If Mario Draghi, the boss of the European Central Bank, appears more downbeat at his regular news conference than usual, it may foreshadow a cut in the bank’s monetary stimulus to the economy. That, at least, is the conclusion of two Japanese researchers who used artificial intelligence (AI) to analyze the tiniest of changes in Mr. Draghi’s facial expressions during his policy meetings.
In their research paper, Yoshiyuki Suimon and Daichi Isami from the University of Tokyo also took a closer look at Mr. Draghi’s Japanese counterpart, Haruhiko Kuroda, in 2017. They claim they identified a link between patterns in his facial expressions and subsequent policy changes. The authors said subtle alterations in Draghi’s look could reflect a sense of frustration the ECB boss might have felt – note the careful phrasing – before making policy adjustments.
The two boffins analyzed Mr. Draghi from video footage of news conferences, capturing his expression every half-second. They then ran those images through “Emotion API,” a program developed by Microsoft. Emotion API uses a visual recognition algorithm to break down human emotions into eight categories: happiness, sadness, surprise, anger, fear, contempt, disgust and neutral.
For Mr. Draghi, the program revealed signs of “sadness” preceding two recent major policy changes: a modest reduction in its bond-buying program in December 2016, and another cutback in quantitative easing in October of last year.
Europe’s most important central banker, of course, is used to fending off such scrutiny. Mr. Draghi’s “emotion scores,” the Japanese researchers said, were smaller than Bank of Japan Governor Kuroda’s. Somewhat surprisingly, this suggests that the Italian-born Mr. Draghi, a former managing director at Goldman Sachs’ international arm, is more inscrutable than his Japanese counterpart. “Draghi maintains more control over his expressions, whether he is doing so consciously or not,” said Mr. Suimon, the lead author of the study.
They also examined the visage of ECB Vice President Vitor Constancio, who sits next to Mr. Draghi at his news conferences. Mr. Constancio showed more happiness even as Mr. Draghi’s score dropped. Kiyoshi Izumi, professor of the University of Tokyo, who specializes in financial data mining and artificial market simulation, said studying simultaneous facial expressions from a team of policymakers, such as Mr. Draghi and Mr. Constancio, provided a stronger basis for drawing conclusions.
“Some people – President Draghi, in this case – are better at poker-facing than Governor Kuroda. So it’s interesting and worth analyzing the news conference as a whole,” Mr. Izumi said.
Jeremy Gray is an editor at Handelsblatt Global. To contact the author: firstname.lastname@example.org