George Evans has investigated a wide range of theoretical and empirical topics in macroeconomics including tests for speculative bubbles, the effect of sectoral imbalance on unemployment, the decomposition of aggregate output into trend and cycle, and models of endogenous fluctuations. In his work on "rational bubbles" he has examined the statistical evidence for the presence of bubbles in foreign exchange rates and in stock prices. Econometric work on business cycles showed how to extract the cyclical component of GDP using multivariate forecasting methods.

Professor Evans is best known for his research on expectational stability and learning in stochastic, dynamic models, including settings with multiple equilibria. In these models, economic agents are assumed to have bounded rationality in making forecasts -- e.g. to use simple or sophisticated forecast rules, which are updated over time using observed data. Because agents learn and adapt to forecast errors, they may, in the long run, approximate full rationality. Adaptive learning techniques can be used, in particular, to determine when the economy can become trapped into cycles, "sunspot equilibria," hyperinflationary paths or deflationary spirals, and how macroeconomic policy can steer the economy away from these inefficient outcomes.

An ongoing project has looked at fiscal and monetary policy when there are pessimistic output and inflation expectations. If the economy is subject to a large negative expectation shock, the economy can be trapped in a deflation or stagnation trap region under adaptive learning, even though the steady state targeted by policymakers is locally stable under learning. This research shows that careful design of policy, and in particular a suitable fiscal stimulus, may be needed to drive the economy out of the stagnation trap and back to the targeted steady state.