Podcast Summary
AI and political events: Stay agile and informed in the face of constant change from technological advancements and political shifts, collaborate with AI in fund management, and stay updated on geopolitical risks.
The future holds uncertainty and potential opportunities, especially in the financial markets. With the increasing role of artificial intelligence (AI) in various industries, even fund management, it's essential for human investors to adapt and collaborate with technology. However, it's crucial not to overlook other factors that can significantly impact markets, such as political events. For instance, recent developments in French politics have led to significant market movements, highlighting the importance of staying informed about geopolitical risks. Overall, the key takeaway is to remain agile and informed in the face of constant change, whether it's from technological advancements or political shifts. PIGIM, as a leading global asset manager, can help investors navigate these challenges with deep expertise and a focus on long-term returns.
Financial Analysis with AI: Large language models like ChatGPT can make stock earnings predictions with accuracy surpassing human analysts, but they only provide predictions for growth direction and not buy/sell recommendations.
A large language model like ChatGPT, when given financial data and statements, can make predictions about stock earnings with some degree of accuracy, surpassing human analysts' predictions in some cases. The University of Chicago business professors, in an experiment, fed the model anonymized balance sheets and income statements, and asked it to analyze them in a financial analyst-like manner. The model's predictions for earnings growth and direction, though not perfect, were better than those of human analysts in the past. The professors even built long-short portfolios based on the model's most certain predictions, which outperformed the broad stock market index. However, it's important to note that the model's predictions were limited to "up a lot," "up by a little," "down a lot," and "down by a little," and it did not provide buy, sell, or hold recommendations. Overall, this experiment suggests that AI, specifically language models, could potentially offer valuable insights in financial analysis.
Language Model in Finance: A study showed that a large language model like ChatGPT outperformed the median stock picker in predicting earnings direction with human-like narrative analysis, offering potential for valuable insights in financial markets.
A study using a large language model like ChatGPT, with some financial information and light prompting, was able to outperform the median stock picker in predicting earnings direction with human-like narrative analysis. The authors were particularly impressed by the model's ability to provide economic logic explanations for financial statement changes, leading to more accurate predictions. This observation, while not groundbreaking, suggests potential for large language models in financial markets. The model outperformed the market by 0.37% a month in backtests, but it's important to note that this was before trading costs were factored in. This level of outperformance, if sustained, could result in significant returns over time. The study is a proof of concept, not a definitive solution, but it indicates that these models might offer valuable insights in financial markets.
Small-cap outperformance in passive index funds: Small-cap stocks may outperform larger ones in passive index funds, but this phenomenon might not always hold true in real life due to factors like trading costs, market efficiency, and model competition. The reasons behind this are unclear.
Small-cap stocks may outperform larger ones when passively managed through an index fund, as evidenced by backtesting results. However, it's important to note that backtesting results may not always translate to real-life success due to factors like trading costs, market efficiency, and model competition. The theory behind this phenomenon is that smaller stocks may be less efficiently priced compared to larger ones, but the authors of the study acknowledge that they don't fully understand why this occurs. Ultimately, while the findings are intriguing, investors should approach them with caution and consider the potential challenges of implementing such a strategy in practice. Additionally, it's worth noting that numerous researchers and investors are exploring similar ideas, so the market for small-cap passive investing may become increasingly competitive.
Computers in stock market analysis: Computers, with their ability to eliminate human biases, can outperform humans in stock market analysis and identification of investment opportunities
Computers, when programmed correctly, can outperform humans in stock market analysis due to their ability to eliminate human biases and focus solely on data. The discussion revolves around researchers testing chat GPT in financial markets and the possibility of unpublished studies that didn't yield the desired results. The speaker also shares his own experience as a stock picker, acknowledging the presence of biases and narratives that can cloud judgment. He emphasizes that computers, unlike humans, are not susceptible to these biases and can identify opportunities more objectively. While this is not a new concept, it's intriguing to consider the implications of computers surpassing human capabilities in stock picking. The speaker himself identifies as someone who struggles with biases and prefers not to label himself as a stock picker. Instead, he appreciates the potential of computers to make unbiased decisions, even if they don't possess intelligence in the traditional sense.
AI and job automation: AI is expected to automate jobs involving data analysis and labor-intensive tasks, but jobs requiring human interaction and physical work are less susceptible to automation, such as nursing, therapy, and construction work.
The advancement of AI technology is expected to automate certain jobs, particularly those involving data analysis and labor-intensive tasks. However, jobs that require human interaction and physical work are less likely to be replaced. For instance, nurses, therapists, and construction workers are less susceptible to automation. In the financial industry, junior analysts who perform data analysis and listen to earnings calls may be replaced by AI. Fund managers, on the other hand, are keen on utilizing AI technology to enhance their work rather than replace them entirely. Ultimately, the economic impact of this shift remains uncertain, and it is essential to consider which professions involve significant human interaction when advising children on career choices.
AI and big calls in asset management: AI and chat GPT may excel in earnings estimates but their ability to make substantial investments based on big-picture fundamentals is yet to be proven
While AI and chat GPT may excel in performing statistical analysis and making earnings estimates in asset management, they may not be equally effective in making the big calls that lead to significant returns. Traditional investors often focus on the fundamental aspects of a business, such as its structural strength and competitive position, rather than minor earnings forecasts. Jim Simons of Renaissance Technologies, a renowned fund, discovered correlations that allowed his computers to make profitable investments, which didn't make sense to humans. However, if these correlations were obvious, someone would have discovered them already. Chat GPT and similar technologies can handle the grunt work of earnings estimates, but their ability to make substantial investments based on big-picture fundamentals remains to be seen.
Role of Technology in Investing: Skilled investors can potentially enhance their performance using technology, as seen in the speaker's hope to improve using it in the FT competition, and the shift in defined benefit plans due to rising interest rates is an example of technology's impact.
Technology may aid skilled investors in enhancing their performance. The speaker acknowledges the existence of successful investors and proposes the potential for technology to augment their strategies. He also shares his own experience of using monkeys to write his column and expresses his hope to improve in the upcoming FT competition. Employers with defined benefit plans have seen a shift due to rising interest rates, requiring less funding to ensure benefits. In the "long short" segment, the speaker shares his negative experience with slugs and offers an unconventional solution. Overall, the discussion revolves around the role of technology in enhancing investment strategies and the current state of employer-based retirement plans.
Commitments and Extraordinary Experiences: Honoring commitments and respecting agreements is crucial, even if past decisions may seem questionable. Embracing new and extraordinary experiences can also enrich our lives.
Even if past decisions seem questionable, it's important to honor commitments and respect agreements. This was discussed in relation to Elon Musk's controversial $56 billion pay package. Despite criticisms from investors, the company and shareholders had agreed to it at the time, and it's essential to uphold such obligations. Additionally, the importance of experiencing new and extraordinary things was emphasized, as was seen in the speaker's attendance at a Youja Wang concert, where they were left in awe of her remarkable piano performances. The event resulted in ten encores, a testament to her talent and the audience's appreciation. Overall, the conversation touched upon the significance of adhering to commitments and enjoying life's unique experiences.
Yuja Wang: Yuja Wang is a remarkable pianist with an engaging conversation on Unhedged, produced by Jake Harper and edited by Brian Erstadt, for FT Premium subscribers.
Key takeaway from this episode of Unhedged is the introduction of Yuja Wang, a remarkable pianist whose name is spelled Y-U-J-A. If you're not familiar with her, take a moment to look her up. The hosts, Rob and Katie, had an engaging conversation about her and shared some fascinating insights. For those interested, you can request an encore of this chat or tune in on Thursday for more. Unhedged is produced by Jake Harper and edited by Brian Erstadt, with additional help from Topher4Heads. The FT's global head of audio is Cheryl Bromley. Special thanks to Laura Clark, Alistomaki, Greta Cone, and Natalie Sideler. FT Premium subscribers can receive the Unhedge newsletter for free, and a 30-day free trial is available to everyone else. So, that's it for today, folks. I'm Katie Martin, and thank you for listening.