Standard rational choice theory assumes agents possess stable, exogenous preferences. James M. Buchanan rejected this axiom, positing an "Artifactual Man" defined by the capacity for self-modification and self-authorship. Interestingly, Buchanan explicitly recognized his affinities with French existentialism regarding this open-ended nature of the self. This paper reconstructs Buchanan’s project through the lens of existentialist ontology to address a central normative challenge in Public Choice and behavioral policy: how to evaluate welfare when the agent is a fluid project rather than a fixed datum. While the "New Paternalism" often attempts to reconstruct a latent rational self to justify interventions, I argue that Buchanan resolves the resulting ontological indeterminacy not by establishing a new fixed standard, but through a fractal architecture of self-binding.  By placing Buchanan in analytic dialogue with Jean-Paul Sartre and Albert Camus, I demonstrate that the "Ulysses contract" at the level of the individual (the Constitution of the I) provides the decision-theoretic template for the social contract at the level of the polity (the Constitution of the We). Aligning with Mario Rizzo and Glen Whitman’s concept of "inclusive rationality" and Malte Dold’s work on the capabilities of the choosing agent, this reframes the "biases" identified by behavioral economics not as errors to be corrected by a choice architect, but as the raw material for a self-imposed technology of aspiration and self-authorship — one that scales from the individual will to the collective constitution. Ultimately, Buchanan emerges as a theorist of democratic self-constitution who offers a rigorous alternative to technocratic paternalism, yet concludes that this project remains perpetually threatened by a twin dilemma: the danger of anarchy, understood as a plunge into unconstrained freedom, and the risk of Bad Faith, defined as the ossification of rules caused by forgetting our inalienable capacity for self-authorship.