Non-monotonic Security Protocols and Failures in Financial Intermediation

Security Protocols as we know them are monotonic: valid security evidence (e.g. commitments, signatures, etc.) accrues over protocol steps performed by honest parties. Once’s Alice proved she has an authentication token, got some digital cash, or casted a correct vote, the protocol can move on to validate Bob’s evidence. Alice’s evidence is never invalidated by honest Bob’s actions (as long as she stays honest and is not compromised). Protocol failures only stems from design failures or wrong assumptions (such as Alice’s own misbehavior). Security protocol designers can then focus on preventing or detecting misbehavior (e.g. double spending or double voting). We argue that general financial intermediation (e.g. Market Exchanges) requires us to consider new form of failures where honest Bob’s actions can make honest good standing. Security protocols must be able to deal with non-monotonic security and new types of failures that stems from rational behavior of honest agents finding thems...