Recent literature shows that firms and professional forecasters are rarely rational when forming their expectations or forecast about the future state of the economy. For instance, they tend to overreact to news when forming their individual forecasts, whereas aggregated (or consensus) forecasts reveal an under-reaction to news (Coibion and Gorodnichenko (2015), Bordalo, Gennaioli, and Shleifer (2018a)). Theories put forward to explain these puzzles are rational inattention, sticky information or, more recently, diagnostic expectations. In this paper, we show that forecasts of central banks can also be subject to non-rationalities. We use official forecasts from the Reserve Bank of New Zealand (RBNZ) and provide evidence that the RBNZ also tends to overreact to news in its forecasts. Our findings are robust to various variables, specifications and approaches, and could have important implications for (the identification of) monetary policy surprises (in the spirit of Romer and Romer (2000) and Miranda-Agrippino and Ricco (2018)) and other monetary policy relevant areas.
Seminar will be live-streamed, allowing for online audience participation (only available during the seminar)