Analysis of the influence of heterogeneous expectations of economic agents on the stability of general equilibrium models with an open economy

The purpose of the publication is to study the influence of bounded rationality of agents on the ability of economic authorities to choose alternative policy rules that stabilize the dynamics of the relevant significant macroeconomic variables by simultaneously analyzing the entire range of model parameters. The scientific novelty lies in the fact that models with an open economy are analyzed, in which economic agents interact with the outside world. The article evaluates and compares behavioral neo-Keynesian models obtained with two alternative ways of introducing heterogeneous expectations. It is assumed that agents can be either short-sighted with a short-term forecast, or far-sighted forecasters. The difference does not matter when the agents have rational expectations, but it does matter when some of them form beliefs about the future according to some heuristics. Bayesian estimates based on the data of the Russian economy show that the behavioral model based on short-term forecasts is better in agreement with empirical data than the model based on long-term forecasts and even compared to the model with rational expectations of agents. Stability and stability analysis was carried out using a numerical procedure — Monte Carlo Filtration Mapping (MCF). This procedure generalizes and supplements the results obtained for a more limited set of parameters of low-dimensional models in which agents do not interact with the outside world. MCF-analysis shows that incorporating heterogeneous expectations reduces the stability and robustness of models. At the same time, a model based on predictors of long-term forecasting is less stable compared to models of short-term forecasting and with rational expectations of agents. An important result is a significant proportion of areas with unstable behavior of the studied models with heterogeneous expectations of agents, in which solutions are characterized by an explosive nature. With the help of Smirnov–Kolmogorov statistics, significant parameters were identified that determine the deterministic behavior of all analyzed models. An interesting result is: the response of the interest rate to changes in the output gap and changes in the real effective exchange rate does not affect the deterministic behavior of the models under study. All obtained results are confirmed by a posteriori Bayesian estimates for these parameters. The findings provide guidance to economists who study the processes of expectation formation with the help of microdata.