Utilising AI
The EFG Hawkishness Indicator seeks to gauge policy bias at Federal Open Market Committee (FOMC) meetings. The latest update provides an innovation on the previous methodology which counted the number of times key words such as “elevated”, “robust”, “high” and “strong” were repeated in published meeting minutes.1
It does so by utilising artificial intelligence (AI). Instead of searching for specific words, the entire minutes of FOMC meetings are assessed and scored on a scale, with a higher score representing a more hawkish policy bias. While the score itself is arbitrary and should not be overemphasised, the fact that the scoring system is internally consistent means that changes over time are worth looking at more closely.
To investigate further, we analysed the minutes of each Fed meeting over the last ten years and compared the EFG Hawkishness Indicator score to the Fed’s preferred inflation measure, core personal consumption expenditure (PCE) inflation, and the Fed funds rate over time. The results are shown in the chart below.