This paper examines the implementability of bail-in―a special bank resolution regime that reorganizes failed banks at the expense of creditors not taxpayers. In the first part of the paper, I consider the statutory bail-in case. I find theoretically the existence of multiple equilibria. If investors believe that the government will choose bailout (bail-in) in case of bank failure, the government indeed chooses bailout (bail-in). I also find that the depositor preference could be a commitment device by which the government can persuade the investors that it will choose bail-in. In the second part of the paper, I consider the contingent convertible bond (or CoCo bond), which is designed as a contractual bail-in tool. I find theoretically the same existence of multiple equilibria. Also, I show that if the government requires CoCo bond issuing banks to use rule-based triggers instead of discretionary triggers, the government can credibly persuade investors that it is tough. As a result, the equilibrium interest rates of CoCo bonds with discretionary triggers are lower than that with rule-based triggers. I confirm this theoretic prediction by conducting an empirical study. Using a CoCo bond issuance dataset from Moody’s and Bloomberg, I find that the interest rate gap is on average 2.5~3%p. The theoretic and empirical analyses imply that CoCo bonds with a discretionary trigger does not work effectively as a contractual bail-in tool.