Three essays on money and banking

The dissertation consists of three studies on money and banking in the presence of uncertainty.;The second paper is an experimental study about depositors' behavior under a demand deposit contract when they face uncertainty over other depositors' actions; and investigates whether bank runs can occur as the result of pure coordination failures. It is found that bank runs can occur as a result of pure coordination failures, but only when coordination is difficult. I compare the experimental results and the simulation results from a learning algorithm modified from Temzelides (1997), and find that learning offers a good approximation to observed lab behavior.;In the third paper, agents face uncertainty over future preferences. The paper takes the mechanism design approach and studies the essentiality of multiple currencies - which act as substitute for the missing record-keeping technology - in the presence of limited commitment and private information about the realization of preferences. When money balances are concealable, a single money is sufficient to solve the problem of limited commitment, and can deal with the private information problem if agents are patient enough; in which case, money balances serve as a preference signalling device. When agents are sufficiently impatient, a second money is essential and allows agents to signal their preferences by holding different monetary portfolios all giving rise to the same total money balances.;Keywords. Money; mechanism design; bank runs; banking crises; experiments; learning.;In the first paper, agents face uncertain future liquidity needs and the bank is formed to provide liquidity insurance to depositors. The bank holds cash reserves to meet depositors' liquidity needs and as an insurance against uncertain return on the bank's assets. The paper analyzes the effect of inflation on banking crises. The main result is that when the bank has access to a stable foreign currency, inflation has a threshold effect on the incidence of banking crises: higher inflation reduces the likelihood of crises when inflation is below the threshold; the reverse occurs when inflation exceeds the threshold. This result appears to be broadly consistent with available evidence.