By leveraging the power of advanced natural language processing (NLP) applied to real-time media, MarketPsych Analytics enables investors to monitor and respond to emerging risks.

Risk Management Insights

There are hundreds of types of risk that exist; credit, supply chain, liquidity, reputational, regulatory… these are often identified and communicated through news & social media channels before all market participants have had the opportunity to exit. MarketPsych monitors information flow and scores risks themes for each security. For instance, see how our ManagementTrust score dropped for Nissan Motor Co after Carlos Ghosn was arrested on suspicion of embezzling company funds (blue) versus the share price (black).

Figure: Trust towards the Management team of Nissan dropping precipitously as it is reported that Carlos Ghosn has embezzled the firm's funds. Companies ranked poorly by Management Trust consistently underperform their respective markets.

Risk-on / Risk-off Research

For equities, tracking sentiment alone can be enough to detect significant events. Our client NN Investment Partners ($300 billion AUM, now owned by Goldman Sachs) reported that media sentiment drops prior to market corrections for major indices, enhancing their risk-on, risk-off models.

A simple depiction of this relationship is seen in graphic above, where two moving averages of market sentiment are superimposed on a price chart of the Russell 1000 from Jan 2023 to Aug 2024. When the shorter-term (30-day) sentiment average drops below the longer term average (90-days), the area between the averages is shaded pink. It is shaded blue when the short-term sentiment is rising faster than the longer-term.

Additionally, according to MarketPsych research, large single-day drops in sentiment associated with major equity indices consistently precede a next-day price decline.

Regime Studies

Identifying periods of severe market stress can allow investors to de-risk by moving portfolio holdings into safer asset classes (risk-on, risk-off trading). In another study, using a Gaussian Mixture Model framework applied to US stock market sentiment, regimes are identified in which the market is likely to underperform, and the model switches to a risk-reduced asset such as a bond index (AGG), resulting in much more stable performance for an index tracker.

Figure: The above risk-related research is among the 100+ academic and internal whitepapers hosted on the MarketPsych website, many with python code available, and available to all data subscribers.