This paper examines the impact of the Federal Reserve’s communication on short-term inflation forecasts. Following the Federal Reserve’s adoption of an explicit inflation target in 2012, SPF respondents’ four-quarter-ahead inflation forecasts display two notable behavioral shifts: (1) increased confidence in their beliefs and (2) less overreactive forecasts to news, aligning more closely with rational expectations. A key factor driving these behavioral shifts is the reduction in uncertainty about trend inflation. To support this claim, I propose a parsimonious inflation expectations model with smooth diagnostic expectations. The model captures changes in both the first and second moments of individuals' predictive densities, providing an explanation for the decrease in short-term forecast disagreement. In line with this mechanism, incorporating the expectations formation framework into the New Keynesian model demonstrates that the Fed’s target announcement contributes to the stabilization of realized inflation, mitigating agents’ overreactive belief updating.
In this study, I use a DSGE model to analyze how economic agents with rational inattention respond to varying levels of economic volatility. Specifically, I compare agent behavior across two periods: the Pre-Great Moderation (1960Q1–1983Q4), characterized by high volatility, and the Great Moderation (1984Q1–2007Q2), marked by lower volatility. I find that economic agents allocate more attention to changes in the economy during periods of high volatility, validating the rational inattention framework across different economic environments. Specifically, households and firms adjust their decisions more quickly when the economy becomes more volatile, which in turn leads to faster behavioral adjustments during the pre-Great Moderation period.
Political partisanship shapes the economic expectations of U.S. households. To examine how electoral outcomes influence forecasts of inflation and unemployment, we conducted a two-wave survey, with the second wave on the morning of November 6, 2024, immediately after the presidential election. Democrats revised their expectations more pessimistically, raising projected inflation and unemployment. Republicans, despite largely anticipating a Trump victory, still revised forecasts more optimistically, often projecting deflation, which most respondents view as favorable. Forecast disagreement narrowed among Republicans but widened among Democrats. To support the empirics, we present an expectations-formation framework where perceived signals are influenced by political bias. We simulate respondent-like personas with a constrained LLM to assess external validity. It recovers the pre-election partisan gap and rapid updating but misses Republicans’ downward revision and dispersion.