In this work, we use measurements of galaxy stellar mass and two-point angular correlation functions to constrain the stellar-to-halo mass ratios (SHMRs) of passive and star-forming galaxies at z ˜ 2-3, as identified in the Spitzer Matching Survey of the UltraVISTA ultra-deep Stripes. We adopt a sophisticated halo modeling approach to statistically divide our two populations into central and satellite galaxies. For central galaxies, we find that the normalization of the SHMR is greater for our passive population. Through the modeling of Λ cold dark matter halo mass accretion histories, we show that this can only arise if the conversion of baryons into stars was more efficient at higher redshifts and additionally that passive galaxies can be plausibly explained as residing in halos with the highest formation redshifts (i.e., those with the lowest accretion rates) at a given halo mass. At a fixed stellar mass, satellite galaxies occupy host halos with a greater mass than central galaxies, and we find further that the fraction of passive galaxies that are satellites is higher than for the combined population. This, and our derived satellite quenching timescales, combined with earlier estimates from the literature, support dynamical/environmental mechanisms as the dominant process for satellite quenching at z ≲ 3.