A hybrid discrete choice model to understand the effect of public policy on fare evasion discouragement in Bogotá's Bus Rapid Transit

Abstract

Listen

Translate article

• The study was carried out with a sample of actual evaders, which is unique in fare evasion studies. • Is critical to understand which observable and non-observable factors motivate individuals to evade fares. • Age, the evasion records, and some personality traits influence the fare evasion behavior. • Contributes with a novel application of discrete choice models to study fare evaders behavior. • Draw relevant policy implications for a large public transport system in South America. Fare evasion is currently a significant concern for public transport systems worldwide. Evasion causes huge economic losses to operators and cities, but also negatively influences the system’s security perceptions. Despite its importance, there are few efforts to understand which observable and non-observable factors motivate individuals to evade fares in a public transport system. Most studies dealing with fare evasion evaluate the effectiveness of several policies, quantify the amount of evasion, or correlate fare evasion with system satisfaction using aggregate data. This paper proposes a Hybrid Discrete Choice model to understand current evaders’ behavior considering sociodemographic, fare evasion records, trip characteristics, user satisfaction with the transport system, and perceptual indicators. The model was estimated using a unique stated preferences survey applied to 324 fare evaders of Transmilenio, the Bogotá (Colombia) Bus Rapid Transit system, in mid-2019. This is an interesting case study because, in 2018, the evasion rate was measured for the first time and it was discovered that evaders comprised approximately 15% of their users on weekdays. The contribution of this research is twofold. First, results highlight the importance of considering observable and latent variables to understand individual behavior in examining fare evasion in similar public transport systems. Age and the evasion records resulted in being the most important observable variables in our case study. Latent variables suggest that the more the user is satisfied with the transport system, the less likely is to evade fares. However, we found that personality traits, such as individual rule-unconsciousness, moderate the effect of satisfaction on the evasion behavior. Second, our results provide evidence that the discussion between certainty and severity of punishment on fare evasion should not necessarily be centered on whether to exacerbate one of these factors. A combination of both, however, should reduce the recidivism rate.