At Monetary Policy Radar, we are making a concerted effort to explore and assess various sources of alternative data to better understand the macroeconomy and what central banks might do next. This also included creating our own alternative data sets in 2024: a measure of core inflation that harnesses machine learning, and a hawks-doves assessment of all central bank speeches using large language models.
This new year has much more in store on the alt data front — we have plans to survey and create more useful series (including using extensive FT textual and traffic data). We will also routinely return to the alt data that has proved valuable in 2024. Here are some of the highlights and what to expect over 2025. Back in July, we compared the market-implied probabilities of central bank rate decisions versus those of superforecasters — a group of laypeople with special talent for forecasting. We found that superforecasters tended to produce better predictions. Looking at the forecasts of FOMC decisions between February 2023 and June 2024 revealed that the market had higher errors (measured using Brier scores — the sum of the squared difference between the forecasted probability and what ended up happening). We will soon run a thorough review of how superforecasters performed in the second half of 2024, including for ECB, BoE and BoJ decisions, but in the meantime we checked how superforecasters did versus the market in December. The market was a touch better on the Fed and ECB decisions to cut, as well as the BoJ vote to hold, but for the BoE the superforecasters were much stronger in predicting a hold. The market put around a 65 per cent chance of an MPC decision to hold rates, whereas superforecasters were far more certain, nearly unanimously expecting Bank rate to remain at 4.75 per cen
Nederland Laatste Nieuws, Nederland Headlines
Similar News:Je kunt ook nieuwsberichten lezen die vergelijkbaar zijn met deze die we uit andere nieuwsbronnen hebben verzameld.
Bron: OilandEnergy - 🏆 34. / 68 Lees verder »