Shifting the Fermi energy in solids by doping, defect formation or gating generally results in changes in the charge density distribution, which reflect the ability of the bonding pattern in solids to adjust to such external perturbations. In the traditional textbook chemistry, such changes are often described by the formal oxidation states (FOS) whereby a single atom type is presumed to absorb the full burden of the perturbation (change in charge) of the whole compound. In the present paper, we analyze the changes in the position-dependence charge density due to shifts of the Fermi energy on a general physical basis, comparing with the view of the FOS picture. We use the halide perovskites CsSnX 3 (X=F, Cl, Br, I) as examples for studying the general principle. When the solar absorber CsSnI 3 (termed 113) loses 50 % of its Sn atoms, thereby forming the ordered vacancy compound Cs 2 SnI 6 (termed 216), the Sn is said in the FOS picture to change from Sn 2+ to Sn 4+ . To understand the electronic properties of these two groups we studied the 113/216 compound pairs CsSnCl 3 /Cs 2 SnCl 6 , CsSnBr 3 /Cs 2 SnBr 6 , and CsSnI 3 /Cs 2 SnI 6 , complementing them by CsSnF 3 /Cs 2 SnF 6 in the hypothetical cubic structure for completing the chemical trends. These materials were also synthesized by chemical routes and characterized by X-ray diffraction, 119 Sn-Mössbauer spectroscopy and K-edge X-ray Absorption Spectroscopy. We find that indeed in going from 113 to 216 (equivalent to the introduction of two holes per unit) there is a decrease in the s charge on Sn, in agreement with the FOS picture. However, at the same time, we observe an increase of the p charge via downshift of the otherwise unoccupied p level, an effect that tends to replenish much of the lost s charge. At the end, the change in the charge on the Sn site as a result of adding two holes to the unit cell is rather small. This effect is theoretically explained as a 'self-regulating response ' [H. Raebiger, S. Lany, and A. Zunger, Nature 453, 763 (2008)] whereby the system re-hybridizes to minimize the effect of the charge perturbation created by vacancy formation. Rather than having a single preselected atom (here, Sn) absorb the full brunt of the perturbation producing two holes/cell, we find that the holes are distributed in a complex pattern throughout the octahedral systems of X 6 ligands, forming hole orbitals with some specific symmetries. This clarifies the relation between FOS and charge transfer that can be applied to a wide variety of materials.2