White‐tailed deer (Odocoileus virginianus) browsing in forest regeneration sites can affect current and future stand structure and species composition. Removal of deer social units (localized management) has been proposed as a strategy to alleviate deer overbrowsing in forest systems. We conducted an experimental localized removal in a high‐density deer population in the central Appalachians of West Virginia, USA, during winter 2002. We removed 51 deer within a 1.1‐km2 area that encompassed 2 forest regeneration sites (14 ha). During the summer following removal, we detected decreases in distance from the removal area in 8 of 30 (26.7%) adult females having pretreatment mean telemetry locations <2.5 km from the center of the removal area. We measured browsing rates during the summers of 2001–2004 from forest regeneration sites to examine efficacy of localized management. Browsing rates declined annually in both removal and control areas, due in part to increased timber harvesting on the larger study site, suggesting that increasing forage availability may be more effective at reducing impacts on forest regeneration than localized reductions in deer populations. Three years after the initial removal, we removed an additional 31 deer from the original 1.1‐km2 removal area. Home range shifts of adjacent deer coupled with the large number of animals collected in the second removal suggests that localized management only produces temporary voids within high‐density deer herds. Localized management may not effectively reduce negative impacts of deer in areas of high deer density.