Connect with us

Finance

Artificial intelligence and asset pricing: The power of transformers

Published

on

The rapid advancement of artificial intelligence (AI) has reshaped numerous fields, and finance is no exception. Asset pricing, a domain traditionally dominated by linear models and factor-based approaches, is now experiencing a transformative shift due to AI’s capacity to uncover persistent predictive patterns in financial data. The introduction of large-scale AI models – particularly transformer-based architectures – has significantly enhanced our capability to model complex relationships among assets and firms (Eisfeldt et al. 2023), leading to improved forecasting and risk assessment.

In this column, we explore how AI-driven asset pricing models, particularly those incorporating transformers, leverage cross-asset information sharing to reduce pricing errors and enhance predictive accuracy. These innovations offer a novel perspective on how financial markets process information and determine asset prices.

Traditional versus AI-driven asset pricing

Most traditional asset pricing models, such as the Fama and French (1993) framework, rely on predefined factors that explain asset returns. While effective, these models assume a fixed and linear relationship between asset characteristics and expected returns. AI, on the other hand, introduces a non-linear, data-driven approach, identifying patterns that are often invisible to traditional methods.

Machine learning models, including tree-based methods and neural networks (Gu et al. 2020), have improved asset return prediction by capturing complex relationships between firm characteristics and returns. However, these models typically focus on ‘own-asset prediction’, meaning that they use only an asset’s individual characteristics to estimate future returns. This approach ignores the broader context in which assets interact.

The role of transformers in asset pricing

In a recent paper (Kelly et al. 2024), we introduce the Artificial Intelligence Pricing Model (AIPM), which embeds transformer networks into the stochastic discount factor (SDF) framework. While initially developed for natural language processing, we show that transformers are remarkably effective in financial applications due to their ability to capture cross-sectional dependencies across assets.

Advertisement

Unlike traditional machine learning models, transformers incorporate the ‘attention mechanism’, allowing them to dynamically adjust the weight placed on different inputs based on their relevance. In the context of asset pricing, this means that the model not only considers an asset’s own characteristics but also how these characteristics interact with those of other assets. This approach significantly enhances predictive power by leveraging market-wide information.

Empirical findings: Performance of AI-based models

We evaluated the performance of the transformer-based AIPM using a dataset of US stock returns and conditioning variables. Compared to traditional asset pricing models and other machine learning approaches, the AIPM demonstrated:

  • Lower pricing errors: The model achieved significantly smaller out-of-sample pricing errors compared to traditional factor-based models and neural networks without attention mechanisms.
  • Higher Sharpe ratios: By integrating cross-asset dependencies, the transformer-based model outperformed existing approaches in terms of risk-adjusted returns out-of-sample.
  • Scalability and complexity gains: We found that increasing model complexity – by incorporating deeper transformer layers – consistently improved predictive performance. This supports the notion that AI models benefit from higher parameterisation when applied to asset pricing.

Empirical evidence from the literature on large language models (e.g. Kaplan et al. 2020) indicates that adding more transformer blocks enhances the capacity of models to effectively represent language. Models with deeper architectures are able to capture more abstract features and longer-range dependencies than shallower models. Each additional layer refines the attention distributions, allowing the model to consider both short-term and long-term relationships. Interestingly, the same benefits of transformer complexity emerge in the context of the asset pricing model.

These findings suggest that AI-driven asset pricing models are more efficient in processing vast amounts of financial data, leading to more accurate and robust predictions.

Implications for investors and policymakers

The integration of AI into asset pricing has profound implications for market participants and policymakers. Investors can benefit from improved portfolio allocation strategies driven by AI’s capability to identify subtle pricing inefficiencies, a concept aligned with research on AI and personality traits shaping economic returns (Makridis 2025). Meanwhile, regulators and policymakers must consider the impact of AI-driven trading on market stability and efficiency.

Moreover, the adoption of AI in asset pricing challenges traditional views on market efficiency, echoing broader concerns about AI’s macroeconomic impact and productivity gains (Filippucci et al. 2024). If AI models consistently outperform classical frameworks, it may suggest that markets are less efficient than previously thought, opening the door for further research into the nature of pricing anomalies.

Advertisement

Conclusion

The fusion of artificial intelligence and finance is revolutionising asset pricing. Our research demonstrates that transformer-based models significantly enhance return predictions by leveraging cross-asset information sharing. As AI continues to evolve, its role in financial decision-making will only grow, offering new opportunities for investors and reshaping our understanding of market dynamics.

References

Gu, S, B Kelly and D Xiu (2020), “Empirical Asset Pricing via Machine Learning”, Review of Financial Studies 33 (5): 2223-2273.

Eisfeldt, A, G Schubert and M B Zhang (2023), “Generative AI and Firm Valuation”, VoxEU.org, 4 June.

Fama, E and K French (1993), “Common risk factors in the returns on stocks and bonds”, Journal of Financial Economics 33(1): 3-56.

Filippucci, F, P Gal, and M Schief (2024), “Miracle or Myth? Assessing the Macroeconomic Productivity Gains from Artificial Intelligence”, VoxEU.org, 8 December.

Advertisement

Kaplan, J, S McCandlish, T Henighan et al. (2020), “Scaling laws for neural language models”, arXiv preprint arXiv:2001.08361.

Kelly, B Kuznetsov, S Malamud and T Xu (2024), “Artificial Intelligence Asset Pricing Models”, NBER Working Paper No. w33351

Makridis, C (2025), “The Role of Personality Traits in Shaping Economic Returns Amid Technological Change”, VoxEU.org, 31 January.

Vaswani, A, N Shazeer, N Parmar, J Uszkoreit, L Jones, A N Gomez, L Kaiser and I Polosukhin (2017), “Attention is all you need”, Advances in Neural Information Processing Systems 30.

Advertisement
Continue Reading
Advertisement
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Finance

Personal Finance: Artificial intelligence is taking cyber scams to a whole new level | Chattanooga Times Free Press

Published

on

Personal Finance: Artificial intelligence is taking cyber scams to a whole new level | Chattanooga Times Free Press

Americans fell victim to $12.5 billion in fraud losses last year, according to the Federal Trade Commission. That represents a startling 25% increase over a year ago. The FBI estimates the losses are even larger, over $16 billion. So, what explains the sharp increase, considering that most consumers are far more attuned to cybercrimes? Like so many other questions, the answer is artificial intelligence.

Forget the Nigerian Prince scam (although that tired, old routine still separated Americans from nearly $1 million last year). And gone are the days when phishing emails screamed “bogus” thanks to typos and bad translations. Artificial intelligence has entered the arena and is assisting criminals in producing ever more believable and compelling appeals. It is getting nearly impossible to spot a fake, so it becomes even more essential to question everything that comes to you unsolicited.

Here are a few examples of state-of-the-art tactics, thanks to generative artificial intelligence.

Enhanced phishing attacks. Phishing attacks involving unsolicited emails or text messages attempt to convince the recipient to provide personal information that can then be used to hack into bank accounts or steal identities. The crooks can now run a draft of their handiwork through applications like ChatGPT to clean up grammar and spelling but also to scour your social media to personalize the message and make it more conversational and therefore more credible.

Deepfakes. This is a general term describing ultra realistic reproductions of documents, voices or even video messages. A common tactic is producing identification documents like driver’s licenses, birth certificates or title papers that can be used to steal your identity. These phony papers often include realistic elements like watermarks or other AI-generated images that convey legitimacy.

Advertisement

It is now simple for a criminal to clone the voice of a familiar person or even a family member. Victims may be persuaded to send money or grant account access, especially if they believe their friend or loved one is under duress and needs help.

Well-made deepfake videos are now becoming nearly impossible to recognize and are proliferating wildly. They may mimic celebrity endorsers or even replicate a family member to spread misinformation or direct the victim to a fake website. Romance scams are particularly insidious, especially among the senior population, and the scale of the technology allows the attacker to carry on multiple “romances” simultaneously.

Endless variety. Schemes pop up faster than law enforcement can track them. One recent caper involves stealing someone’s identity, enrolling in an online college course using their name and pocketing some of the student loan funds. In some cases, AI chatbots even submitted homework and took exams to maintain the ruse, and some legitimate students have been crowded out of classes because the chatbots filled the seats. And the cyber crime arms race is just heating up.

What to do if you believe you have been victimized. If you suspect that you have been targeted by an internet scammer, it is essential that you report the incident. Security experts believe that most victims fail to report the crime, often out of fear or embarrassment.

Begin by filing an online report with the Federal Trade Commission at ReportFraud.gov. The commission will log your case and provide you with a list of next steps to take to pursue a recovery and to reduce your chances of being scammed again.

Advertisement

If the scam involves your bank account or credit cards, contact the financial institution to notify them of the loss. You may need to close your old accounts and open new ones. Also remember that you are not responsible for fraud losses on credit cards if you report the event promptly.

Ironically, but hardly surprisingly, scammers are impersonating the Federal Trade Commission itself. Note that the FTC will never threaten you or suggest that you transfer or withdraw funds.

You should also report the details to the Internet Crime Complaint Center, known as IC3. This is a central repository run by the FBI that compiles data that is used by law enforcement agencies to investigate cybercrimes, and your input is valuable.

If the attack involved identity theft or if you believe the attacker obtained some of your personal information, visit IdentityTheft.gov (another Federal Trade Commission resource) to report your case and obtain information on how to reclaim control of your information.

Take steps now to reduce your risk. The internet, email and text messaging are places where you should trust no one. Never respond to unsolicited offers, requests or threats. If you are concerned about ignoring potentially valid communications, look up the contact information separately and reach out directly to the company or agency to confirm the communication.

Advertisement

Always use multi-factor identification, like a validation text (preferred) or email to complete a sign in process. Never give your passwords to anyone and be sure to use a unique password for every website you sign into. Many if not most people fail this one. There are also very user friendly applications called password keepers that will track your disparate login information for you.

Finally, it is well worth the effort to initiate a credit freeze with the three major credit reporting bureaus, Experian, Transunion and Equifax. This will block any attempts to access your file and can easily be lifted if you need to apply for credit.

Cyber criminals are constantly innovating, and the old days of clumsy, easily spotted phishing scams are long over. Artificial intelligence has made scams harder to detect and call for even greater vigilance.

Christopher A. Hopkins, CFA, is a co-founder of Apogee Wealth Partners in Chattanooga.

Advertisement
Continue Reading

Finance

Lument Finance Trust, Inc. Declares Quarterly Cash Dividends for its Common and Preferred Stock

Published

on

Lument Finance Trust, Inc. Declares Quarterly Cash Dividends for its Common and Preferred Stock

NEW YORK, June 20, 2025 /PRNewswire/ — Lument Finance Trust, Inc. (NYSE: LFT) (“LFT” or the “Company”) announced the declaration of a cash dividend of $0.06 per share of common stock with respect to the second quarter of 2025. The dividend is payable on July 15, 2025, to common stockholders of record as of the close of business on June 30, 2025.

The Company also announced the declaration of a cash dividend of $0.4921875 per share of 7.875% Cumulative Redeemable Series A Preferred Stock. The dividend is payable on July 15, 2025 to preferred stockholders of record as of the close of business July 1, 2025.

James P. Flynn, Chief Executive Officer, said, “We approached this quarter’s dividend decision with thoughtful deliberation and a clear-eyed view of our near-term earnings outlook. We determined it was prudent to adjust the dividend to reflect present realities, preserve book value and support our long-term earnings potential. Our focus remains on maximizing our flexibility, in order to achieve positive asset management outcomes and responsibly manage our liquidity. We believe these efforts will best position us to create long-term value for our shareholders.”

About LFT

LFT is a Maryland corporation focused on investing in, financing and managing a portfolio of commercial real estate debt investments. The Company primarily invests in transitional floating rate commercial mortgage loans with an emphasis on middle-market multi-family assets. LFT is externally managed and advised by Lument Investment Management, LLC, a Delaware limited liability company.

Advertisement

Additional Information and Where to Find It

Investors, security holders and other interested persons may find additional information regarding the Company at the SEC’s Internet site at http://www.sec.gov/ or the Company website www.lumentfinancetrust.com or by directing requests to: Lument Finance Trust, 230 Park Avenue, 20th Floor, New York, NY 10169, Attention: Investor Relations.

Forward Looking Statements

 Certain statements included in this press release constitute forward-looking statements intended to qualify for the safe harbor contained in Section 27A of the Securities Act of 1933, as amended, and Section 21E of the Securities Exchange Act, as amended. Forward-looking statements are subject to risks and uncertainties. You can identify forward-looking statements by use of words such as “believe,” “expect,” “anticipate,” “project,” “estimate,” “plan,” “continue,” “intend,” “should,” “may,” “will,” “seek,” “would,” “could,” or similar expressions or other comparable terms, or by discussions of strategy, plans or intentions. Forward-looking statements are based on the Company’s beliefs, assumptions and expectations of its future performance, taking into account all information currently available to the Company on the date of this press release or the date on which such statements are first made. Actual results may differ from expectations, estimates and projections. You are cautioned not to place undue reliance on forward-looking statements in this press release and should consider carefully the factors described in Part I, Item IA “Risk Factors” in the Company’s Annual Report on Form 10-K for the year ended December 31, 2024, which is available on the SEC’s website at www.sec.gov, and in other current or periodic filings with the SEC, when evaluating these forward-looking statements. Forward-looking statements are subject to substantial risks and uncertainties, many of which are difficult to predict and are generally beyond the Company’s control. Except as required by applicable law, the Company disclaims any intention or obligation to update or revise any forward-looking statements, whether as a result of new information, future events or otherwise.

SOURCE Lument Finance Trust, Inc.

Advertisement
Continue Reading

Finance

Investors in Apollo Commercial Real Estate Finance (NYSE:ARI) have seen returns of 30% over the past three years

Published

on

Investors in Apollo Commercial Real Estate Finance (NYSE:ARI) have seen returns of 30% over the past three years

As an investor its worth striving to ensure your overall portfolio beats the market average. But if you try your hand at stock picking, you risk returning less than the market. We regret to report that long term Apollo Commercial Real Estate Finance, Inc. (NYSE:ARI) shareholders have had that experience, with the share price dropping 12% in three years, versus a market return of about 57%.

With that in mind, it’s worth seeing if the company’s underlying fundamentals have been the driver of long term performance, or if there are some discrepancies.

AI is about to change healthcare. These 20 stocks are working on everything from early diagnostics to drug discovery. The best part – they are all under $10bn in marketcap – there is still time to get in early.

While markets are a powerful pricing mechanism, share prices reflect investor sentiment, not just underlying business performance. One way to examine how market sentiment has changed over time is to look at the interaction between a company’s share price and its earnings per share (EPS).

Apollo Commercial Real Estate Finance has made a profit in the past. On the other hand, it reported a trailing twelve months loss, suggesting it isn’t reliably profitable. Other metrics may better explain the share price move.

Advertisement

It’s quite likely that the declining dividend has caused some investors to sell their shares, pushing the price lower in the process. The revenue decline, at an annual rate of 19% over three years, might be considered salt in the wound.

You can see how earnings and revenue have changed over time in the image below (click on the chart to see the exact values).

NYSE:ARI Earnings and Revenue Growth June 20th 2025

It’s probably worth noting that the CEO is paid less than the median at similar sized companies. But while CEO remuneration is always worth checking, the really important question is whether the company can grow earnings going forward. So we recommend checking out this free report showing consensus forecasts

As well as measuring the share price return, investors should also consider the total shareholder return (TSR). The TSR is a return calculation that accounts for the value of cash dividends (assuming that any dividend received was reinvested) and the calculated value of any discounted capital raisings and spin-offs. Arguably, the TSR gives a more comprehensive picture of the return generated by a stock. As it happens, Apollo Commercial Real Estate Finance’s TSR for the last 3 years was 30%, which exceeds the share price return mentioned earlier. And there’s no prize for guessing that the dividend payments largely explain the divergence!

Advertisement

Continue Reading

Trending