The rapid advancement in the development of non-fullerene acceptors has led to single-junction polymer solar cells with efficiencies over 18%. Even with these novel acceptor materials, the choice of the donor polymer remains important. Tuning of the donor and acceptor compatibility in terms of absorption, frontier orbital energy levels, mixing enthalpy and charge carrier mobility is routinely performed by side chain variation. Fluorination presents an additional powerful approach to optimize these parameters. Although significantly less studied, chlorination can give rise to similar effects, while donor-acceptor phase separation due to fluorophobic interactions is less of an issue. Moreover, from a material synthesis point of view, the introduction of chlorine groups is in many cases much more straightforward. In this work, we present a series of push-pull type benzo[1,2-b:4,5-b']dithiophene-alt-quinoxaline donor polymers and compare the behavior of the non-halogenated, fluorinated and chlorinated derivatives in polymer solar cells when combined with small molecule and polymer type non-fullerene acceptors. The solar cell efficiencies vary from 2.4 to 8.4%, elucidating the large impact of these small structural variations. Best results are achieved for the chlorinated donor polymer, affording a high open-circuit voltage, balanced charge carrier mobilities and favorable donor-acceptor interactions. Combined with the easier synthesis of chlorinated materials, this suggests that more emphasis should be put on chlorination as a valuable approach to tune the properties of organic semiconductors for solar cell blends (and other optoelectronic applications).