‘Little red dots’ cannot reside in the same dark matter haloes as comparably luminous unobscured quasars

ABSTRACT The JWST has uncovered a new population of candidate broad-line active galactic nucleus (AGN) emerging in the early Universe, named ‘little red dots’ (LRDs) because of their compactness and red colours at optical wavelengths. LRDs appear to be surprisingly abundant (${\approx} 10^{-5} \, {\rm cMpc}^{-3}$) given that their inferred bolometric luminosities largely overlap with those of the ultraviolet (UV)-luminous quasars identified at high z in wide-field spectroscopic surveys. In this work, we investigate how the population of LRDs and/or other UV-obscured AGN relates to the one of unobscured, UV-selected quasars. By comparing their number densities, we infer an extremely large and rapidly evolving obscured:unobscured ratio, ranging from ${\approx} 20{:}1$ at $z\approx 4$ to ${\approx} 2300{:}1$ at $z\approx 7$, and possibly extending out to very high (${\approx} 10^{47}\, {\rm erg}\, {\rm s}^{-1}$) bolometric luminosities. This large obscured:unobscured ratio is incompatible with the UV-luminous duty cycle measured for unobscured quasars at $z\approx 4\!-\!6$, suggesting that LRDs are too abundant to be hosted by the same haloes as unobscured quasars. This implies that either (a) the bolometric luminosities of LRDs are strongly overestimated or (b) LRDs follow different scaling relations than those of UV-selected quasars, representing a new population of accreting supermassive black holes emerging in the early Universe. A direct comparison between the clustering of LRDs and that of faint UV-selected quasars will ultimately confirm these findings and shed light on key properties of LRDs such as their host mass distribution and duty cycle. We provide a mock analysis for the clustering of LRDs and show that it is feasible with current and upcoming JWST surveys.