DEA and benchmarking: DataEnvelopmentAnalysis (DEA) has been recognized as an excellent method for analyzing performance and modeling organizations and operational processes, particularly when market prices are unavailable. Unlike the statistical regression method that tries to fit a regression plane through the center of thedata, DEA floats a piecewise linear surface to rest on top of thedata by linear programming techniques (Cooper et al., 2007). DEA, on the other hand, produces an efficient frontier consisting of the set of most efficient performers, allowing a direct comparison to the best performers. But, the statistical regression method estimates the parameters inthe assumed functional form by a single optimization over all Decision Making Units (DMUs) whereas DEA uses optimizations for different DMUs without a priori assumptions on the underlying functional forms. Because of this unique feature, DEA has been applied to various areas of efficiency evaluation. In DEA, the ratio of weighted outputs and inputs produces a single measure of productivity called relative efficiency. Let there be n DMUs whose efficiencies have to be compared. Let us take one of the DMUs, say the kth DMU and maximize its efficiency according to the formula given following:
This paper focuses the issue of efficiency inthe Brazilian motor carrier industry using both DEA (DataEnvelopmentAnalysis) and SFA (Stochastic Frontier Analysis), the two dominant approaches to modern benchmarking (BOGETOFT; OTTO, 2010). Moving from DEA to SFA, there are some main distinguishing features. In terms of methods, the DEA approach has its roots in mathematical programming, whereas the SFA approach is much more directly linked to econometric theory. Given that, while the slack analysis of DEA provides insight for increasing or reducing input resources to improve efficiency scores, the SFA method focuses on the economic justification of a given production function and subjecting it to further hypothesis testing (LIN; TSENG, 2005). Besides, SFA is a parametric approach, thus implying some advantages and disadvantages over DEA, mainly related to the assumption of a stochastic relationship between the inputs used and the output produced.
commercial banks’ performance. For instance, Casu and Molyneux (2000) employed the DEA approach to investigate the efficiency in European banking systems. They attempted to examine whether the productive efficiency of European banking systems has improved and converged towards a common European frontier, following the process of EU legislative harmonization. Noulas (2001) studied the effect of banking deregulation on private and public-owned banks by usingDataEnvelopmentAnalysis. The results showed that the private banks were more efficient than the public-owned, although the gap between levels of efficiency is not relevant from a statistical viewpoint. Barr et al. (2002) evaluated the productive efficiency of U.S. commercial banks. Study results revealed a close interdependence between efficiency and independent measures of performance, including confidential ratings made by bank examiners. Jemric (2002) investigated the efficiency of banks in Croatia. The main results showed that foreign banks are, on average, the most efficient; also banks that recently entered the market are more efficient than those operating for a long time. Also, small banks are more efficient than large ones. Wu (2005) examined productivity and efficiency of banks in Australia during 1983-2001. The main results reported that efficiency increased in times of deregulation. Loukoianova (2008) made a comparison of the banking sectors in Western Europe, the U.S. and Japan depending on the specialization of banks.
previous work, Kittelsen and Førsund [33] did not acknowledge explicitly in their model that the efficiency of the courts might be influenced by exogenous factors as they have only considered the workforce of courts (number of posts as judges and number of office staff) as an input. However, they did acknowledge that the cases judged by the courts are neither homogenous in relative labour intensities nor in total time required. Therefore, the outputs of their model followed the conventional functional division of cases used inthe judiciary service. To this purpose, seven major categories of cases were considered: Civil cases; B-cases; Examination and summary jurisdiction cases; Ordinary criminal cases; Registry cases; Cases of duress; Probate and bankruptcy cases. Based on this model, their analysis found a fairly efficient court sector, with the average technical efficiency for all courts above 90%. Furthermore, their study found that most of the inefficiency was scale inefficiency with inefficient courts, on average, being smaller than optimal. Contrary to the work of Lewin et al. [32], they also performed a dynamic analysis by assessing productivity changes inthe sector. Usingthe Malmquist Productivity Index (MPI), they found that the improvement for the six year period was rather weak at about 6%, with 4% due to catching up and 2% due to technology shift.
Drake (2001) analyses relative efficiencies within the banking sector and the productivity change inthe main UK banks over the period 1984 to 1995. The results obtained provide important insights into the size-efficiency relationship inthe considered sample of banks and offer a perspective on the evolving structure and competitive environment within which the banks are currently operating. Webb (2003) utilizes DEA window analysis, to measure the relative efficiency levels of large UK retail banks during the period 1982-1995, mostly finding that the overall long run average efficiency trend is falling, and also that all banks inthe study show reducing levels of efficiency over the entire time period.
and output variables, we used Delhi method based on opinion of 30 experts in health networks. Data were analyzed usingdataenvelopmentanalysis technique. In this method, technical, scale, and managerial efficiency are calculated based on input and output variables. We selected input-orientation approach and variable returns to scale (VRS) model for dataanalysis. In input-orientation approach, decision-making units (DMUs) can change their inputs (4, 5, 6). When the efficiency measurement is based on the input-orientation approach (minimizing the production factors), a value of one show perfect efficiency. In this study, we efficiency range was included: poor efficiency (0.5>), moderate efficiency (0.51- 0.8), and good efficiency (0.81-1). Finally, Data was analyzed by SPSS.18 and DEAP.2 software.
In this paper we systematically compare the output from the health system of a set of OECD countries with resources employed (doctors, nurses, beds and diagnostic technology equipment). Usingdataenvelopmentanalysis (DEA), we derive a theoretical production frontier for health. Inthe most favourable case, a country is operating on the frontier, and is considered as efficient. However, most countries are found to perform below the frontier and an estimate of the distance each country is from that borderline is provided – the so-called efficiency score. Moreover, estimating a semi-parametric model of the health production process using a two-stage approach, we show that inefficiency inthe health sector is strongly related to variables that are, at least inthe short- to medium run, beyond the control of governments. These are GDP per capita, the education level, and unhealthy lifestyles as obesity and smoking habits.
In recent years, DEA has been used in agricultural enterprises: In an earlier study, Fraser and Cordina [14] applied DEA to evaluate the technical efficiency of input use for irrigated dairy farms in Australia. They reported that DEA was a useful tool in helping to benchmark the dairy industry, which is continually striving to improve the productive efficiency of farms. Subsequently, DEA was used to investigate the efficiency of individual farmers and to identify the efficient ones in citrus production in Spain [15]. In another study [16] the technique was applied to benchmark the productive efficiency of irrigated wheat area in Pakistan and India based on three inputs of irrigation, seed and fertilizer. Nassiri and Singh [17] applied the DEA technique to thedata of energy use for paddy production in India. They assumed energy equivalents of different inputs as input variables and the paddy yield as output variable. Finally, Omid et al. [18] employed this technique to analyze the technical and scale efficiencies of greenhouse cucumber producers in Iran.
After being generated, the electricity is transported via high- voltage lines. This transport of energy features the transmission segment, which encompasses a range from 230 kV and 750 kV. According to ABRADEE (2015), there are 77 utilities to perform the energy transport and together they manage more than one hundred thousand kilometers of lines scattered throughout Brazil. The power transmission occurs through national interconnected system and is essential to ensure the continuous supply of electricity inthe country, since it allows seasonal and regional complementation. Transmission segment transports electrical energy to the distribution sector, where the voltage is lowered to <230 kV, in order to make the connection to final users. Distribution segment has 64 power distributors firms, being 60% private and 40% public, and achieved approximately 77 million “consumer units” in 2015, of which 85% are residential (ABRADEE, 2015). Inthe Brazilian electricity market, the purchase and sale negotiations occur in two environments: Regulated contracting environment (RCE) and free contracting environment (FCE). In RCE, distribution companies buy electrical energy from sellers in public auctions under set prices. On the other hand, in FCE, distributors are free to negotiate their own bilateral contracts with their suppliers outside auctions. In order to deal with those environments, the chamber of electricity trading (CCEE) was created (Araújo et al., 2007; Melo et al., 2011).
Figure 6 presents our candidate DEA models: a sustainability model (a), and an efficiency model (b). The sustainability model (a) measures the efficiency of a country (or subnational entity or company) for generating construction GVA and maximizing the CDW Recovery Rate (CDWRR) while at the same time reducing Non-Hazardous Mineral CDW. It is an input-oriented model that, for a given construction GVA and CDWRR, looks for minimizing Non-Hazardous Mineral (NHM) CDW. There is no input inthe model, as CDW is an undesirable output treated as an input. The reason is that, so far, all the inputs considered in related work, such as labor force and gross capital formation [21], add little value to the model. Alternative inputs, such as national GVA, buildings’ construction year distribution and distribution of population have been tested by us, but these inputs either add little value (GVA is highly correlated with construction GVA) or consider efficient a significantly higher number of countries. In a sector where data quality might be modest/poor and outliers are frequent, the usage of additional variables not directly related to the efficiency metric sought (such as using labor salary for estimating CDW production) causes the efficiency metric to lose discriminatory power. Finally, it is expected that the returns of the model will be varying in scale (thus VRS will be used), and different construction turnover and CDWRR might show different efficiencies when reducing the input.
The Portuguese Post Offices have suffered, since their inception in 1520, profound changes in their structure and inthe services provided to the population. Anyone who visits today the company CTT Correios de Portugal, SA, whether visiting a post office, a postal distribution center or a post treatment center will certainly be surprised not only with all the technology that supports internal operations, but also with the professionalism and proactive attitude of the employees of this company. All this evolution perceived by customers is the result of five centuries of history. The aim of this study is to explore the potential of DataEnvelopmentAnalysis (DEA) to assess the efficiency of the post offices and postal distribution centers (PDCs) inthe south of Portugal. To this effect, we collected data from 84 post offices and 42 PDCs. Our results show significant differences among efficiency scores in both groups and emphasize the importance of identifying efficient units. These efficient units can serve as benchmarks for learning, revealing the type of structures and processes that can be applied in other units in order to make them efficient and sustainable. Our results also show the utility of DEA as a tool to support decision-making in this company, as this technique can assist managers inthe identification of the units that have the greatest potential to improve their performance. Furthermore, the fact that DEA allows the decomposition of efficiency in two components (pure technical efficiency and scale efficiency) is very useful in order to identify the type of restructuring that can be most efficacious in each unit. Lastly, a preliminary analysis of the impact of seasonality inthe efficiency of the units revealed that this can be one of the factors that contribute to explaining variations in performance in some of the units. This result suggests that, in order to remain efficient, some units may need to adjust their capacity according to the season.
the context of HIV prevention may result in substantial inaccuracies as they found compelling evidence that efficiency increased (unit costs decreased) with scale, across all countries and interventions examined. Brandeau et al. [5] and Brandeau and Zaric [46], also discuss the issue of scale in HIV prevention programmes, pointing out that the relationship between investment in HIV prevention and HIV infections averted may not be linear, which indicates that increased spending on a prevention programme may not always be cost effective. Our choice of a VRS assumption is also consistent with the publication of Hollingsworth and Smith [47], which warns that when ratios are used, as is the case of the variable political stability, the BCC formulation [18] should be adopted. We have also used a weight restriction on the input side, in order to prevent countries from attributing a null weight to the variable related to spending on PMTCT. This restriction imposes that the sum of the virtual weights given by each country to the non- controllable inputs cannot exceed the virtual weight given to the controllable input. The development of this virtual weight restriction follows the approach proposed by Sarrico and Dyson [48]. Please refer to the appendix for more details on the DEA model used.
Abstract : An efficiency analysis of the commercial dredge fleet operating along the South coast of Portugal between 2005 and 2007 sought to determine the efficiency of the vessels usingdataenvelopmentanalysis models, considering fixed inputs (vessel power, length, tonnage, and an indicator of stock biomass) and a variable input (number of days at sea). The annual quota per vessel was also included inthe model as a contextual factor. Inthe technical-efficiency (TE) analysis, outputs were defined by the catch weight for each of the three target species (bivalves). Using price data for each species inthe wholesale market, revenue efficiency was also estimated to complement the TE analysis. The advantage of the approach lies inthe ability to separate technical aspects from allocative aspects inthe efficiency assessment, allowing two-dimensional graphic representation of vessel performance. The procedure allows the identification of benchmark vessels, which maximized the catch weight of the species landed, given their inputs, as well as the vessels that selected the appropriate target species to maximize the revenue of the fishing activity, given output prices. The approach also allowed the specification of targets for inefficient vessels that correspond to the catch by species, permitting revenue maximization from fishing.
Farrell's empirical work had been confined to single-output cases when Charnes et al. (1978) extended his work and defined a linear program in order to obtain efficiency values, for a production system characterized by multiple inputs and multiple outputs. They also introduced an innovation to Farrell’s measure of efficiency. They introduced slack measures to account for non-radial adjustments to the frontier. According to the production metaphor used by DEA, each homogeneous DMU is engaged in a transformation process, where by using some inputs (resources) it is tries to produce some outputs (goods and services). DEA uses observed inputs and outputs in order to construct the production frontier. The relative efficiency is defined as the ratio of the total weighted output to the total weighted input (the ratio ranges from zero to one). A DMU is considered relatively efficient if it achieves a score of one. The technique allows each unit to choose the optimal weight structure in order to maximize the ratio. In addition, DEA allows each unit to identify a benchmarking group, in other words, a group of units that are following the same objectives and priorities, but performing better. There are two main advantages in adopting this technique, first it can handle multiple inputs and multiple outputs, and second it does not require a precise knowledge of the form of the production function.
Wanke and Barros (2014) measured efficiency in Brazilian banking using a two-stage process where inthe first stage, the number of branches and employees were used to attain a certain level of administrative and personnel expenses per year. Inthe productive efficiency stage, these expenses permitted the consecution of two important net outputs including equity and permanent assets. They applied the network-DEA centralized efficiency model to optimize both stages, simultaneously. They reported that Brazilian banks were heterogeneous, with some concentrating on cost efficiency and others on productive efficiency. In addition, cost efficiency was described by marketing and administration (M&A) as well as size, while productive efficiency was described by M&A and public status. Liu et al. (2009) applied DEA technique to measure the relative efficiencies on a bank in Taiwan and studied the performance and productivity changes when banks implement financial electronic data interchange. They included 18 branches of the performance for implementation of financial electronic data interchange of the overall efficiency, pure technical efficiency, scale efficiency, analysis of reference groups and the potential to improve the value of analysis for different branch performance assessments. The empirical results shown that case bank could adopt the DEA evaluation model as references to upgrade the overall operating performance effectively for creating competitive advantages. Wang et al. (2014) utilized network DEA method to evaluate efficiencies of the Chinese commercial banks.
With regards to the system design, a good example of the flexibility involved is the decision to install or not install Oil Circuit Recloser (OCR) equipment inthe overhead distribution lines. This equipment will interrupt the flow of power if there is a fault or short on the power lines, isolating the impact of power interruption to a smaller part of the power line, until the problem has been fixed and electricity distribution can be restored. Although a significant investment, it is considered that this piece of equipment can minimise the effects of interruptions, because, according to some studies (see for example, Weedy 1972: 26), around 90% of faults on overhead power lines are transient and can be cured by autoreclosing. EDPD has invested significantly in this technology, spending 4.4 million Euros in 2009, 8 million Euros in 2010 and 9.5 million Euros in 2011 (EDP – PDIRD 2009-11), and still plans to increase its investment. It is expected that the installation of OCR equipment may contribute to improving the quality of the service by lowering the duration and spread of power interruptions. However, once installed, this type of equipment also requires regular maintenance, whose costs should be taken into account. It is therefore of major interest to investigate the impact of this technology on the efficiency of lines.
This paper follows this latter strand of literature and tests banking efficiency across EU countries inthe wake of the recent crisis, using both Stochastic Frontier Analysis (SFA) and DataEnvelopmentAnalysis (DEA) estimates and comparing the results obtained for a panel comprising the “old” EU-15 countries (Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden and UK) and another panel comprising all of the current EU-27 members. The main conclusions point to the existence of statistically important technical inefficiencies that increased slightly after 2004 with the inclusion of the 12 new member-states. The obtained country efficiency rankings also allow us to conclude that countries that performed well inthe EU-15 panel maintain their strong positions inthe enlarged EU-27 panel. Furthermore, theanalysis of the convergence process with the estimation of a beta-convergence model clearly shows that while there is convergence in banking efficiency across EU countries, it is a very slow process and not only the new member-states but also some of the “old” EU countries are still facing difficulties in adapting to the new market conditions.
34 treatment success rates is missing for numerous countries for the year of 2013 and those who have information, sometimes it does not comprehend all types of TB or patients’ categories. Moreover, information available for most of the countries, covers just a single year or, when it covers more than one year, it is not for consecutive years. Because of this, we were unable to perform a dynamic analysis of the performance of countries but we believe that the present DEA model allowed some relevant findings. We also advise some caution inthe interpretation of the results for Guatemala, Haiti, Kenya, Lesotho, Namibia, Nigeria and Peru (countries marked with the sign * in Tables 6 and 8) as for these countries output 1 does not include the relapsed cases. These cases are accounted in output 2. Considering that none of these countries was considered effective, that means that their TB treatment success rates did not impact the effectiveness scores of the other countries. However, considering that the prognosis of treatment for new and relapsed cases is better that for previously treated TB cases, excluding relapses, it is likely that the effectiveness scores for these countries are overestimated.
The objective of this study was to measure the technical and scale efficiencies of milk-producing farms inthe state of Minas Gerais, considering different production levels, and also to identify the determining factors of their technical efficiency. The analyses were carried out using both DataEnvelopmentAnalysis (DEA) and an econometric Tobit model. Thedata consisted of information collected in 2005 relating to 771 milk-producing farms. The results indicated that most of the farms exhibit technical inefficiency problems. Small farmers have the potential to expand their production and productivity, increasing technical efficiency, since they were performing with increasing returns to scale. The large farmers presented the best measures of technical ef- ficiency, which is explained, partly, by factors such as access to rural credit, training and technical support.
Energy inefficiencies inthe Ugandan SME foundry units is attributed to the technology employed, poor operations and maintenance practices, and the poor quality scrap inputs. There is need for improvement of technology, which could give better yields and energy efficiencies. Foundry employees should get acquaintance with better operation and management practices which embrace efficient energy management. Comparing the inputs and outputs of the foundries understudy usingthe DEA method indicate that the energyimprovement potentialfor the SME metal casting units in Ugandais about 42%.Various energy conservation measures need to be applied in order to reduce energy use, as indicated inthe previous study [18].The model also highlights DMU1 (foundry A) and DMU4 (foundry D) as the foundries that can be emulated by foundry E. This study could be extended to other energy intensive industries like the processing industries and also benchmark energy conservation practices used inthe developed countries to enable the Ugandan industries increase their productivity and competitiveness.