Mechanism design focuses on setting up contractual space, understood as games played by parties, in such a way that desired outcome can be sustained as Nash equilibrium. This implies that some desired outcomes are hard to sustain due to lack of possibility of enforcing contracts that will guarantee the desired outcome. Some outcomes can be sustained in case of repeated interaction between parties and possible "punishment" in the future.

Distributed ledger technology allows for setting up contracts between parties that will be executed automatically upon meeting specific requirements. In particular, smart contracts are digital contracts that allow terms contingent on a decentralized consensus which is self-enforcing and tamper-proof through automated/algorithmic execution. Such form of executing a smart contract enlarges contracting space and furthermore expands the set of sustainable equilibria. Unless there is an easy way of finding out whether conditions were met to execute the contract, its execution will depend on third parties reaching consensus.

In the paper, we focus on contracting in a principal-agent model where one party (agent) acts on behalf or makes decisions that impact another party (principal). The key issue is that those two parties may have different interests or asymmetric information. The outcome may be principal not entering the relationship with an agent or incurring additional cost necessary to ensure that agent will act in the best interest of the principal (agency costs). Our aim in this paper is to show how the introduction of smart contracts will impact the relationship between principal and agent. We will change principal-agent model to accommodate the existence of smart contracts and use the game-theoretic framework to demonstrate their relevance in the stated context.

First, we demonstrate that smart contracts may considerably lower agency costs and allow for contracting in situations where, in the absence of smart contracts, the principal would not enter a contractual arrangement with an agent. Then we discuss how a process of generating decentralized consensus may impact the outcome. We focus on two scenarios: consensus generated by several entities with limited credibility/knowledge and consensus generated by a few credible parties. The first setting is closed to consensus mechanism in Ethereum network while the latter is closer to the way R3 Corda network operates.

We conclude with a discussion how consensus mechanism may further impact principal-agent setting, what may be regulatory implications, and how smart contracts may challenge existing mechanism design models.