Why AI won’t take your fixed income job

CP+ Why AI won't take your fixed income job article supporting image

A conversation with ChatGPT.

There's been so much talk about how AI could replace human workers across virtually every industry sector, from telecoms to retail to trucking and beyond. Lately, following the proliferation of generative AI, the conversation has turned to us, and whether AI might affect fixed income jobs. So, I figured I should have a conversation with ChatGPT, the ubiquitous AI chatbot, to see what’s in store for my fixed income career. ChatGPT is a bit long winded, so I’ll just provide the interesting bits.

Me: Will AI take my fixed income job?

ChatGPT: AI is best suited for tasks that involve repetitive, rule-based processes, such as data analysis, pattern recognition, and decision-making. These are tasks that can be automated to increase efficiency and accuracy, potentially reducing the need for human intervention.

However, there are other aspects of fixed income jobs that require human skills and judgment, such as relationship building with clients, communication, and problem-solving.

It sounds like AI can’t do soft skills and problem solving, which is good for me since that’s the fun part of my job. Perhaps AI is more of a helper than a replacement—something that can handle the time consuming or mundane parts of our jobs, freeing up time for us to innovate and collaborate with clients.

Me: Which fixed income tasks can AI help automate?

ChatGPT:

1. Portfolio Management: AI can assist with portfolio management tasks, such as identifying investment opportunities, conducting risk assessments, and optimizing portfolio allocations based on market conditions.

2. Risk Management: AI can analyze data related to fixed income securities and identify potential risks that could impact investment returns.

3. Trading: AI can automate fixed income trading by using algorithms to analyze market data and identify trading opportunities.

4. Credit Analysis: AI can assist with credit analysis by analyzing financial statements, credit ratings, and other relevant data to assess the creditworthiness of issuers of fixed income securities.

5. Compliance Monitoring: AI can assist with compliance monitoring by analyzing transactions and other data to identify potential regulatory issues or violations.

I’m not sure if ChatGPT is bragging or what, but that’s an impressive list. At MarketAxess, I work on CP™+, our AI-driven real-time fixed income pricing, which is used in the first three areas. Large dealers and algo funds use CP+ data in their own AI models to optimize trading strategies. Similarly, clients are creating AI models with our CP+ data to identify patterns in the market to inform their investment decisions and track portfolio risk in real-time. With more than 50 million levels per day, price updates as fast as every 15 seconds and 100% coverage in its markets, CP+ is taking these workflows to new levels of efficiency and scale.

My fixed income career may still have some legs! But there are a lot of choices for fixed income data out there, so what’s the difference? One would think that better quality data would lead to better AI results…

Me: Does AI need good data to work?

ChatGPT: Yes, AI needs good data to work effectively. In fact, the quality of the data used to train an AI model is one of the most important factors in determining the accuracy and performance of the model. If the data used to train an AI model is biased, incomplete, or otherwise flawed, the model's predictions and decisions may be inaccurate or unreliable. On the other hand, if the data is of high quality, the AI model will be better equipped to learn the patterns and relationships needed to make accurate predictions and decisions.

Our data products leverage unique and quantifiable trading insights from our platform (think depth of book). This data eclipses what you see in trades alone. In Europe, we also use our TraX market data as a key input for CP+. Unique data + strong models (yes, we use AI models too) = high quality data products. The numbers support it too: last year more than 20% of trades on the MarketAxess platform used CP+ Auto-X™ trading protocols.

While it dawned on me that ChatGPT (a.k.a. the AI talking robot) has some major conflicts of interest in this discussion, I think the conclusion is irrefutable—AI is driving greater efficiency in the fixed income market. But it’s not going to take your job, or mine, just yet.