Source: Cheng & Golshan (2025) — “Silent Suffering: Using Machine Learning to Measure CEO Depression”
At first this sounds like a technical finance question. Then you look closer and realise it is also a question about attention, incentives, and whether people are using information or just being impressed by it.
This paper sits at an unusual intersection: clinical psychology, vocal acoustics, machine learning, and corporate governance. The researchers introduce a measure of CEO depression derived from acoustic features of conference call recordings, validated against clinical assessments, and then examine whether the measure predicts anything about firm behaviour. CEO depression, as measured by the model, correlates with subsequent corporate outcomes including investment decisions, risk-taking, and firm performance. The effect is not trivial — mental state, it turns out, leaves measurable traces in both voice and firm-level data. The measurement approach is worth understanding. Clinical depression has known vocal signatures: changes in pitch, pace, monotonicity, and other acoustic properties that clinicians have studied for decades. The researchers adapt machine learning models originally developed for clinical depression detection and apply them to the professional context of earnings calls. The pre-registration of this study through the Journal of Accounting Research’s registered reports process is notable — the hypotheses were committed to before seeing the results, which substantially reduces the risk of data-mining. The broader point is about what AI measurement enables in governance research. CEO health and mental state has long been treated as unobservable — or at least as something that would require invasive data collection to study systematically.
So I would not read this as a neat technology story. It is a messy information story. The tools improve, but people still have to decide what deserves attention. Annoying, but true.