An objective function for simulation based inference on exchange rate data

摘要: The assessment of models financial market behaviour requires evaluation tools. When complexity hinders a direct estimation approach, e.g., for agent based microsimulation models, simulation estimators might provide an alternative. In order to apply such techniques, objective function is required, which should be on robust statistics the time series under consideration. Based identification foreign exchange rate in previous research, derived. This takes into account stylized facts about unconditional distribution returns and properties conditional distribution, particular, autoregressive heteroscedasticity long memory. A bootstrap procedure used obtain estimate variance-covariance matrix different moments included function, as base weighting matrix. Finally, are analyzed two market, simple GARCH-model stochastic volatility model using DM/US-$ benchmark. It also discussed how results inference purposes.