“…A natural starting point for our inquiry is the existence of a strong empirical regularity that appears to connect attitudes towards compound risk and ambiguity. Starting with Halevy (2007), a growing experimental literature documents a large correlation between ambiguity aversion and compound risk aversion 2 (Abdellaoui, Klibanoff, and Placido, 2015;Armantier and Treich, 2016;Prokosheva, 2016;Chew, Miao, and Zhong, 2017;Aydogan, Berger, and Bosetti, 2019;Dean and Ortoleva, 2019;Gillen, Snowberg, and Yariv, 2019;Schneider and Schonger, 2019;and Berger and Bosetti, 2020). For example, controlling for measurement error and the associated attenuation bias, Gillen, Snowberg, and Yariv (2019) find that the correlation between compound risk aversion and ambiguity aversion is 0.86.…”