In order to find an explanation for the radiative quiescence of supermassive black holes in the local universe, the most accurate estimates for a sample of nearby galaxies are collected for the mass of a central black hole (MBH), the nuclear X-ray luminosity LX,nuc, and the circumnuclear hot gas density and temperature, by using Chandra data. The nuclear X-ray luminosity LX,nuc varies by ~3 orders of magnitude and does not show a relationship with MBH or with the Bondi mass accretion rate B; LX,nuc is always much lower than expected if B ends in a standard accretion disk with high radiative efficiency (this instead can be the case of the active nucleus of Cen A). Radiatively inefficient accretion as in the standard advection-dominated accretion flow (ADAF) modeling may explain the low luminosities of a few cases; for others, the predicted luminosity is still too high, and, in terms of Eddington-scaled quantities, it is increasingly higher than that observed for increasing B. Variants of the simple radiatively inefficient scenario including outflow and convection may reproduce the low emission levels observed, since the amount of matter actually accreted is reduced considerably. However, the most promising scenario includes feedback from accretion on the surrounding gas; this has the important advantages of naturally explaining the observed lack of relationship among LX,nuc, MBH, and B, and evading the problem of the fate of the material accumulating in the central galactic regions over cosmological times.