Essays on Macroeconomic Policies: Experiments and Simulations

The first paper of this thesis (Chapter 2) explores how expectations of inflation and output are influenced by central bank forward guidance within a learning--to--forecast laboratory macroeconomic environment. Subjects are incentivized to forecast the output gap and inflation. An automated central bank forms projections about the economy assuming subjects form expectations following the REE solution. The central bank communicates output and/or inflation projections, interest rate projections, or no information. Communicating about future output or inflation generally reduces the degree to which subjects rely on lagged information and increases their reliance on the REE solution. Interest rate projections, by contrast, do not significantly alter subjects' forecast accuracy or disagreement. Central bank credibility significantly decreases when the central bank makes larger forecast errors when providing forward guidance about either output and inflation, but not when they provide a dual projection. Our findings suggest that expectations are best coordinated and stabilized by communicating output and inflation forecasts simultaneously. The second paper of this thesis (Chapter 3) evaluates the central bank communication of its future inflation and output expectations to reduce economic variations in the event of a demand or cost--push shock. Four communication strategies are tested: no communication, communicating output, communicating inflation, and communicating output and inflation. Two Taylor rules are considered: (a) central bank interest rate responds to inflation and output (flexible inflation targeting [IT]), and (b) central bank responds only to inflation (strict IT). We find that with a demand shock, communicating future inflation reduces output variations and increases inflation and interest rate variations; however, communicating output stabilizes inflation and interest rates and destabilizes output (the interest rate rule did not matter qualitatively). With a cost--push shock, communicating future output decreases inflation and interest rate variability, irrespective of Taylor rule qualitatively. In order to stabilize output, a central bank should be uncommunicative under flexible IT but should communicate future inflation under strict IT. The third paper of this thesis (Chapter 4) studies the effect of the degree of information observability on bank runs in a sequential laboratory environment. We conduct an experiment with ten depositors in a queue who are randomly ranked to submit their decisions. The depositors decide between withdrawing their deposit or waiting and leaving their deposit in a common experimental bank. Two treatments are considered: a sequential high-information treatment, and a sequential low--information treatment. In both treatments, depositors who are in the front of the queue tend to withdraw more than those at the back of the queue. Moreover, depositors are responsive to the information about the preceding withdrawals within a period. We find that in the sequential high--information treatment the possibility of observing preceding withdrawals increases the likelihood of bank runs compared to a sequential-low information treatment.