Today the most commonly used electronic identity (eID) credentials offer low assurance, while the strongest and most reliable credentials are associated with the most rigid electronic identity management (eIDM) systems and are therefore rarely used outside of the framework defined by the credential issuing organisation.
In an ecosystem for digital identities as envisaged for the future, all these types of credentials are supposed to become interoperable though at different levels of assurance. In this work we argue that achieving such credential interoperability requires the rethinking of the way that credentials should be evaluated in such an ecosystem. We also point out several steps on how to achieve this in a bottom-up approach as required in rapidly changing technological environments and slowly changing and underdeveloped regulatory environments.