We show that current clustering observations of quasars and luminous AGN can be explained by a merger model augmented by feedback from outflows. Using numerical simulations large enough to study clustering out to 25 comoving h -1 Mpc, we calculate correlation functions, biases, and correlation lengths as a function of AGN redshift and optical and X-ray luminosity. At optical wavelengths, our results match a wide range of current observations and generate predictions for future data sets. However, at redshifts z > 3, our results do not match the rapid evolution of the correlation length observed in the SDSS, a discrepancy that is at least partially due to differences in the scales probed by our simulation versus this survey. In the X-ray, our simulations produce correlation lengths similar to what is observed in the Chandra Deep Field (CDF) North, but not the significantly larger correlation length observed in the CDF South.