The correlation-factor approach to exchange and correlation is based on a simple real-space ansatz in which the exchange--correlation hole is written as \(\rho_{xc}(\mathbf r,u)=f_c(\mathbf r,u)\,\rho_x(\mathbf r,u)\) (6,7). Within this framework, the Becke--Roussel model (2) plays a central role because it provides a compact, non-oscillatory representation of the exchange hole that can be constrained through the on-top value, the curvature, the normalization, and the exchange energy per particle. In this way, it offers a flexible bridge between exact-exchange information and hole-based density-functional modeling, in the spirit of Becke's emphasis on real-space constructions of exchange and correlation, including his coordinate-space correlation model and his later real-space models for nondynamical, static, and strong correlation (1-5). In this contribution, we will give a short account of the correlation-factor model and of the way in which the Becke--Roussel construction underlies several of its developments, including extensions designed to incorporate exchange plus static correlation from multiconfigurational reference states. We will then turn to a broader viewpoint on correlation suggested by recent work (8): strong correlation and the derivative discontinuity can be understood within a common framework in terms of charge stiffness and the subdivision of Hilbert space into weakly communicating charge sectors. In this perspective, strong correlation corresponds to the suppression of collective charge fluctuations within a sector, whereas the derivative discontinuity appears as the non-analytic switching between competing sectors.
Finally, we will discuss the role of the complex domain in mean-field electronic-structure theory (9). An approximate exchange--correlation functional \(E_{\mathrm{xc}}[\rho]\) can in principle generate stationary densities that are complex. This suggests that approximate density functionals possess a broader variational landscape than is usually assumed, with additional branches of self-consistent solutions that may carry physical meaning.

1) A. D. Becke, J. Chem. Phys. 88, 1053 (1988).
2) A. D. Becke and M. R. Roussel, Phys. Rev. A 39, 3761 (1989).
3) A. D. Becke, J. Chem. Phys. 119, 2972 (2003).
4) A. D. Becke, J. Chem. Phys. 122, 064101 (2005).
5) A. D. Becke, J. Chem. Phys. 138, 074109 (2013).
6) R. Wang, Y. Zhou, and M. Ernzerhof, J. Chem. Phys. 150, 084107 (2019).
7) P.-O. Roy, E. Cuierrier, and M. Ernzerhof, J. Chem. Phys. 152, 211101 (2020).
8) M. Ernzerhof, Derivative discontinuity and strong correlation, a unified approach. In preparation (2026).
9) M. Ernzerhof, M. Loutis, P.-O. Roy, and D. Mayou, J. Chem. Phys. 163, 214105 (2025).