Patently Risky: Framing, Innovation and Entrepreneurial Preferences
24 Pages Posted: 29 Jun 2017 Last revised: 23 Jul 2017
Date Written: June 26, 2017
Innovation policy balances static monopoly rights against dynamic entrepreneurial incentives. In striking this balance, it is commonly presumed that decision makers in innovative settings react to their economic environments in a manner similar to their counterparts in other contexts. This paper reports on a series of experiments that call this presumption into question. Subjects were offered a choice between a sure thing and a risky choice, where our principal manipulation was to alter the decisional frame. Subjects in the control group confronted an unadorned choice between safe and risky options; subjects in the Invest Frame, in contrast, were told that the risky choice was tantamount to an investment in an innovation-related project. In all other respects, we controlled for both the language of the experimental instrument as well as the economic stakes entailed. In some subgroups, the risky choice also included a potential monetary loss. We administered the experiments across three subgroups of settings/subjects: a brick-and-mortar lab using university students as subjects, an Internet protocol again using university students as subjects, and an Internet protocol using Mechanical Turkers as subjects. Our main result, which appears quite strong and robust, is that the Invest Frame induced subjects to manifest greater degrees of risk tolerance on average, across all three settings, and across specifications involving positive and negative payoffs. We calibrate our results to an estimate of a downward “shock” that the invest frame introduces to subjects’ coefficient of relative risk aversion, benchmarking our results against Holt and Laury’s (2002) well documented risk-aversion elicitation scale. A secondary result concerns subject pool comparisons: Mechanical Turk (M-Turk) subjects consistently manifested greater risk aversion than student subjects on the Internet and in the lab (regardless of frame or stakes). Controlling for such baseline differences, however, M-Turkers responded similarly to our manipulation, suggesting that M-Turkers can be used reliably to elicit and test the types of framing questions explored here.
Keywords: Patent, Intellectual Property, Experiment, Risk Preferences
JEL Classification: C9, C92, L2, L26, K1, O34
Suggested Citation: Suggested Citation