session participants for a constructive and stimulating discussion. Rolf Färe as always provided inspired comments and suggestions. Svante Ylvinger supplied several related materials and Finn Førsund gave useful guidelines that both help me in revising the paper. Last but not least, the paper has benefited of valuable comments from two anonymous referees of this journal.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Karagiannis, G. On structural and average technical efficiency. J Prod Anal 43, 259–267 (2015). https://doi.org/10.1007/s11123-015-0439-x
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11123-015-0439-x