Intrinsic motivation for honesty is perceived as an important determinant of large and persistent variation in cheating behavior. However, little is known about its actual role due to challenges in obtaining precise measures of motivation for honesty, as well as field outcomes on cheating. We fill these gaps using a unique setting of informal milk markets in India. A novel behavioral experiment, which combines a standard die roll task with Bluetooth technology, is used to measure motivation for honesty of milkmen at both extensive and intensive margins. We then buy milk from the same milkmen and show that cheating in the field, measured by the amount of water added to milk, widens significantly with a milkman’s degree of dishonesty. Additional analyses show that conventional binary measure of motivation for honesty suffers from measurement errors, resulting in underestimation of this association.
SAFE Working Paper No. 134