COSTING OF BUS OPERATIONS

Abstract

The objective of this Academic Exercise is to ascertain cost behaviour of public bus operations. Such a study will aid in gauging service performance and hence assist in evaluating operational and investment decisions. Various methods used in calculating individual service costs are surveyed. In particular, economic approach of identifying fixed and variable cost can be used to decide whether a service should be implemented or withdrawn. Other approaches involve cost accounting techniques, which attempt to solve the problem of joint costs. Difficulties are encountered in using each of these methods and our study suggest that there is no superior method. Each of these methods are subject to the characteristics of cost of a particular operator and the purpose for which such costing system is used for. An attempt was made to analyse the cost structure of Singapore Bus Services. This was not possible due to the lack of available information at the micro level. The academic exercise then proceeded to develop an economic theoretical framework of costing with the objective of analysing cost behaviour at the macro level. This involved identifying cost functions of bus operators. Such functions are useful in ascertaining economies of scale, which will determine the appropriate fleet size of efficient operations Various functional forms like the linear, log-linear and Trans log are employed and discussed. Our findings favour the use of the Translog form as it places little restriction on the underlying production technology of the operator and allows for the testing of various functional forms. An empirical exercise is carried out with data on the Singapore Bus Services Limited (SBS). The data used consist of time series observations from 1979 to 1992 obtained from various sources. The results focus on the cost behaviour of the SBS particularly in the area of scale economies. Results of the study presented indicate that the Translog cost model gives a better explanation of the SBS technology than other more restrictive models like the linear and log-linear form. It also appears that the SBS is reaping economies of scale. Results pertaining to factor separability and homotheticity are also presented.