Sentiment Heatmap

The Sentiment Heatmap gives uses the most granular view of dovish and hawkish sentiment of the FOMC. Each box represents a currently active FOMC board member.

• The size of each box represents that specific governor’s relative influence over the committee.
• The color of each box represents the relative level of hawkish or dovish sentiment of that governor which can be visualized with the arrow that appears on the hawkish / dovish scale at the bottom when mousing over a specific governor.
• Mousing over a governor provides detailed statistics about each governor.

Clicking on a specific governor allows the user to drill down into the specific topics most commonly associated with that particular governor as well as the level of dovish and hawkish sentiment associated with that topic in the context of the specific governor.

• The size of each topic box represents a weighted value for how important that topic is in relation to the others within the context of the specific governor.
• The size of each topic box represents a weighted value for how important that topic is in relation to the others within the context of the specific governor.
• The color of each topic box demonstrates how dovish or hawkish the sentiment is in relation to that topic when viewed through the context of the governor in question.

Crossover Likelihood Heatmap

The Crossover Likelihood Heatmap helps users focus their attention on the specific Fed governors that are likely to be the most interesting in terms of changing their decisions on the committee. Just like “swing states” absorb most of the attention during a presidential election, similarly, the governors with the highest Crossover Likelihood are the ones most interesting to the market because they are the least set in their ways. The Crossover Likelihood view shows us how close each governor is to the middle-line between dovish and hawkish. The closer to the middle-line, the more likely he or she will switch from dovish to hawkish, or vice versa, and correspondingly change their vote on rates. The size of each box represents the specific governor’s relative influence over the committee.

Crossover Likelihood helps filter down a user’s analysis to the specific governors that are most likely to cross from one side of the spectrum to the other.

Sentiment over time [Temporal Chart]

Our Sentiment Over Time chart allows users to see the changes in the Sentiment Heatmap over time for both individual governors, and as a weighted average of all governors. Utilizing this view, users can see the trends for specific governors over time, whether they have crossed from dovish to hawkish, or vice versa, and how often they cross.

Clicking on a specific governor in the legend, we can similarly see a multi-line chart of the topics associated with that governor over time so that we can closely see which topics are driving hawkish or dovish sentiment for a specific governor.

Hawkish / Dovish Sentiment vs Inflation + Employment Benchmark

The charts in this module allow the user to see a more aggregated view of hawkish and dovish sentiment than what the Fed governor heatmap view displays. Each chart has a benchmark (black line) which defines the context through which individual articles of unstructured data are analyzed. The benchmark is built up from three different combinations of data sets from the BLS (Bureau of Labor Statistics). The benchmark can represent 1) Inflation Only, 2) Employment Only, or 3) a combined benchmark of both inflation and employment.

The user can select which benchmark view to compare against. The system will display the level of hawkish or dovish chatter of the overall market within the context of that benchmark (Market Chatter - purple line). Additionally, the system will display the level of hawkish or dovish chatter emanating solely from official FOMC communications within the context of the selected benchmark (FOMC Sentiment - green line). Comparing these two lines, the user can determine when market chatter is in sync with official FOMC communications and vice versa. When there are larger gaps, the market is more likely to be taken by surprise by what the Fed says or does.

Accrete Synthetic Fed Futures Index vs Fed Futures & T-Bill

The chart in this module primarily displays the Accrete Synthetic Fed Funds Futures yield (in green), which continuously provides users with more accurate interest rate forecasts by estimating a fair value of the 30 Day Fed Funds Futures yield (in black). This fair value is based on MACI’s continuously evolving interpretation of official FOMC Communications, FOMC Governor Sentiment & Influence and FOMC Governor Crossover Likelihood. In addition to providing users with more accurate interest rate forecasts, the index forecasts FOMC Policy Decisions along with real time probability scores. The higher the probability score, the less likely the market is to be surprised by the FOMC, and vice versa.

To demonstrate one use of the synthetic future, also displayed below in the chart is a historical and live trading system which generates buy and sell signals based on the USGG3M 13 week generic US Govt. 3-month yield (in blue). The bubbles indicate when to trade, and the real-time probability score indicates how strong the signal is and thus how aggressively to size the specific position.

• The black line is the 30 Day Fed Fund Futures yield for the upcoming month.
• The blue line is the T-bill yield, not its actual price.
• The green line is the Accrete Synthetic Fed Futures yield for the upcoming month.
• The green and red bubbles indicate specific entry and exit points on the T-bill yield.

Awareness Cloud

Since our system analyzes thousands of articles daily, there is no way for a user to possibly read everything the system finds. Our awareness cloud helps give users a taste of the most relevant articles from within the universe of all relevant articles analyzed. The awareness cloud displays the articles it finds most relevant in a scrolling ticker. On that ticker, the user can click on the headline to read the original article. To save time, the user can also click on the “summary” button which uses our system’s intelligence to summarize the article into roughly one paragraph. These summaries allow the user to easily read the key points of many articles in the same amount of time that would be spent reading just one article.

Relevant Articles Analyzed

The bar chart in this module shows a running total of how many articles were analyzed each day over the last 5 days, including today. The chart gives a good sense of the level of breadth the system covers while also showing spikes in overall content creation activity within the context of inflation and employment.

Expert Profiles

Our expert profiles table gives the user a sense of which experts contributed to the creation of the ontologies that provide the contextual seed for the system and their respective levels of contribution. Importantly, each of these experts was sourced, highly vetted, and formally engaged by Accrete on an independent basis. The importance of sourcing these experts independently is critical because only by sourcing them independently and putting them to work on the same task without any knowledge of each other can we find the unbiased ground truths that are necessary to ensure the overall system is bias-free. We always strive to find experts from a diverse set of backgrounds, to bring a breadth of perspective into the creation of each product.