Technology Innovation and Diffusion as Sources of Output and Asset Price Fluctuations
Executive Summary — A central challenge to modern business cycle analysis is that standard macro models are unable to generate fluctuations in the stock market with the amplitude, persistence, and lead-lag pattern observed in the data. At the same time, standard macro models predict that good news about future, such as those received during 1994-1995 on the arrival of IT, lead to recessions rather than expansions. HBS professor Diego Comin and coauthors develop a model that overcomes these two problems by explicitly incorporating an endogenous speed of diffusion of technologies that is increasing in the resources spent in adoption. Revisions in beliefs about future profits generate fluctuations in the stock market with the amplitude and lead over output observed in the data. The firms' investment decisions in adoption leads to a shift in labor demand that increases hours worked and output. Key concepts include:
- In the model, news about future growth prospects produces movements in current output and hours that are positively correlated with the news.
- The mechanism described here is also potentially relevant to business fluctuations driven by factors other than news about future technological prospects. In particular, endogenous technology adoption amplifies the effects of these shocks vis-à-vis a model without this mechanism.
- The framework also broadly captures the cyclical pattern of stock price movements. It can account for the run-up of stock prices in the mid 1990s and also some of the decline preceding the most recent recession.
We develop a model in which innovations in an economy's growth potential are an important driving force of the business cycle. The framework shares the emphasis of the recent "new shock" literature on revisions of beliefs about the future as a source of fluctuations, but differs by tieing these beliefs to fundamentals of the evolution of the technology frontier. An important feature of the model is that the process of moving to the frontier involves costly technology adoption. In this way, news of improved growth potential has a positive effect on current hours. As we show, the model also has reasonable implications for stock prices. We estimate our model for data post-1984 and show that the innovations shock accounts for nearly a third of the variation in output at business cycle frequencies. The estimated model also accounts reasonably well for the large gyration in stock prices over this period. Finally, the endogenous adoption mechanism plays a significant role in amplifying other shocks. Keywords: Business Cycles, Endogenous Technology Adoption, News Shocks, Stock Market. JEL classifications: E3, O3. 58 pages.