With economists having to react to news cycles getting even shorter and market events happening quicker than before, they can no longer count solely on the traditional widely disseminated macro-metrics and tools to provide actionable analysis, according to Michael McDonough, Chief Economist for Financial Products at Bloomberg.
On October 30, Bloomberg in partnership with Coalition Greenwich released a new study that assesses how U.S. economists and strategists are utilizing data, analytical tools and emerging technology in anticipation of the most impactful macro events, including the U.S. Presidential Election.
The study reveals that economists from leading asset managers, top banks and broker research firms, NGOs and government agencies are adopting new predictive tools, alternative data and generative AI.
McDonough said that for example, in the first presidential debate, people were caught off guard by what happened.
“Relying on traditional polling data means waiting days or even weeks to assess the impact,” he told Traders Magazine.
“Predictive markets, however, offer an immediate sense of how significant an event might be, allowing for a faster understanding of its implications,” he said.
“Broadly speaking, traditional data is very backward looking, and people are trying to get a read on where we are now, not where we were a month ago,” he added.
McDonough said that another key finding, which supports this, was that economists rank alternative data—particularly social media sentiment analysis and real-time consumer transaction data—and Generative AI-enhanced analytics as the #2 and #3 most important tools to analyze macro drivers over the next 12 months.
“We’ve offered prediction market data on the Terminal since 2020, so while we were aware of the demand, we underestimated just how significant that interest would become,” he commented.
According to the findings, early career economists highly value generative AI for use in forecasting models as the single area of their workflow that may be transformed in the coming 2-3 years.
Kevin McPartland, Head of Market Structure and Technology Research at Coalition Greenwich, commented that experienced economists and strategists are also looking to complement traditional tools with innovative solutions, as shown through their adoption of alternative datasets in their daily workflows.
“In our research, 53% of economists with 15-years or more of experience reported utilizing social media sentiment data daily, 47% utilized real-time consumer transaction-level data, 47% utilized web traffic and app usage data, 35% event driven feeds, 29% geolocation data (including foot traffic data) and 24% satellite/imagery data,” he shared.
“Considering this, when analyzing the U.S. economy, 56% of economists with 15-years or more of experience still view point-in-time data as more important than alternative data,” he said.
McPartland told Traders Magazine that forecasting is only one of AI’s early applications, and economists are already adopting AI in complementary ways beyond prediction as they try to determine its value.
For example, he said, the study found that AI is widely used in summarization tools, with 33% of economists and strategists identifying it as the most transformative AI application in their workflow, and 18% viewing sentiment analysis as most transformative.
“Specifically when trying to figure out risk analytics and the sensitivity of a position to the election for a portfolio, macro investors can utilize these AI tools and alternative metrics to help gauge the potential risks, he said.
The findings also suggest that fast, robust, and easily accessible data is key.
“Economists today are faced with an overwhelming array of data, and the ability to quickly access and synthesize this information is essential,” McDonough said.
He stressed that with the sheer volume of data out there today, economists and strategists need to seamlessly integrate it into their workflow and connect all of it to find the signal in all this noise.
“That’s where AI plays a critical role to seamlessly connect the data, model it, present it in a way where they can then extract insights and signals,” he said.
“This is going to become very crucial for economists and strategists in the future, especially as they continue to have to analyze larger raw and calculated data sets than ever before,” he concluded.
