Finance
Artificial intelligence and asset pricing: The power of transformers
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.
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.
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.
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.
Finance
How much will Social Security go up next year? See latest forecast
How to find your Social Security Number safely
Learn how to safely find your Social Security Number with the official Social Security website.
Problem Solved
Before Social Security payments are posted this week, many retirees are looking ahead at the potential Cost of Living Adjustment for 2027 with an advocacy group predicting a similar increase to 2026.
On April 10, The Senior Citizens League — a nongovernmental advocacy group for seniors — released its monthly COLA forecast for 2027, saying data showed a 2.8% increase is likely.
“Over the last seven weeks, crude oil prices have soared, and fuel prices have followed suit. Consumers are getting pinched at the pump as gas prices soar, while businesses are paying more for transportation and/or production costs. This energy price shock is beginning to show up in the monthly U.S. inflation report, and it’s having a tangible impact on 2027 COLA forecasts,” The Motley Fool, a financial and investing advice company, and USA TODAY content partner, reported on April 18.
The official announcement will come in October, as it’s based on third-quarter inflation data.
According to Consumer Price Index data published last week, the annual inflation rate reached a two-year high of 3.3%, up 0.9% over the last month. This is largely due to soaring oil prices caused by the war in Iran.
Social Security payments are always scheduled on Wednesdays, with the final wave of this month scheduled for April 22, according to the Social Security Administration. The schedule is based on the birth dates of the recipients — retired, disabled workers or survivors.
Here’s who will get a Social Security check this week and more on the 2027 COLA forecast:
When is the final Social Security in April 2026?
Social Security benefits are sent out based on the recipients’ birth dates. Wednesday, April 22, is the final wave of payments for those with birth dates between the 21st and the 31st of April.
What is the 2027 COLA forecast?
The 2027 COLA increase is forecast to be 2.8% due to continuing inflation prices, according to The Senior Citizens League’s April 10 press release. If the SSA approves that rate of increase, average payment for retired workers would go up by $56 per month in January 2027.
The SCL releases a COLA prediction each month based on the Consumer Price Index, Federal Reserve interest rate and the National Unemployment rate from the U.S. Bureau of Labor Statistics.
Beneficiaries who want to stay updated with the monthly predictions may visit the SCL’s “COLA Watch” webpage that includes the forecast, calculations, historical trends and more.
The official COLA increase for 2027 will be announced in October 2026.
What were the big Social Security changes in 2026?
At the beginning of 2026 recipients received a 2.8% COLA for Social Security and Supplemental Security Income (SSI) payments, according to the SSA’s COLA Fact Sheet and American Association of Retired Persons, increasing payments about $56 per month.
Here are more details on the 2026 COLA increase, per the SSA:
- The maximum amount of earnings subject to the Social Security tax increased to $184,500.
- The earnings limit for workers who are younger than full retirement age (67 years old) increased to $24,480. (There will be a $1 deduction for each $2 earned over $24,480.)
- The earnings limit for people reaching their full retirement age in 2026 increased to $65,160. (There will be a $1 deduction for each $3 earned over $65,160, until the month the worker turns full retirement age.)
- There is no limit on earnings for workers who are at full retirement age or older for the entire year.
What should I do if I don’t get my Social Security payment?
According to the SSA, if you don’t receive your payment on the scheduled date, wait three days additional days, then call their office.
Where are the Social Security offices in Michigan?
There are 48 offices in Michigan, and to find an office near you, recipients may use the office locator via the Social Security’s website by entering your zip code for office hours, numbers, available services and more.
How can I replace my Social Security card?
The personal account, “my Social Security” allows recipients to manage their personal records, including a request for a replacement Social Security card and benefit statements for taxes and more. New accounts are created using ID.me or Login.gov as a multifactor authentication.
When will I get my checks in May? Full 2026 schedule
USA TODAY Contributed
Contact Sarah Moore @ smoore@lsj.com
Finance
Hong Kong reasserts role as safe haven in global finance amid Iran conflict
The seven-week military conflict in the Middle East will redefine Hong Kong’s role as a global financial centre, positioning the city as a safe harbour for capital and investments.
Anecdotal evidence suggested that more banks had turned to Hong Kong to protect their businesses and committed themselves to expanding their presence in the city. At the same time, inquiries about adding allocations of mainland Chinese assets among global investors had recently increased, potentially enlarging the customer base for the city’s asset-management industry and family offices and driving demand for offshore yuan-linked financial products.
For years, Hong Kong’s status as a financial centre in the Asia-Pacific region has been challenged by Dubai, which has risen to prominence as a gateway linking Asia and Europe in capital flows, transport and logistics. With the war destabilising the Middle East – at one point forcing the closure of the Dubai International Airport and sending stocks in the Gulf region plunging – Hong Kong has re-emerged due to its geographical location, a pegged exchange rate, free capital flows and support from China’s economic strength.
“In that context, China and Hong Kong are attracting renewed attention,” said Gary Dugan, CEO of The Global CIO Office in Dubai, which advises family offices and ultra-high-net-worth individuals globally. “There is growing interest among some clients in increasing exposure to China and Hong Kong. It is less a simple flight to safety and more a reassessment of where investors see relative value, policy consistency and long-term strategic opportunity.”
Dubai now relies on trade, tourism and finance as the pillars of its economy, reflecting the success of its four-decade diversification away from oil for sustained growth. The United Arab Emirates city is home to Jebel Ali Free Zone, the biggest free-trade zone in the Middle East, and the second-largest stock market in the region, with combined market values of US$1.01 trillion. The city, also a global hub for gold trading, has a population of 4 million, about 80 per cent of which are foreign expatriates. Dubai’s economy grew by 4.7 per cent in the January-to-September period last year.
Finance
Budget crisis is top concern for MPS leader Cassellius | Opinion
Before seeking a new referendum MPS needs to rebuild trust in the community through completing state audits, putting in place controls to prevent overspending and routine reports to the public.
For MPS Superintendent Brenda Cassellius, who just wrapped up her first year leading Milwaukee’s public school system, her tenure has been punctuated by some very big numbers.
The first is $252 million. That is the amount of new spending voters narrowly approved in an April 2024 referendum to support operations in Wisconsin’s largest school district. Just months later, MPS was rocked by revelations the district was months behind in filing key financial reports to the state, which led to former Superintendent Keith Posley’s resignation.
The second is $1 billion. MPS faces a deferred maintenance backlog exceeding $1 billion. The district’s enrollment has declined 30% over the last 30 years, leaving many schools at less than 50% full. That, in part, is driving a plan to close some schools and to improve others to help lower costs.
The final is $46 million, the deficit MPS was running for the 2024-25 school year, an unexpected shortfall which has led to hundreds of staff layoffs.
Getting the district’s accounting, budgeting and financial reporting back on track has dominated Cassellius’s first year at MPS. In an April 15 interview with the Journal Sentinel’s editorial board, she talked in detail about the challenges putting that into order and progress she sees in restoring transparency into its operations.
State funding and aging buildings create budget nightmares
Cassellius says state needs to keep up its share of school funding
In an interview with the Journal Sentinel editorial board, MPS leader Brenda Cassellius says budgets and buildings are her two top worries.
Cassellius said the on-going budget crisis is her top concern. She said the state’s failure to live up to its share of funding is exacerbating MPS’ budget woes. A group of school districts, teachers and parents filed suit against the state Legislature and its Joint Finance Committee claiming the current state funding system is unconstitutional and prevents schools from meeting students’ educational needs.
Funding for special education is especially critical. About 20% of MPS students have disabilities, almost twice the share of the city’s charter schools, and the average of 14% across Wisconsin.
“What’s keeping me up now, you know, is really just the budget crisis we’re in, with not only this year but multiple years going out without additional state aid, we’ve been not getting funding for what our needs are for our students, and particularly our students with special needs,” she said.
Although the state budget increased special education funding to a 42% reimbursement rate, the actual rate has been about 35%. Another component to the budget headache is the age of MPS buildings. The average age is 85 years-old compared to 45 across the nation.
“We have just kicked this can down the curb or kicked it down the street or whatever you call it for too long. And it’s time that we really take on a serious conversation about the conditions of the learning environments in which we send our children,” she said. “Particularly in Milwaukee Public Schools, we serve the most vulnerable children. Children who have language barriers, children who have disabilities, children in high-concentrated poverty.”
What needs to happen before MPS seeks another referendum
Voters need to be comfortable MPS has made tough budget decisions
In an interview with Journal Sentinel editorial board, Brenda Cassellius said voters will need to see budget improvements before seeking more spending
Cassellius said MPS will definitely need to go back to voters for a new referendum in the future. In addition to the 2024 measure, voters approved an $87 million plan in 2020.
Before doing that, she said the district first needs to rebuild trust in the community through completing required state audits, putting into place controls to prevent overspending and routine reports to the school board and public about finances.
“I don’t think that the voters are going to want us to bring something forward until they feel comfortable that we have done the cleanup that is necessary,” she said. “And we’ve built the trust that we have the sufficient controls in place.”
In the interim, she’s hoping the state will meet its constitutional responsibility to adequately fund public schools.
“What the public expects is you know where the money is, you’re spending it as close as you can to children, you’re getting good on the promise around art, music, and PE, and the things the public said they wanted to fund,” Cassellius said. “And they want their kids to have so that they have a quality education and an excellent education in Milwaukee Public Schools, and that they had the right amount of staff that they actually need. In the school to be safe and to run a good operation.”
Rebuilding finance staff in wake of $46 million in overspending
MPS is rebuilding school finance staff in wake of reporting lapses
In an interview with the Journal Sentinel editorial board April 15, MPS superintendent discusses accountability for district’s financial problems.
The $46 million budget shortfall from the 2024-25 school year started coming into view last fall and was confirmed in mid-January. Cassellius noted that in addition to hiring a new superintendent, MPS also parted ways with its comptroller and CFO.
“We are really rebuilding the personnel and staff of the finance department. That is what’s critical, is having the right people in the right seats doing the work,” she said. “Also critical is making sure that you have the right controls in place. The audit findings found that we did not have proper controls in place and now we have those proper controls in place and when we find things we put new SOPs in place and that is what any business does.”
Identifying that shortfall, though painful, was the result of better accounting.
“Being three years behind in auditing means that you don’t have full sight on your actual revenues and expenditures. And so we have now full sight of our revenues and our expenditures and that’s why we were able to see this new deficit of $46 million,” she said. “And we still continue to work with DPI on those processes to make sure that every month we’re doing monthly to actuals and doing those accounting, reporting that to the board. In a way that is consumable to the public that they can understand.”
Jim Fitzhenry is the Ideas Lab Editor/Director of Community Engagement for the Milwaukee Journal Sentinel. Reach him at jfitzhen@gannett.com or 920-993-7154.
-
Politics6 minutes agoGovernor’s race wildly unpredictable two weeks before Californians receive ballots
-
Sports17 minutes agoRod Martin, Raiders Super Bowl hero and USC standout, dies at 72
-
World30 minutes ago‘Predators’: Amnesty slams Netanyahu Putin, Trump, as human rights decline
-
News59 minutes agoThe Onion has agreed to a new deal to take over Infowars
-
New York2 hours agoGotti Grandson Is Sentenced to 15 Months for Covid Relief Fraud
-
Detroit, MI3 hours agoWhy a Detroit family’s $300 brick repair job turned into a fraud investigation
-
San Francisco, CA3 hours agoSea lion pup found in San Francisco’s Outer Sunset malnourished but ‘feisty’
-
Dallas, TX3 hours agoDallas Cowboys 2026 NFL Draft debate heats up