Frontier studies in Taiwan

The DEA studies concerning Taiwanese bus transit system are reviewed as follows.

Chang and Kao (1992) employed the data envelopment analysis method to evaluate the efficiency of the five bus firms in Taipei city. When vehicle kilometers (revenue or the measure combining vehicle kilometers, revenue and the number of traffic trips on routes) was used as the output measure, it concluded that the publicly owned Taipei Municipal Bus had increased (not increased) its technical efficiency after the government liberalized the urban bus market. This article also found that in both the one output (vehicle kilometers) and three outputs cases, Taipei Municipal Bus had, on an average, lower efficiency scores than the private firms, and that while each firm usually employed a linear production technology for several, consecutive years the private firm were more flexible in adopting different technologies.

Cheng and Shiau (1994) extended the traditional one-stage performance approach, which only considered the relationship between input and output, to the two-stage performance approach that combines the concepts of efficiency and effectiveness to evaluate the relative aggregate performance among highway bus operators (HBOs). The DEA model modified by CCR (Charnes, Cooper and Rhodes) was applied to develop a two-stage performance model that includes:

1. a relative efficiency evaluation model that explored the relationship between input and output, and

2. a relative effectiveness evaluation model that explained the relationship between output and consumption.

Operation data of 32 HBOs were collected (1987 to 1991) and were used to perform the relative performance evaluations at cross sections of time, and over the series of time. These results of production structure showed that the production technologies were quite different among HBOs, and generally inflexible in the short term. Regarding the efficiency of resource use, most of the HBOs went from bad to worse in the recent years. In effectiveness of output utilization, the output of the HBOs which served in the central, southern and eastern Taiwan were utilized less effectively than that of the HBOs in the northern Taiwan. In addition, this study also found that the HBOs with better relative performance in the one-stage model presented the worse results on either relative efficiency or relative effectiveness of the two-stage model. This result indicated the two-stage performance approach could more precisely evaluate and distinguish between the efficiency of resource use and the effectiveness of output utilization. This study could help an HBOs to identify future directions for improving operation performance.

Change and Sun (2001) presented a two-stage model for measuring the performance of Taipei urban bus companies. The bus transit performance is measured in terms of overall performance, operation efficiency, operation effectiveness, service effectiveness, and cost efficiency. This paper concludes that:

1. Shoutu and Hsinhsin (1998). Tayo (1997-1999), Sanchung (1997, 1999), Sintien, Fuho and Hsinbo (1999) reached overall efficiency;

2. Hsinho (1997-1999) reached operation efficiency;

3. HsinHsin (1997-1999), Fuho, Hsinho (1999) reached operation effectiveness;

4. Hsiho (1999), Tamsui (1997), Fuho (1997,1999) and Tayo (1997,1998) were rated service efficiency;

5. Tayo (1998). Fuho and Hsinho (1999) reached cost efficiency;

6. High efficiency means high service effectiveness while high efficiency doesn’t mean high operation effectiveness; and

7. High overall performance means high cost efficiency.

Chang (2003) indicated that TMTC was the largest public intercity bus company in Taiwan. With a cumulative debt over 40 billions NT dollars, TMTC was one of largest financial burdens of the government and then forced to privatize in July 2001. The objective of this study is to examine the impacts of privatization. The operating data of TMTC before and after privatization. From the operator perspective, before and after analysis was applied first to compare the major performance measures. The results show that all the performance measures are significantly improved. From the user perspective, a questionnaire was designed to survey the perceived changes by the users. The results show that most users agree that the services have been significantly improved after privatization. In order to closely examine the impacts of privatization on operating performance, the DEA was also applied to compare the relative efficiency for the 14 major service lines. The Tobit regression analysis was followed to analyze the service variables and characteristics that can significant influence operating efficiency. The results indicate that service distance over 150 kilometers, freeway lines, the number of competitors, and providing demand responsive service can significantly increase service efficiency. Overall, privatization provides positive impacts on intercity bus services.

Cho and Fan (2003) empirically examines the property right theory in economics comparing the performance of TMTC before and after privatization. The criteria of both the Malmquist TFP Index and its decomposing techniques are used to estimate this performance change in the framework of the DEA. This study has found that TMTC’s privatization has had a positive impact on the productivity enhancement and hence confirmed the property right

theory, but no evidence was found that the technical efficiency improved because of privatization. Therefore, this study presents some management problems with regard to inefficiencies existing within the TMTC’s privatization. Future works are also suggested.

Cho and Fan (2004) propose three categories of productivity measures to investigate the changes of production productivity, service productivity as well as consumption productivity following TMTC’s privatization. The results indicate that privatization had a striking impact, in terms of TFP, on various productivity growths. The decomposing results demonstrate that the technical change was the most important factor for the new owner’s (KKTC’s) productivity progress while efficiency of effectiveness change had little contribution to this growth. However, further decomposing results suggest that the insignificant efficiency of effectiveness in the newly-privatized firm may be attributed to either incorrect selection of input combinations or inappropriate returns to scale or to both.

In summary, all the literature reviewed above, can be divided into three groups in terms of measure, that is,

1. technical efficiency,

2. Malmquist total factor productivity (TFP) index, and 3. efficiency and effectiveness (or cost efficiency).

These are listed in Table 3.1 below. In terms of research topics, the literature reviewed can be classified roughly into five categories:

1. Privatization (Cowie 2002; Cho and Fan 2003, 2004);

2. Regulation (Button and Costa 1999; Nolan et al. 2002; Dalen and Gomez-Lobo 2003);

3. Owership (Change and Kuo 1992; Jorgensen 1997; Pina and Torres 2001; Odeck 2003; Cho and Fan 2003, 2004);

4. Subsidy (Obeng 1994; Sakano and Obeng 1995; Nolan 1996, 2002; Sakano et. al.

1997; Jogensen 1997; Odeck and Alkadi 2001; Obeng and Sakano 2002);

5. Others (Chu et al. 1994; etc).

Table 3.1 Classification of the Reviewed Literature

Technical efficiency Malmquist TFP index Efficiency and effectiveness

Non-parametric Button and Alkadi (2001), Costa (1999),

Nakanishi and Norswrthy (2000), Pina and Torres (2001), Odeck& Alkadi (2001),

Cho and Fan (2003,2004)¹

Chu et al. (1992),

Technical efficiency Malmquist TFP index Cost efficiency³

Parametric

Viton (1986),

Thiry & Tulkens (1992), Sakano and Obeng (1995), Sakano et al. (1997)², Jorgensen et al. (1997), Mata&Raymond (1998)²

Obeng et al. (1992), Sakano et al. (1997), Obeng&Sakano (2002)

Fazioli et al. (1993), Battacharyya etal. (1995), Dalen&Gomez-Lobo (2003)

Note: ¹ DEA study of Taiwanese bus transit industry

2 with scale efficiency

3 costs are used as the input variable