UC Berkeley

Cycling is gaining relevance globally since it is environmentally, economically, and socially sustainable and provides a wealth of health benefits. In Latin American cities, cycling is becoming more popular due to large cycling infrastructure investments implemented in pioneering cities with the purpose of making cycling safer and more attractive. However, despite its growing relevance, studies on cycling in Latin America are still scarce. Most literature on cycling examines North American, European, and Asian cities, and there are relatively few studies that analyze Latin American cities. Furthermore, much of the literature in Latin America examines just a single city. Thus, I address this gap in the literature by developing measures of bicycle accessibility using the concept of level of traffic stress or comfort. I examine the distribution of bicycle accessibility across different socioeconomic groups to understand it from an equity point of view. In addition, theory suggests that higher accessibility should be associated with additional trips; yet there is no evidence of this for bicycle accessibility in Latin America. I also address this limitation in the literature by examining associations between bicycle travel and accessibility. Finally, although many studies have identified differences in self-reported cycling comfort by gender, there is limited research at the individual level understanding differences in behaviors in Latin America. This is important because the social construction of gender identities and roles is largely caused by cultural characteristics and societal expectations. I, therefore, also examine associations between gender and cycling behaviors in large metropolitan areas of Latin America to elucidate city-specific associations between bicycle infrastructure and actual cycling.

This is the first large-scale study analyzing bicycle accessibility, its distribution across different sociodemographic groups, its association with cycling, and gender-specific factors associated with cycling for multiple metropolitan areas in Latin America. The research presented in this dissertation examines cycling in six large Latin American cities: Asuncion (Paraguay), Bogota (Colombia), Buenos Aires (Argentina), Mexico City (Mexico), Santiago (Chile), and Sao Paulo (Brazil). I make three distinct contributions:

• I measure bicycle accessibility by applying the level of traffic stress (LTS) to examine how the level of bicycle accessibility differs among different sociodemographic groups;• I examine associations between bicycle accessibility and the frequency of cycling trips; and • I investigate whether gender acts as a modifier of the infrastructure-cycling relationship. Here, gender differences in the environmental and sociodemographic characteristics associated with cycling were examined.

Level of traffic stress (LTS) is a novel measure that has emerged to address cycling comfort concerns. However, onerous data requirements have made calculating the LTS for many cities around the world impractical. I applied a reliable and generalizable method of calculating LTS utilizing open data to examine how sociodemographic characteristics relate to bicycle accessibility in the six metropolitan areas in Latin America. Residents' sociodemographic characteristics were obtained from household travel surveys in each metropolitan area. I examine accessibility to points of interest (POIs) using isochrone metrics based on 10, 20, and 30-minute travel time from home for several types of bicycle networks determined by different values of LTS and topography. The results show that POIs and bicycle infrastructure tend to be concentrated in and around the CBD, with housing submarkets unaffordable to lower-income residents. Accordingly, negative binomial regression models consistently show that income and educational attainment are positively associated with bicycle accessibility across cities. Specifically, high-income individuals enjoy between 1.08 and 6.10 times higher accessibility than low-income individuals. The accessibility index of individuals with college-level education is 1.12 to 4.15 times higher than that of individuals with primary education. The deployment of bicycle infrastructure in high-density, low-income neighborhoods greatly impacted the distribution of accessibility benefits in some cities. The results suggest that conducting equity studies for transportation planning is vital for Latin American cities, to address inequitably distributed access to opportunities.

Although cycling is a sustainable transportation alternative for cities, little is known about how accessibility measures are associated with cycling. I examined how isochrone accessibility is related to the number of cycling trips in the sample cities. The models show mixed results, contrary to theory and previous studies, which have suggested positive associations between accessibility and cycling. Although I found some positive associations between accessibility and cycling, numerous outcomes show the opposite: accessibility's negative associations with cycling. These opposite outcomes are found mostly in cities with a greater spatial mismatch between where cyclists reside and where the potential opportunities are. This trend is observed even after controlling for sociodemographic and physical environmental variables. Furthermore, in most models, the relationship between accessibility and cycling is statistically non-significant. This highlights the complexity of travel behavior and that the relationship between accessibility and cycling is not straightforward. Furthermore, it underscores the importance of interpreting accessibility measures in the context of the location patterns of individuals and uneven cycling infrastructure. On the other hand, the models show that sociodemographic variables better explain cycling trips than accessibility and physical environmental variables.

In Latin America, women's cycling rates are much lower than men's. Despite this difference, few studies have been published on the gender-specific effects of factors associated with cycling in this region. I address this gap by investigating gender differences in the environmental and sociodemographic characteristics associated with cycling in the sample cities. Specifically, I examined the factors associated with self-reported bicycle mode choice and cycling distances. I found that increased bicycle lane infrastructure, measured as the total km of bicycle lanes within a 500 m radius from home, is associated with more cycling and longer cycling distances, for both women and men. What is striking about the results is that individuals with dependent children, particularly women, tend to cycle more. This finding was unexpected because some scholars suggested that child caring is one of the main obstacles to women’s bicycle use. Nonetheless, women's lower rate of driver's licenses and limited access to motorized modes compared to men's, and women's higher trip rates during the child-raising period may explain this result. Furthermore, older women cycle significantly less than both women of other ages and their male counterparts. Thus, expanding the comfortable and protected cycling network to include nonwork destinations can increase women's cycling rates, meet the needs of women who engage in household care-focused activities, and address the wider gender gap in cycling at older ages. Because cyclists in these cities tend to have lower incomes and no access to automobiles, investments in cycling infrastructure not only address the current gender gap in cycling, but are also likely to favor disadvantaged individuals.

Latin America is a highly unequal region in terms of the spatial distribution of accessibility benefits. The spatial mismatch between where cyclists live and where opportunities are is evident in some cities, because of socioeconomic segregation and a non-diverse cycling population. For urban regulations and policies, conducting equity studies is vital to change this situation. Furthermore, providing cycling infrastructure for a diversity of people, including women and older people, is important to increase overall cycling shares in this region.