With the recent recovery of large carnivores in Europe, preventive measures to protect livestock are on the rise. Fences that exclude carnivores from grazing areas have been proven as effective, but they can be costly as well as posing a barrier for wildlife. We studied the effect of exclosures of > 10 km2 to protect sheep Ovis aries on the distribution and density of moose Alces alces using fecal pellet group counts in two study areas in southeastern Norway. During the summer grazing season, the fences were powered. Outside of the grazing season, one exclosure remained fenced while the other fence was demounted. This quasi‐experimental setting allowed us to investigate whether fences had a barrier effect for moose, and/or whether moose density was affected by interactions with sheep (competition or facilitation) or large carnivores (refuge hypothesis). During winter, moose pellet group density was about equal inside and outside of the exclosure with demounted wire strands, but less than half inside the permanently fenced exclosure compared to outside, indicating a potential fragmentation effect of the fence. During the grazing season, when wire strands were powered, moose pellet group density was equal or doubled inside as compared to outside both exclosures. Moose may have sought refuge from large carnivores inside the fences. Fecal pellet group densities of moose and sheep inside the fence were neither positively (facilitation) nor negatively (competition) correlated. However, moose used young forest, the most used habitat type by sheep, to a lesser extent inside than outside of the exclosures, maybe due to interference competition. Our study demonstrates that livestock protection fences can have an impact on more than the targeted wildlife species. To understand the mechanisms behind direct and indirect effects of fences, monitoring the movement and survival of individuals by means of GPS and camera traps would be needed.