Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Products often bundle together many functions, e.g., smartphones. The firm develops the big idea (which functions to bundle) and then chooses one supplier per function. We assume there is a holdup problem and prove that the firm’s bargaining power (Shapley value) is declining in the number of suppliers. Greater scope as measured by the number of suppliers exacerbates holdup, but this can be partially offset by the appropriate choice of vertical integration or outsourcing. Our main result flows from the empirical observation that the number of functions varies across products within an industry (firm heterogeneity). We introduce the notion of an ‘ideas-oriented’ industry as one in which more productive firms have higher marginal returns to introducing a new function. This leads to two testable hypotheses. More productive firms will (1) have more suppliers and (2) be more likely to integrate those suppliers. We take this to the data by training a multilayer perceptron to predict whether or not each of 29 million PATSTAT patent applications involves new/improved functions. We merge these patents with S&P Capital IQ data on 55,000 companies and their supplier networks. We show that in industries where patents are skewed towards new or improved functions, more productive firms have more suppliers and are more likely to integrate these suppliers.