Testing Metcalfe's law: Pitfalls and possibilities

Abstract

A small but burgeoning body of literature has tried to assess whether Metcalfe's law provides a realistic yardstick for the value of specific networks. In this paper, I uncover a number of flaws in the extant tests. First, a proper test of Metcalfe's law—or of any of the competing “laws”—requires correct identification of the type(s) of network effects involved and the relevant market(s). Second, a multi-market setting typically calls for scaled network sizes. Third, controlling for intertemporal changes in network quality may be imperative. Finally, indicators at the individual and aggregate levels should not be mixed. Armed with these insights, I re-examined Madureira et al. (2013)’s results. Unlike Madureira et al., I found that Metcalfe's law fits the data better than Briscoe's law.