Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
The proposed blended approach combines identification via heteroskedasticity with sign/narrative restrictions, and instrumental variables. Since heteroskedasticity can point identify shocks, its use results in a sharp reduction of the potentially large identified sets stemming from other approaches. Conversely, sign/narrative restrictions or instrumental variables offer natural solutions to the labeling problem and can help when conditions for point identification through heteroskedasticity are not met. Blending these methods together resolves their respective key issues and leverages their advantages. We illustrate the benefits of the approach in Monte Carlo experiments, and apply it to several examples taken from the literature.