Performance-based transit subsidy allocation scheme to fulfill multiple policy criteria

ABSTRACT

Performance-based subsidy allocation schemes are typically intended to satisfy a single policy goal (e.g. social welfare maximization) and often do not include objective approaches to quantifying bus service efficiency. This paper proposes a framework for network data envelopment analysis and introduces an innovative transit subsidy allocation scheme with a fixed budget that fulfills multiple policy criteria. The proposed basic scheme minimizes the maximum deviation from efficient allocation in proportion to the operating scales across all routes in a two-stage production system. Multiple policy criteria are then added to the basic scheme as additional goal components. The proposed scheme is evaluated by using an empirical dataset of 344 subsidized intercity bus routes collected from 20 Taiwanese bus operators. The results demonstrate that the subsidies allocated on the basis of the proposed scheme can satisfy the policy criteria incorporated into the scheme. The proposed modeling framework is flexible, and therefore, local governments can easily incorporate policy criteria into the scheme according to their own interests and needs.

KEYWORDS:

• Transit subsidy
• performance-based subsidy
• subsidy allocation
• network data envelopment analysis
• efficiency
• multiple objectives

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. Lim and Zhu (2016) examined the matter of obtaining frontier projections and divisional efficiency scores through the use of dual models within a two-stage NDEA model. Chen et al. (2013) suggested that the envelopment-based NDEA model be employed to establish frontier projections for inefficient DMUs and that the multiplier-based NDEA model be used to determine divisional efficiency.

2. Nondiscretionary factor problems occur when certain inputs and outputs, or factors that are crucial for measuring efficiency, are beyond the control of management (Banker and Morey 1986).

3. A share-input variable is an input in the production process that is directly related to output. Because of this direct relationship, share-input variables are directly incorporated into the production function, whereas subsidies, which less directly affect output, are incorporated as additional factors and should not be treated as input variables. Rather, subsidies involve centralized allocation and should be handled in a centralized manner, as was reported in Li et al. (2019).

4. According to the definition in Amirteimoori, A., Kordrostami, S (2005) Allocating fixed costs and target setting: A DEA-based approach. Applied Mathematics and Computation 171, 136–151.

5. Model (8) and model (13) involve multiple objectives, and thus, they cannot be solved directly through standard methods or solved easily. The computation algorithm proposed by Li et al. (2019) can be used to solve model (8) and model (13). Further details can be found in Li et al. (2019).