Social Push and the Direction of Innovation

Research by Elias Einio¹, Josh Feng², and Xavier Jaravel³

¹VATT Institute for Economic Research, ²Unviersity of Utah, ³London School of Economics

Several historical examples suggest that innovators' personal experience may shape their entrepreneurial vision, and in turn inequality across the socio-demographic groups who benefit from these innovations. Despite the lack of opportunities for most American women in the 19th century, Josephine Cochrane managed to receive a U.S. patent for her “Dish Washing Machine” in 1886; she wanted to protect her fine china and avoid having to hand-wash them herself. Madam C.J. Walker, an African-American entrepreneur and the first female self-made millionaire in America, made her fortune in the late nineteenth century by developing and marketing a line of cosmetics and hair care products for Black women; she suffered severe dandruff and other scalp ailments herself from a young age. Louis Braille invented a world-famous reading and writing system for visually impaired people; he was himself blinded at the age of three as a result of an accident in his father's workshop.

In our work, we build several datasets linking consumer characteristics to innovators' gender, parental income, and age to present new facts about the direction of innovation and innovators' socio-demographic backgrounds. Consumer characteristics are measured in comprehensive consumption surveys, in detailed scanner data for consumer packaged goods, and in a new data set covering mobile phone applications. Innovators and their backgrounds are identified from patent records, start-up and venture capital databases, registries of firms, and administrative tax records. Using this comprehensive dataset, we document that innovator-consumer homophily is a very common feature of modern innovation systems, holding in both the United States and Finland for all types of innovations we study (new phone applications, new consumer goods, patents, and new firms) and across several demographic measures (gender, age, SES, and geography).

We then build an equilibrium growth model that accounts for innovator-consumer homophily and the differential rates of participation in the innovation system across the same demographic measures. We then run counterfactuals to assess the welfare impacts of increased participation in innovation.

(a) Predictions of the model for growth rate from increased participation in innovation, (b) predictions of the model for welfare differences from increased participation in innovation, and (c) empirical evidence on homophily from phone applications

We used CHPC resources to compute the equilibrium growth model and find the parameters that best matched moments from the data. CHPC allowed for extensive parallelization, which helped greatly speed up each model calibration.

Attribution: This content was provided by the researchers and edited for style by staff at the CHPC.