The impact of ownership on the cost of bus service provision: an example from Italy
Summary (2 min read)
1 Introduction
- The aim of this paper is to investigate the influence of ownership on the cost characteristics of a sample of bus companies operating local public transport (LPT) in Piedmont, a North Western region of Italy, and to draw some policy conclusions from such evidence.
- A translog cost frontier is estimated using the model in Battese and Coelli (1995) and the public ownership dummy, included in the inefficiency model, is always positive and significant pointing to higher inefficiencies for public companies.
2.1 Measuring inefficiency
- The aim of the paper is to give some evidence on the role of ownership on the efficiency of LPT firms.
- While c(.) is the deterministic cost structure, exp(v) represents the effect of exogenous shocks and exp(u) is the inefficiency.
- In a fixed effects model (firstly proposed by Schmidt and Sickles, 1984, in a frontier context) no distributional assumptions are imposed on the inefficiency term while since Pitt and Lee (1981) seminal contribution, random effects estimators are based on the imposition of a particular distribution on the inefficiencyii .
- Under the assumption of no correlation among the error terms and the regressors, distributional assumptions are imposed to both v and u terms.
- In their empirical application the authors decided: (a) to impose the influence of a set of external factors on the mean value of the truncated distribution of the inefficiency term and (b) to test the validity of the inclusion of another set of external factors into the cost technology.
2.2 The translog cost function
- A total cost function is estimated: TC = c(Y, N, pL, pM, pK, τ, Type of service) where total cost TC is a function of output Y, network dimension N, input prices for labour, other variable inputs and capital respectively (pL, pM, pK) and a time trend which approximates technology, τ.
- Summing up public firms are characterised by larger size and mixed urban and intercity services, while private firms mainly supply intercity services and they are smaller in terms of number of employees, vehicles and supplied bus-kilometres.
4.2 Density and scale economies in the bus industry
- They may produce interesting results as an automatic byproduct.
- Input prices are held fixed at the mean value of the whole sample.
- The number of vehicle-kilometres is much lower for intercity transport firms and this evidence allows for the existence of higher economies of density, that cannot be exploited as long as the type of service does not change and is constrained by a low level of demand.
- They present an inefficiency score of 1.033, whereas public firms in the range 150-250 employees, have an inefficiency score of 1.30.
- Xii Public companies show on average small economies of scale (1.072) while for private firms economies of scale are more relevant (1.151).
4.3 Inefficiency indicators
- Table 5 reports the estimates of the mean cost-inefficiency for the whole sample and separately for the two groups of public and private firms by different size classes to examine the possible correlations between firm size and inefficiency.
- The methodology itself is different as Fraquelli et al. compute the long run scale economies using the total number of places offered times km run as output variable and network size is not included.
- Characterised by particularly low levels of inefficiencies are excluded, (1.21 vs. 1.30, see footnote xiv)xv.
- Estimated inefficiencies are much higher under specification 1 (mean inefficiency 18%) but this result is not surprising as long as the two specifications give different measures of inefficiency.
- The influence of ownership on the provision of local public transport appears to be a relevant issue for policy makers, with particular attention to efficiency and economies of scale of the two groups, which should be carefully taken into account in the design of a transport policy.
Acknowledgments
- The authors wish to thank Luca Sanlorenzo for excellent assistance on the data and Massimo Filippini and Massimiliano Piacenza for helpful comments.
- The authors are also grateful to the participants to the Hermes Workshop on Local Public Transit, Moncalieri, July 18, 2005 and to the 4th North American Productivity Workshop (NAPW), New York, June 27-30.
- Financial support from Hermes (Higher Education and Research on Mobility Regulation and the Economics of Local Services) Research Centre is acknowledged.
6 References
- Battese G.E. and Coelli T.J., (1993), A Stochastic Frontier Production Function Incorporating a Model for Technical Inefficiency Effects, University of New England, Working Papers in Econometrics and Applied Statistics, n. 69.
- Coelli T. J. , Rao D. S. and Battese G.E., (1998), An introduction to efficiency and productivity analysis, Kluwer Academic Publisher, Boston.
Did you find this useful? Give us your feedback
Citations
128 citations
81 citations
62 citations
53 citations
45 citations
References
7,616 citations
"The impact of ownership on the cost..." refers background in this paper
...…model in Battese and Coelli (1995), which allows for the estimation of firm specific inefficiencies that vary over time, is thus implemented (see Coelli et al., 1998; Kumbhakar and Lovell, 2000, for a review of the theoretical and empirical literature on productivity analysis; see Piacenza,…...
[...]
5,783 citations
"The impact of ownership on the cost..." refers background or methods in this paper
...A translog cost frontier is estimated using the model in Battese and Coelli (1995) and the public ownership dummy, included in the inefficiency model, is always positive and significant pointing to higher inefficiencies for public companies....
[...]
...Following the model in Battese and Coelli (1995), the cost frontier is thus: Cit ¼ cðYit, pit; Þ expðvit þ uitÞ ð2Þ where uit Nþ( 0Zit, 2u), i.e. uit’s are nonnegative truncations of a normal distribution and they are independently but not identically distributed across time and firms and external…...
[...]
...However Kumbhakar et al. (1991) and Battese and Coelli (1995) addressed the inconsistency of the twostage method and suggested the simultaneous estimation of the production (cost) frontier and the inefficiency model....
[...]
...Bhattacharyya et al. (1995b) propose a two step procedure for the estimation of firm- and time-specific inefficiencies that takes into account firm- and time-specific variances, while we adopt the one step procedure introduced by Battese and Coelli (1995)....
[...]
...The panel structure of the data is exploited and the model in Battese and Coelli (1995), which allows for the estimation of firm specific inefficiencies that vary over time, is thus implemented (see Coelli et al., 1998; Kumbhakar and Lovell, 2000, for a review of the theoretical and empirical…...
[...]
3,396 citations
3,378 citations
"The impact of ownership on the cost..." refers background in this paper
...The formula for the computation of the expected value CÎit¼E(exp(uit) j it¼ it), where it¼ (uitþ vit), is a generalization of the results by Jondrow et al. (1982) and can be found in Battese and Coelli (1993)....
[...]
Related Papers (5)
Frequently Asked Questions (10)
Q2. What are the contributions in this paper?
For more Information regarding the PEER-project see: http: //www. peerproject.
Q3. What is the impact of ownership on the provision of local public transport?
The influence of ownership on the provision of local public transport appears to be a relevant issue for policy makers, with particular attention to efficiency and economies of scale of the two groups, which should be carefully taken into account in the design of a transport policy.
Q4. What are the two sets of estimators for cost frontier functions?
In the empirical literature on stochastic frontier functions, where panel data are available, at least two sets of estimators have been proposed (see Sickles, 2005 and Greene, 2002, 2004 for recent surveys on panel estimators for parametric and non parametric frontiers): fixed effects and random effects estimators.
Q5. How many employees does the dummy variable represent?
In order to control for the possibility of spurious relationships, the authors decided to include a set of dummy variables: SMALL, that equals one if the firm has less than 50 employees, MEDIUM, equal to one if the firm has more than 50 and less than 150 employees and LARGE that equals one if the firm has more than 150 employees.
Q6. What is the main caution when using a set of different variables as external influences?
The authors are going to use a set of different variables as external influences, assuming thatthey all are exogenous (i.e. out of the firm control) at least in the five years of their panel.
Q7. What is the significance of the interaction terms with PUBLIC dummies?
However the estimations for the interaction terms with PUBLIC*INTERCITY and PUBLIC*URBAN dummies are negative and bigger than 0.069 in absolute values.
Q8. What is the dummy in the inefficiency model?
A translog cost frontier is estimated using the model in Battese and Coelli (1995) and the public ownership dummy, included in the inefficiency model, is always positive and significant pointing to higher inefficiencies for public companies.
Q9. What are the results of the papers by Mizutani and Urakami?
The results are mixed: while Filippini and Prioni (2003) do not reach clear-cut outcomes, the papers by Mizutani and Urakami (2003) and Roy and Billion (2005) point to higher costs and higher inefficiencies respectively for publicly owned bus companies.
Q10. What is the significance of the parameter estimates for PUBLIC dummy?
The parameter estimates for PUBLIC dummy is positive (0.069) indicating the presence of significant higher inefficiencies for publicly owned companies.