We extend Fagin’s result on the equivalence between functional dependencies in relational databases and propositional Horn clauses. It is shown that this equivalence still holds for functional dependencies in databases that support complex values via nesting of records, lists, sets and multisets. The equivalence has several implications. Firstly, it extends a well-known result from relational databases to databases which are not in first normal form. Secondly, it characterises the implication of functional dependencies in complex-value databases in purely logical terms. The database designer can take advantage of this equivalence to reduce database design problems to simpler problems in propositional logic. An algorithm is presented for such an application. Furthermore, relational database design tools can be reused to solve problems for complex-value databases. Key Words: Logic in Databases, Functional Dependency, Implication Problem, Complex values, Horn clause