Introduction This paper attempts a rehabilitation of the kinked demand approach (Hall and Hitch, 1939; Sweezy, 1939) in explaining collusion among producers of homogeneous goods. The kinked demand approach is one of the most known and still used approaches to deal with collusion (Maskin and Tirole, 1988; Sen, 2004; Currarini and Marini, 2011; Garrod, 2012), although it has been strongly criticized because of its behavioural assumption of asymmetric reaction of competitors to price changes (Stigler, 1947; Primeaux and Bomball, 1974; Bhaskar et al., 1991). However, this assumption is not necessary for having a kink: actually, Salop (1979) shows that kinked demand curves can exist in Hotelling models of product differentiation with Nash-like firms. However, neither Salop nor successive scholars have investigated the implications of this result for collusion.