Firms statistically discriminate (redline) against job candidates based on where they live. We conducted a correspondence test by sending three identical fictitious resumes to every nonprofessional job offer posted in two main job vacancy newspapers in Bogota. The only difference between the resumes was the residential address in which the applicants lived. Two of the three resumes sent in each trio were located at the same commuting time (and geographical distance) from the job, but one resided in a low-crime neighborhood and the other in a high-crime neighborhood. The third resume was for a fictitious individual located in a low-crime neighborhood that is further away (longer commuting time and greater distance). Our experimental design allows us to explore whether employers discriminate against potential employees based on where they live, and if they do, which mechanisms are behind their discriminatory preferences. Building on the urban economics literature, we test two potential mechanisms: statistical discrimination due to negative signaling neighborhood effects and statistical discrimination based on commuting time to work. If any of these holds, we would expect employers to offer interviews to job applicants who reside in deprived or distant neighborhoods less often. We find that employers statistically discriminate (redline) based on commuting time to work. In particular, living one hour away from the vacancy reduces the callback rate by 32 percent while holding the attributes of the place of residence constant. We did not find evidence that employers respond to negative signaling effects or engages in taste based-discrimination.