Why do some technologies lead to speculation while others do not? Why were there speculative bubbles in stocks of early radio producers and broadcasters, “aeroplane” manufacturers and airlines, internet storefronts, electronics producers, electric automobile manufacturers, and transcontinental railroads, but not in the stocks of producers of lasers, northeastern railroads, antibiotics, nylon, rayon, cellophane, or televisions?
Our proposed work aims to rectify an important methodological flaw in current studies of speculative crises: one cannot identify the causes of bubbles by examining a single case, nor can one identify bubbles based on an analysis of bubbles throughout history. ...view middle of the document...
Our finding is consistent with the hypothesis that “noise traders” can move markets (Delong, Shleifer, Summers & Waldmann, 1990), as well as the argument that novices drove the events of 1929 (Galbraith, 2009 ). Importantly, we exploit variation in business model uncertainty in our sample to explain differences in the likelihood of an asset bubble.
Fourth, we find that speculation is generally inhibited in the case of strong intellectual property protection.
These findings are based on preliminary analysis of 12/58 of the total sample of innovations. If our findings hold for the complete sample, we will recast past financial manias and crises, such as the bull market of the 1920s and the Great Crash and the “Tronics” boom and bust of the 1960s, as well as recent crises such as the internet bubble and crash, as consequences of market democratization.
Three important disclaimers are in order. First, we do not set out to explain banking crises – at least not directly. Instead our focus is on asset bubbles associated with new technologies. Second, our goals are modest with regard to the question of investor rationality, in the sense that we are interested in identifying conditions that increase the propensity for rapid swings in asset prices, but understand that one can, in most cases, identify assumptions that would lead markets with rational actors to generate observed price patterns. Finally, in the spirit of Goldfarb, Kirsch & Miller (2007) our analysis is stochastic. While we identify conditions that are associated with a greater likelihood of bubbles, in no place do we claim that these conditions are sufficient to generate asset bubbles.
Method: We turn to the historical record to identify two distinct timelines.
Identifying Trends in Market Democratization: First, we compile a timeline that identify influxes of inexperienced investors into markets. Democratization of investment is generally a complex evolution of a multiple of institutions and requires careful identification of meaningful influxes of new investors. We know of no systematic account of market democratization over the entirety of our sample period. Instead, we piece together existing scholarly accounts of individual episodes and supplement these accounts with primary research and contemporary accounts as needed.
Identifying Candidate Technologies & Speculative Episodes:
Candidate Technologies: We require a sample of technologies that were not selected based upon their generation of market speculation, but were sufficiently novel and important to have that potential. To this end, we turn to a series of inventories of technologies that were hypothesized to significantly impact economic growth (Kleinknecht, 1984; Mensch, 1979 [from Clark, Freeman & Soete, 1984]; and Jewkes, Sawers & Sillerman, 1969). We aggregate the results of the three lists. If a technology appears on two or more of these lists, we include it in our study. There are 177 technologies on...