We present a sample of over 50 luminous field bulges (including ellipticals) found in the Groth Strip Survey (GSS), with 0.73< z < 1.04 and with bulge magnitudes I <= 23. The exponential disk light is removed via decomposition of HST images using GIM2D. We find that 85% of these bulges are nearly as red as local E/S0's and have a shallow slope and a small color dispersion in the color-luminosity relation, suggesting roughly coeval formation. The surface brightnesses of these bulges are about 1 mag higher than local bulges. These results are explained adopting a "drizzling" scenario where a metal-rich early formation is later polluted by small amounts of additional star formation. Almost all disks have the same or bluer colors than their accompanying bulges, regardless of the bulge-disk ratio and bulge luminosity, as expected from semi-analytic hierarchical galaxy formation models. We present evidence that the few blue bulge candidates are not likely to be genuine blue ellipticals or bulges. Our deeper, more extensive, and less disk-contaminated observations challenge prior claims that 30% to 50% of field bulges or ellipticals are in a blue, star-forming phase at z < 1. We conclude that field bulges and ellipticals at z ~ 1, like luminous early- type cluster galaxies at the same redshift, are already dominated by metal-rich, old stellar populations that have been fading from a formation epoch earlier than z ~ 1.5. (abridged)