A performance benchmarking study of Indian Railway zones
The Authors
Sajeev Abraham George, M.H.S.S. College of Engineering, Mumbai, India
Narayan Rangaraj, Indian Institute of Technology, Mumbai, India
Abstract
Purpose – The paper aims to carry out a performance benchmarking study of the zones of Indian Railways (IR) to develop an alternate approach for measurement of aggregate operational performance of the railway zones and to envisage its operations in a supply chain perspective, so as to gain academic and practical insights.
Design/methodology/approach – A case study research employing data envelopment analysis (DEA) methodology has been used, with the help of data obtained from the IR annual statistical statements published by the Ministry of Railways, Government of India.
Findings – Within the set of inputs and outputs considered, the exercise identified the best performing railway zones over the years and the efficiency trends. Some weaknesses of the conventional DEA were addressed by including the concept of cross-efficiencies along with self-efficiencies, by analyzing longitudinal data spread over four years and also by comparing the efficiencies with the operating ratios. To an extent, this study has also helped to understand the impact of the recent restructuring of the zones on their performance.
Originality/value – The study enables the reader to gain some valuable insights from a managerial perspective for IR so as to formulate strategies of its zones to foster better performance.
Article Type:
Case study
Keyword(s):
Benchmarking; India; Data analysis; Railways; Supply and demand.
Journal:
Benchmarking: An International Journal
Volume:
15
Number:
5
Year:
2008
pp:
599-617
Copyright ©
Emerald Group Publishing Limited
ISSN:
1463-5771
Introduction
Transport infrastructure is the backbone of a nation's economy and has become one of the most critical elements of economic liberalization strategies all over the world (Pangotra and Raghuram, 1999). The primary modes of transportation in India are road and rail. In the case of rail, currently all activities are being handled by one single organization, namely the Indian Railways (IR). Although, IR is a public sector monopoly, there is a growing demand on it to realize the full scope of its assets through proper use of systems and technology and also to address the lack of its customer orientation. Therefore, there is a strong need to devise a performance benchmarking system of its supply chain operations.
Performance measurement plays an increasingly important role in the regulation and management of public sector infrastructural organizations. The main objectives of performance measurement in the public sector, as noted by Radnor and McGuire (2004) are to improve the public services through increased economy and effectiveness in service delivery and to reinforce accountability so that organizations are clearly held to account for the resources they use and the outcomes they achieve. Performance measurement of infrastructural services in the public sector is more complex than that of private sector business organizations. These organizations may need to perform dual roles as a commercial undertaking and as provider of public utility services with a social service obligation. As a result, rather than simply focusing on measures of profitability and other financial measures, the performance measurement system should focus on measures associated with inputs, processes, outputs and outcomes. The three dimensions of performance, as pointed out by Raghuram (2001) namely; the efficiency and effectiveness in asset creation, asset management and service delivery need to be addressed. A service-oriented supply chain perspective would help in better understanding these complexities and performance implications.
Benchmarking has become one of the most popular exercises adopted by organizations to understand how well they are performing relative to their peers. It enables organizations to identify the key processes that need improvement and to search for applicable solutions from the best in class (Lee et al., 2006). As reported by Wynn-Williams (2005), though the concepts and principles of benchmarking in the public sector have received limited attention, they are recognized as having greater potential in this area. McGaughey (2002) classified benchmarking as: internal, external and best practice and stated that internal benchmarking should be the starting point of a company's benchmarking program. There is also a strong demand for effective methods for analyzing benchmarking results, which can be used for design, analysis and improvement of processes (Talluri, 2000). Development of performance benchmarks offers an indirect way of introducing competitive pressure on infrastructure services. This can be used by governments, regulators and others to identify areas where there is potential to improve performance.
Data envelopment analysis (DEA) is a non-parametric mathematical programming technique that allows for the simultaneous evaluation of multiple inputs and multiple outputs to calculate a single comprehensive measure of efficiency. It has become an increasingly popular management tool for performance measurement and benchmarking, since it overcomes many drawbacks of traditional performance measurement systems. An examination of the literature reveals the application of DEA for performance benchmarking in a variety of settings such as banks, public sector enterprises, public utility services, health care units, educational institutes, charity organizations and business firms (Forker and Mendez, 2001; Mukherjee et al., 2002; Min and Park, 2005; Karlaftis, 2004; Verma and Gaverneni, 2006; Duffy et al., 2006; Jong Joo et al., 2007, Dharmapala and Saber, 2007).
This paper reports a study conducted on the supply chain performance benchmarking of IR zones using DEA. The remaining part of the paper is structured as follows. The next two sections describe the background of the study setting and attempt to provide a supply chain perspective of the railway operations. Further, DEA models which are used in this study and subsequently the application of these models for performance benchmarking of IR zones have been presented. The last section contains discussion and conclusions of the study.
Background
IR has been rated as the world's largest railway system under a single management. With over 1.4 million employees, it is the biggest government undertaking and the largest employer among public sector companies of the country. It has an extensive network, which is spread over 63,465 route kilometres, and 7,031 stations. Approximately, 27 percent of the network is electrified (Annual Report and Accounts 2004-05, Indian Railways, Ministry of Railways, 2004). The sheer magnitude, complexity and mix of activities that IR is involved in, make it a difficult organization to manage professionally. It is therefore managed in a multi-tiered manner with the whole of IR being divided into zones, the zones into divisions and the divisions into sections. Prior to 2003, IR was divided into nine broad zones namely: Central, Eastern, Northern, North Eastern, North East Frontier, Southern, South Central, South Eastern and Western.
As part of the initiative to further decentralize the IR operations, given the constraints of geography, regional considerations and other compulsions, these nine zones were restructured in the year 2003 to make the number of zones at present to 16 as: Central, Eastern, East Central, East Coast, Northern, North Central, North Eastern, North East Frontier, North Western, Southern, South Central, South Eastern, South East Central, South Western, Western and West Central.
The seven new zones namely, East Coast, East Central, North Central, North Western, South East Central, South Western and West Central zones were formed by reorganizing some of the divisions from the previously existing nine zones. East Coast was constituted of divisions from South Eastern, East Central from North Eastern and Eastern, North Central from Northern and Central, North Western from Northern and Western, South East Central from South Eastern, South Western from Southern and South Central and West Central from Western and Central zones (Government notification on new zonal railways, Ministry of Railways, 2002).
The entire IR is primarily concerned with the efficient and effective movement of passenger and freight traffic. Its basic objective is to optimize the service levels of its operations within the resource constraints at the same time generating sufficient surplus to finance its own developmental activities, taking into account the complex socio-economic structure of the nation. The IR has played an important role in the social and economic development of the country. Despite financial difficulties, IR has recorded significant growth in route kilometres, rolling stock, signalling and telecommunication, electrification, modernization of diesel and electric traction, large-scale application of IT and employment generation (Agarwal and Makker, 2002). Although, IR is a public sector monopoly, in the recent past it has been facing stiff competition from public and private sector players offering services in other modes of transport such as road and air and hence has been experiencing a continuous decline in its market share. In the year 2001, an expert group headed by Dr Rakesh Mohan reported that IR was on the verge of a financial crisis and was therefore heading for a terminal debt trap. According to them one of the root causes for the financial problems confronting the IR was the lack of adequate productivity increases that are commensurate with real wages over time and they maintained that it was urgently required for the railways to adopt a strategic perspective to achieve higher growth in both the passenger and freight segments. Surprisingly in the last couple of years, IR has pulled out a financial turnaround, which attracted a lot of academic and practical interest from both India and abroad. The reason for this commendable feat is attributed to the result of a number of the top level strategic initiatives by the railway management. They include the efforts to increase its freight volumes by enhancing the varying capacity of the wagons and by reducing the wagon turnaround time. Currently, there is increased performance expectation from the railways and high-profile initiatives for achieving superior performance. Therefore, there is a strong motivation to analyze the operational performance of IR and benchmark the operations of its zones with respect to is peers so as to drive improved performance and to gain academic and practical insights.
A supply chain perspective of IR operations
IR is the largest public sector organization of the country. Owing to the large magnitude and the mix of the activities that IR is involved in, it has been divided into zones and zones into divisions and divisions into sections and thus is managed in a multi-tiered manner with a geographical divisional perspective. This structure basically involves the decomposition of management functions by physical jurisdiction of the railway network. At the same time, there is also a decomposition of activities based on the functional areas of railways as mechanical, electrical and civil engineering functions coupled with staff, traffic and commercial activities right from the topmost railway board level to the zonal, divisional and sectional levels. Therefore, in that perspective IR organization can be considered as having a matrix type of structure with multiple commands from the functional and geographical jurisdictions. For example, the chief electrical engineer of a zone not only reports to the general manager of the zone, but also to his head in the railway board in charge of the electrical engineering function. Thus, IR is managed in a similar manner as that of large multi-divisional organizations.
The transport logistics in a supply chain is usually an intermediary that facilitates the physical flow of goods from the point-of-origin to the point-of-destination. The railway operations are mainly concerned with the effective movement of passengers and goods from one point to other and the associated services. The freight or passenger to move from one point to another has to pass through the originating zone, the different zones through which it will be carried and then to the destination point in the terminating zone. Each of these zones is involved in the configuration, coordination and improvement of a set of sequentially-related activities and therefore may be considered to constitute a service-oriented supply chain. We can consider each zone as a decision-making unit (DMU) which transform a set of inputs to a set of useful outputs. Each railway zone has its own set of multiple inputs and outputs. They consume number of resources at varying degrees to produce the useful outputs. Both the passenger and freight traffic may be considered to follow a supply chain network across these zones. Each stage in the chain adds value to the customers as they pass through these sequentially-related stages. Each zone has its own performance objectives to achieve within the limited resources at their disposal and these individual objectives should support the overall objectives of IR.
The loading of goods, the movement of the same through the different zones till it reaches the destination and then unloading of the same at the destination are sequentially related activities related to the freight traffic. Similarly, in the case of passengers, it is their effective and efficient movement from a point-of-origin to a point-of-destination that constitutes the sequentially related activities of the supply chain. From the customer perspective, IR can be perceived as a single entity and there is no need to be concerned about its internal organization while specifying the operational and commercial requirements. However, though the customer is not paying separately to the various zones that he is passing through, there is an internal system in the railways for the allocation of costs and revenues. As per an agreed upon formula there exists a centralized apportionment system of the railways by which the originating zones retain a part of the revenues generated by them and the remaining part will be apportioned between the other zones that contribute to the movement of this traffic. Thus, the total apportioned earning of a zone is the sum of its retained share and inward share from the other zones. Similarly, there is a centralized system for allocation of investments, which at present is not interlinked with the performance but in the form of budgetary support that depend on social and other political compulsions.
A fundamental SCM principle is to manage the physical, informational and financial flows so that they are addressed in a consistent and effective manner with that of the strategic objectives. The success of SCM is essentially reliant on factors the current performance metrics are not sufficiently adequate to measure (Saad and Patel, 2006). Therefore, is important to understand the flows of material, information and cash along the supply chain and to ensure that these flows fit with the supply chain objectives in the divisional, zonal and corporate levels of the IR. The sequentially-related stages in the chain add value to the services offered to the customers of railways.
Data envelopment analysis
DEA is a performance measurement technique developed by Charnes et al. (1978) which can be used to evaluate the relative efficiencies of a set of DMUs with the same set of inputs and outputs. DEA is an effective approach to assess the efficiency of organizations, where traditional performance measurement either fail or are difficult to apply. Such situations typically occur especially within public sector or non-profit organizations, where no strict functional relationships can be formed, priory between factors of production, and where relative weights of these inputs and outputs are not well-defined (Golany and Roll, 1989). Though, DEA generates a large quantity of information to the decision makers, its main strength lies in overcoming the pitfalls of traditional performance measurement problems as pointed out by Djerdjouri (2005) such as:
- being able to handle disproportionate multiple inputs and outputs;
- not requiring the decision maker any priory arbitrary weights; and
- being able to combine multiple inputs and outputs to a single comprehensive measure of relative efficiency.
We now provide a brief review of the basic, super efficiency and cross-efficiency DEA models that are used in this paper for data analysis.
Basic DEA model
DEA defines the efficiency of each DMUs as the ratio of the weighted sum of outputs to the weighted sum of inputs. The outputs are the products and or services produced by the units and inputs are the resources used to produce these outputs. A unit with an efficiency score of 1(100 per cent) is considered as efficient and a score of less than one indicates that the unit is inefficient. Each unit is allowed to select the optimal weights that maximize its efficiency, subject to the condition that the efficiency of all the units in the set when evaluated with these weights are not allowed to exceed one. In the basic DEA model developed by Charnes et al. (1978), the objective is to maximize the efficiency value of a test DMU; p from among a reference set of n by selecting the input and output weights associated with the inputs and outputs. Therefore, the weights for the inputs and outputs are the decision variables. The original mathematical model is formulated as follows: Equation 1 where, k=1, … , s (outputs); j=1, … , m (inputs); i=1, … , n (DMUs); y ki = amount of output k produced by DMU i; x ji = amount of input j utilized by DMU i; v k = weight given to output k; u j = weight given to input j.
This fractional program can be solved as an LPP by setting its denominator equal to some arbitrary constant and maximizing its numerator. Alternatively it can be solved by setting the numerator equal to some constant and minimizing the denominator. Therefore, the equivalent LPP, which can be solved by commercial LP software, is formulated as follows: Equation 2 In some DEA implementations, for computational convenience, the dual of the above program is formulated and solved. This problem is run n times to identify the relative efficiency scores of all the units. Depending on whether inputs and outputs are controllable, a DMU can have either an input orientation or output orientation.
Reduced CCR model
The advantage of the reduced CCR model (RCCR) or super efficiency formulation proposed by Andersen and Petersen (1993) is that it allows for the ranking of the efficient units within themselves. In the RCCR model, the unit whose efficiency is to be maximized is excluded from the constraint set, thus allowing its relative efficiency score to exceed the value of 1. The model is run n times where n is the number of units. The RCCR model solely bases its calculations on how large a score a unit can achieve, without violating the efficiency constraints of the other units in the set. The efficiency score of the inefficient units remains same in both CCR and RCCR models and it is only for the efficient units that have got a possibility of achieving scores greater than 1. The model is formulated as: Equation 3
Cross-efficiency model
One of the main shortcomings of the traditional DEA approach is that each unit tries to optimize its own weights to get maximum efficiency while there could be other units which could perform even better with these weights. This sometimes tends to overestimate the efficiency of some of the units resulting in false positive (maverick) efficiency scores. A false positive unit weighs heavily on a single input or output making that unit more efficient than others. A procedure for discriminating between truly efficient and false positive units is by analyzing the cross-efficiencies (Sarkis and Talluri, 2004). The aggressive cross-efficiency formulation by Doyle and Green (1994) not only maximize the efficiency of the target unit but also minimize the efficiency of the composite units constructed from other (n−1) units. The cross-efficiency is a more democratic concept, where the efficiency of a unit is obtained by using its peer group's ratings and their optimal weights are substituted to find out the relative efficiency of all the others. We can thus calculate the cross-efficiency matrix (CEM) of each unit as seen by others, where the diagonal elements are a unit's own self-appraisal and the non diagonal elements give its peer-appraisal figures. The mean cross-efficiency scores can be used to rank the units. The aggressive cross-efficiency model formulated by Doyle and Green (1994) is represented as shown below: Equation 4 where, DMU p is the target DMU; ∑ k (v k ∑ i≠p y ki ) is the weighted output of the composite DMU; ∑ j (u j ∑ i≠p x ji ) is the weighted input of the composite DMU; pp is the simple efficiency of the DMU p.
Cross-efficiency measures bring a democratization process among the units, where they not only do self-appraisal but each individual unit is cross-checked with its peer appraisals. Thus, we can eliminate the tendency of the units to over-rate them and thus can make the efficiency figures more realistic. Also, we will be able to identify the units that enjoyed the maximum relative increment when we shift our focus from peer-appraisals to self-appraisal and term them as maverick units. Maverick index (MI) of a DMU is defined as the ratio of the optimal efficiency score from CCR formulation to the mean cross-efficiency score of the test unit.
Application to benchmarking of IR zones
The above described DEA models are used for the performance benchmarking of the IR zones. Basically, the implementation of DEA involves identifying the inputs and outputs of the units being assessed, identifying measures for the inputs and outputs, collecting data on the inputs and outputs, solving the appropriate models and interpreting the results (Thanassoulis et al., 1987).
Inputs and outputs
The assessment of operational efficiency using DEA starts with selection of appropriate input and output measures that can be aggregated into a composite index of overall performance. Inputs are basically the different resources that the zones consume for its operations while the outputs represent a set of quantitative measures of results expected from these zones. DEA methodology requires the inputs and outputs characterizing the performance to be identical for each zone, except for differences in magnitude. If the number of units for analysis is high, the probability of capturing high-performance units that determine the efficiency frontier will be high and more inputs and outputs can be incorporated in the analysis. The input-output set should confirm to the exclusivity, exhaustiveness and exogeneity requirements and should involve as few variables as possible (Thanassoulis, 2001). In this study, they were chosen based on previous DEA studies reported in literature and also on the availability of data.
Input 1: operating expenses (OE)
One of the major inputs of any organization is the capital. This input represents the aggregate capital inputs to the different zones in the form of grants which include, operating expenses (OE) for rolling stock and equipments, the repairs and maintenance expenses, OE for traffic and fuel, the staff wages, welfare and other amenities, etc. The input was represented in units of thousands of rupees.
Input 2: tractive effort (HP)
Another input considered was the tractive effort needed for the locomotives to be in operation which would determine to a great extent the capacity of the zones to handle the volumes of traffic. This covered the locomotives for both the passenger and goods trains in the zone and also included the locomotives running in the broad gauge, metre gauge and narrow gauge tracks. There are also three types of locomotives used by the railways namely the diesel (hydraulic and mechanical type), diesel-electric type and the electric type. The tractive effort was aggregated as the total horse power (HP) consumed by the different types of locomotives running on the different gauges of tracks.
Input 3: equated track kilometres (ETK)
One of important resources for the railway operations is the track assets. The input represents the equated total kilometres of track contained by the zones from the three gauges of broad, meter and narrow gauges. Each gauge was divided into electrified and non-electrified sections with electrified sections further into AC and DC tractions.
Input 4: number of employees (EMP)
The total number of staff employed in each zone was taken as one of the inputs for the zone.
Input 5 and 6: number of passenger carriages (PC) and number of wagons (WG).
The wagon holding and carriage holding of the zones indicate the level of their goods stock and coaching assets for freight and passenger traffic, respectively. These were measured as the average number on the line in terms of four wheelers.
Output 1: passenger kilometres (PK)
The main function of railways is the efficient and effective movement of passengers and goods. A good indicator of the carrying capacity of the zones is the throughput. The throughput refers to the total amount of traffic carried in a particular time not in transport supply units like number of trains, but in demand units like number of passengers (Rangaraj and Srivastava, 2001a). Passenger kilometres (PK) and ton kilometres are the most commonly used operating measures for passenger and freight traffic, respectively, (Ramanathan, 2003). Therefore, the first output is taken as the number of PK carried by each zone in thousands. PK is defined as the product of the number of passengers carried and average distance travelled. This number represented the aggregate of all the three gauges and all the various passenger classes such as A.C first class, A.C sleeper, A.C three tier, first class, A.C chair car, sleeper class and second class. Suburban and non-suburban railways were considered wherever applicable.
Output 2: ton (freight) kilometres (TK)
IRs major revenue earning is from the freight traffic. It is observed that the passenger traffic earned approximately 30 percent of the total earnings as compared to 65 percent from freight traffic. The remaining 5 percent earnings are from parcel traffic and other miscellaneous revenues. The bulk of the freight traffic include coal and other raw materials for steel plants and thermal powerhouses, iron ore for export, cement, food grains, fertilizers, mineral oils and other commodities. Ton kilometres (TK) is the product of the tons carried and the average distance. This output represented in units of 1,000 tons aggregates the net-freight kilometres of goods and proportion of mixed traffic of the different zones for all types of tractions and gauges.
Data analysis
The main source of the data used in this study is the Indian Railways Annual Statistical Statements published by the Ministry of Railways (1997-1998, 1998-1999, 2004-2005), Government of India. This statistical statement contains the zone-wise financial and operational information of the railways. Four years data were collected for the years 1998 and 1999 with nine zones and 2004 and 2005 with the present 16 zones. Preliminarily, four models were analyzed with the PK and TKs as the outputs and the following as the inputs:
- OE and tractive effort;
- OE, tractive effort and equated track kilometers;
- OE, equated track kilometers, no. of passenger carriages (PC) and no. of goods wagons; and
- no. of employees, equated track kilometers, no. of PC and no. of goods wagons.
The data of the above inputs and outputs for all four years are shown in Table I.
The DEA exercise of model 1 and model 2 revealed that the scores were identical in most cases for all the four years. The correlation analysis of inputs also showed that equated track kilometres (ETK) have been strongly correlated to OE. Similarly, the close agreement between DEA scores of model 3 and model 4 and the high correlation between the inputs OE and number of employees (EMP) enabled us to exclude the input EMP from the input set without loss of generality. Considering the satisfactory agreement of the DEA scores of model 1 with the other three models it was inferred that the two inputs OE and tractive effort along with the two outputs PK and ton kilometres are able to capture the operating efficiency of the system to a good extent and therefore was decided to use only model 1 for the subsequent data analysis.
Before the data are executed to determine the efficiency scores, the original absolute values were mean-normalized by dividing each value of a zone by the mean value of all the zones for that respective input or output factor so as to lessen the impact of scaling difficulties (Sarkis, 2000).
Simple and super efficiencies
Firstly, the input-oriented basic DEA model described before was run as many times as the number of zone to identify the relative efficiency score of each unit. The efficient units were further ranked by using the RCCR model described before and the super efficiency scores were obtained from this analysis. The results of the analysis are shown in Table II.
The results indicate that for the years 1998 and 1999, only four zones namely Central, North Eastern, South Eastern and Western zones were relatively efficient with a DEA score of 100 percent. It can also be observed that out of the 16 zones in the year 2004, seven of them, namely Central, East Coast, North Eastern, North Central, North Western, South East Central and Western were the efficient zones with a score of 100 percent. In 2005, also six zones obtained efficiency score of 100 percent. Central, East Coast, North Eastern, North Central, South East Central and Western remained efficient while the North Western zone efficiency came down from 100 percent to 90.81. Both these years the South East Central Zone has been identified as the best performer with the highest super efficiency score. Southern, North East Frontier, Eastern, South Eastern and South Western are the poorly performing zones. It can be observed that Central, North Eastern and Western zones have been consistently efficient for all the four years considered whereas the South Eastern zone which was efficient before restructuring of the zones turned to be an inefficient one.
Cross-efficiency analysis
The cross-efficiencies of the zones were analyzed using the formulation earlier discussed. A 9 × 9 CEM was generated by formulating and solving all the 81 LPPs for year 1999 and 16 × 16 CEM for the year 2004 (Tables III and IV).
For example, 0.56 the value in the first row second column of Table IV represents the cross-efficiency score of Eastern zone by the Central zone. The mean cross-efficiency scores calculated using this formulation can be used to rank the zones. The mavericks or the false positives are defined through a MI score was calculated using the expression: Equation 5 where MI P is the maverick index for the test zone k, pp is the optimal score from CCR formulation for the test unit and e p the mean cross-efficiency score of the test zone.
The cross-efficiencies for the efficient zones of Central, North East, South Eastern and Western for the year 1999 are 0.65, 0.73, 0.64 and 0.82, respectively, which means South Eastern Zone benefited the most and Western zone the least by moving from peer to self-appraisal. Therefore, Western zone can be considered as the best performer for this year. Also among the efficient zones, the MI of South Eastern zone is the highest. Since the cross-efficiencies of the zones are within a close range, we could conclude that there is a good agreement between self- and peer-appraisal scores of these units.
A comparison of DEA scores and operating ratios
Railways use the operating ratio (ratio of the working expenses to the gross earnings) as a major indicator of its financial performance and that of the zones. For the year 1999 and 2004, the IR had an average operating ratio of 93.31 and 92.22, respectively. The zone-wise operating ratio percentage with the CCR scores and cross-efficiencies for the years 1999 and 2004 are shown (Tables V and VI) to make a comparison between them. Zones are ranked in the descending order of their operating ratio.
For the year 1999, it can be observed that the efficient zones South Eastern, Western and Central Railways have relatively good operating ratios. But in the case of the CCR efficient North Eastern zone, the operating ratio is too large, thus showing a wide disagreement between the two measures. The same holds true in the case of 2004 data as well, though its cross-efficiency is the lowest among efficient zones.
The DEA evaluation of some of the zones does not seem to be closely in agreement with that of their perceived performance level in terms of the operating ratio. It brings into light in some cases the weaknesses in the financial measures and in other cases could be weakness of the input output specification of the DEA. For the year 1999, the disagreement is most obviously reflected in the case of North Eastern Railways which by DEA evaluation is one among the efficient zones but has a very poor operating ratio. Similarly, for the year 2004 major disagreement between DEA evaluation and operating ratio is observed in the case of both North Eastern and North Western zones. There could be two possible reasons for this disparity. Firstly, it can also be observed from the input output data that both these zones predominantly handle passenger traffic. The freight-oriented railway zones have greater leverage on the operating ratios than passenger-oriented zones that results in zones with high scores in terms of the passenger traffic would still be losing out in operating ratio due to its lesser freight traffic. This is because passenger tariffs have been cross-subsidized by railways due to political and social compulsions. Secondly, another point to be considered here is the originating traffic of the zones (ton originating and passenger originating), since those zones from where the traffic is originating share a major portion of the revenues. The proportion of originating goods traffic to ton kilometres carried by the above zones were lower compared to the other zones which again got reflected in their inferior operating ratios. However, efficiency of traffic carried by the zones is equally important in a supply chain perspective since it is the part of the sequentially related activities of the supply chain that add value to customer and in turn helps to attract more business in future. It is also interesting to note that the other passenger traffic-dominated zones of Eastern and Southern zones have both very poor DEA scores (self- and cross-efficiencies) and operating ratios. This further supports the result that North Eastern zone is an efficient zone in terms of the passenger traffic.
Discussion and conclusions
The study has demonstrated how IR could evaluate the operational efficiency of its zones using DEA methodology. DEA is potentially attractive because it overcomes a number of drawbacks of traditional performance measurement systems and provides a single integrated measure of relative efficiency. A variety of performance measures sometimes make it difficult to draw broad inferences about the relationships between the initiatives and performance outcomes. The DEA methodology gives us an aggregate measure of performance and the system approach followed helps us to look at the operations in a supply chain perspective that involve all the sequentially related stages to satisfy the needs of the end customers. Fundamentally, supply chain management involves the management of physical, informational and financial flows in an efficient manner at the same time being consistent and effective with the end customer needs. This can well be applied to both passenger and freight operations of the railways. However, the nature of the supply chain is different in both this cases with the passenger business being predominantly supply driven and freight business demand driven and multi-modal (Rangaraj and Srivastava, 2001b). These complexities make the systems approach all the more relevant to analyze the performance outcomes at various levels and provide feedbacks for improvements.
The aggregate operational efficiency measures presented here serves as a valuable indicator for identifying bottlenecks and formulating potential strategies for improvement of performance by information integration and allocation of resources in the supply chain. Within the set of inputs and outputs considered, the exercise identified the best performers over the years and the efficiency trends. The approach helped to not only identify the inefficient zones but also to provide information on how much improvement they could make and also their benchmarks to emulate for performance improvement. Some weaknesses of the conventional DEA were also addressed by including the concept of cross-efficiencies along with self-efficiency and also by analyzing longitudinal data spread over four years and comparing the efficiencies with the operating ratios.
To an extent, this study has also enabled us to understand the impact of restructuring of the zones on their performance. In fact, one of the major reasons for the restructuring of the zones was the fact that zones like Northern, Western and Central Railways were becoming unwieldy in terms of size, operational response to certain crisis situations and sheer work force (Rangaraj, 2006). Results of this study indicate that these zones have more or less maintained the same performance levels even after the restructuring. Paradoxically, although these have been affected by the change, they are not the ones most affected. The maximum efficiency drop was observed in the case of the South Eastern zone, which was formerly a relatively efficient one. In fact, it was South Eastern that was split into three zones namely East Coast, South East Central and South Eastern zones. The new zones East Coast and South East Central zones have been identified as relatively efficient while South Eastern zone has experienced a drop in its efficiency. Thus, it can be observed that although the restructuring has not affected the operational performance of some of the zones, it has detrimentally affected the performance of many other ones. Based on our studies it can very well be observed that Central and Western zone have been consistently performing well both before and after restructuring and among the new zones East Coast, North Central and South East Central appear to be good performers. Existing zones Southern, South Central, Eastern, North East Frontier and Northern which were inefficient zones before restructuring continue to remain so after restructuring with even further drop in their efficiency scores. Among newly formed zones of East Central, South Western and West Central, it appears that there is a vast scope for efficiency improvements. However, further analysis incorporating other inputs and outputs and also data of more number of years would be necessary to confirm these findings. Nevertheless, the study has attempted to gain some valuable insights from a managerial perspective for IR so as to formulate strategies of its zones to foster better performance.
Equation 1
Equation 2
Equation 3
Equation 4
Equation 5
Table IInputs and outputs
Table IIEfficiency scores
Table IIICross-efficiency matrix 1999
Table IVCross-efficiency matrix 2004
Table VOperating ratio comparison 1999
Table VIOperating ratio comparison 2004
References
Agarwal, A.L., Makker, A.K. (2002), "Railway finances off the track", in Kanda, A., Vrat, P., Agarwal, A.L., Deshmukh, S.G. (Eds),Transportation Systems: Status and Directions, Phoenix Publishing House, New Delhi, pp.68-89.
Andersen, P., Petersen, N.C. (1993), "A procedure for ranking efficient units in DEA", Management Science, Vol. 39 pp.1261-4.
Charnes, A., Cooper, W.W., Rhodes, E. (1978), "Measuring efficiency of decision making units", European Journal of Operational Research, Vol. 2 pp.429-54.
Dharmapala, P.S., Saber, H.M. (2007), "Resource allocation within academic units: DEA and goal programming approach with an application to an Arabian Gulf University", Benchmarking: An International Journal, Vol. 14 No.2, pp.202-10.
Djerdjouri, M. (2005), "Assessing and benchmarking maintenance performance in a manufacturing facility: a DEA approach", INFOR, Vol. 43 No.2, pp.121-33.
Doyle, J., Green, R. (1994), "Efficiency and cross efficiency in DEA: derivations, meanings and uses", Journal of Operational Research Society, Vol. 45 pp.567-78.
Duffy, J.A.M., Fitzsimmons, J.A., Jain, N. (2006), "Identifying and studying ‘best performing’ services: an application of DEA to long-term care", Benchmarking: An International Journal, Vol. 13 No.3, pp.232-51.
Forker, L.B., Mendez, D. (2001), "An analytical method for benchmarking best peer suppliers", International Journal of Operations & Production Management, Vol. 21 No.1/2, pp.195-209.
Golany, B., Roll, Y. (1989), "An application procedure for DEA", OMEGA International Journal of Management Science, Vol. 17 No.3, pp.237-50.
Jong Joo, S., Stoeberl, P.A., Kwon, I.G. (2007), "Benchmarking efficiencies and strategies for resale operations for a charity organization", Benchmarking: An International Journal, Vol. 14 No.4, pp.455-64.
Karlaftis, M.G. (2004), "A DEA approach for evaluating the efficiency and effectiveness of urban transit systems", European Journal of Operational Research, Vol. 152 pp.354-64.
Lee, Y.P., Zailani, S., Soh, K.L. (2006), "Understanding factors for benchmarking adoption: new evidences from Malaysia", Benchmarking: An International Journal, Vol. 13 No.5, pp.548-65.
McGaughey, R.E. (2002), "Benchmarking business-to-business electronic commerce", Benchmarking: An International Journal, Vol. 9 No.5, pp.471-84.
Min, H., Park, B.I. (2005), "Evaluating the inter-temporal efficiency trends of international container terminals using data envelopment analysis", International Journal of Integrated Supply Management, Vol. 1 No.3, pp.258-77.
Ministry of Railways (1997-1998, 1998-1999, 2004-2005), Indian Railways Annual Statistical Statements, Government of India, New Delhi, .
Ministry of Railways (2002), Notification on New Zonal Railways, Government of India, New Delhi, .
Ministry of Railways (2004), Annual Report and Accounts 2004-05, Indian Railways, Government of India, New Delhi, .
Mukherjee, A., Nath, P., Pal, M.N. (2002), "Performance benchmarking and strategic homogeneity of Indian banks", International Journal of Bank Marketing, Vol. 20 No.3, pp.122-39.
Pangotra, P., Raghuram, G. (1999), "Resource mobilization strategies for financing of transport infrastructure and services", in Raghuram, G., Jain, R., Sinha, S., Pangotra, P., Morris, S. (Eds),Infrastructure Development and Financing towards a Public-private Partnership, Macmillan India Ltd, New Delhi, pp.327-55.
Radnor, Z., McGuire, M. (2004), "Performance management in the public sector: fact or fiction?", International Journal of Productivity and Performance Management, Vol. 53 No.3, pp.245-60.
Raghuram, G. (2001), Integrated Transport: India Infrastructure Report 2001, Oxford University Press, New Delhi, .
Ramanathan, R. (2003), An Introduction to Data Envelopment Analysis. A Tool for Performance Measurement, Sage, New Delhi, .
Rangaraj, N. (2006), "Implications of the zonal and divisional jurisdictions of the Indian Railway organization on their operations", Technical Note, IE&OR, IIT Bombay, Mumbai, .
Rangaraj, N., Srivastava, A.K. (2001a), "Network capacity planning and operational control in Indian Railways", Conference Proceedings, National Connectivity Vision: Surface Transport and Communication, Indian National Academy of Engineering, New Delhi, pp.19-34.
Rangaraj, N., Srivastava, A.K. (2001b), "Some IT imperatives for Indian Railways in a supply chain management environment", Conference Proceedings, National Connectivity Vision: Surface Transport and Communication, Indian National Academy of Engineering, New Delhi, pp.369-78.
Saad, M., Patel, B. (2006), "An investigation of supply chain performance measurement in the Indian automotive sector", Benchmarking: An International Journal, Vol. 13 No.1/2, pp.36-53.
Sarkis, J. (2000), "An analysis of the operational efficiencies of major airports in the United States", Journal of Operations Management, Vol. 18 pp.335-51.
Sarkis, J., Talluri, S. (2004), "Eco-efficiency measurement using DEA: research and practitioner issues", Journal of Environmental Assessment Policy & Management, Vol. 6 No.1, pp.91-123.
Talluri, S. (2000), "A benchmarking method for business-process reengineering and improvement", International Journal of Flexible Manufacturing Systems, Vol. 12 No.4, pp.291-304.
Thanassoulis, E. (2001), Introduction to Theory and Applications of Data Envelopment Analysis, Kluwer Academic Publishers, New York, NY, .
Thanassoulis, E., Dyso, R.G., Foster, M.J. (1987), "Relative efficiency assessment using DEA: an application to data on rates departments", Journal of Operations Research Society, Vol. 38 No.5, pp.397-411.
Verma, A., Gaverneni, S. (2006), "Measuring police efficiency in India: an application of DEA", Policing: An International Journal of Police Strategies & Management, Vol. 29 No.1, pp.125-45.
Wynn-Williams, K.L.H. (2005), "Performance assessment and benchmarking in the public sector", Benchmarking: An International Journal, Vol. 12 No.5, pp.482-92.
Further Reading
Raghuram, G., Rangaraj, R. (2000), Logistics and Supply Chain Management: Cases and Concepts, Macmillan India Limited, New Delhi, .
About the authors
Sajeev Abraham George is an Assistant Professor in Production Engineering in the M.H.S.S. College of Engineering, Byculla, Mumbai, affiliated with the University of Mumbai. He received his BTech in Mechanical Engineering and MTech in Industrial Engineering from the National Institute of Technology, Kozhikode, Kerala. He is pursuing a PhD in the area of supply chain management in the interdisciplinary programme, industrial engineering and operations research from the Indian Institute of Technology, Bombay.
Narayan Rangaraj is a Professor and Convener of the Interdisciplinary Programme in Industrial Engineering and Operations Research at the Indian Institute of Technology, Bombay. His areas of research interest include logistics and supply chain management, performance of small firms, transport management and railway operations. He has co-authored a book on logistics and supply chain management and has published papers in major journals. Narayan Rangaraj is the corresponding author and can be contacted at: narayan.rangaraj@iitb.ac.in