This paper analysis the statistical distribution of the markup in publicly-listed companies in the US for the years 1962-2019 and provides theoretical explanations for its origin and change. Fitting into the ongoing discussion about statistical rather than general equilibria in economics, this paper branches the discussed concepts further into the field of industrial organization. While the work is part of a general critique of the concepts of perfect competition, my approach is agnostic in finding the best theoretical fit for the empirical distribution. Settling for a two-component mixture distribution of a Log-Normal and Exponential distribution I provide two different, however consistent, explanations and theoretical concepts for the data-generating process in a winner/loser competition and a maximum entropy model. I conclude that changes to the markup over the last 59 years indicate that the competitive system of capitalism has not changed over time, but rather the possibilities and form of interaction of firms developed. The empirical results indicate a tendency of increasing markups while the competitive process decelerated, leading to an enhanced stickiness of the markup.