The Power of Print

Beginning with the seminal work by Kydland and Prescott, standard business cycle theory has credited technology shocks with being a major source of business cycle fluctuations. Unfortunately, these shocks, by their nature, are hard to identify and measure. This paper utilizes new measures of technological progress based on publications in the field of technology to address two questions. First, what impact did information and communications general purpose technologies have on the U.S. economy in the postwar period? Second, if there is a significant impact, what sectors were most affected by these types of technology? To examine how information and communications technology (ICT) shocks affect the economy, I first create new measures of ICT technological change based on new information from the Library of Congress’ database, and examine their properties. The rationale for using these new book indicators is that, like patents, the introduction of new titles (excluding new editions) in the field of technology should be related to technological progress. Moreover, as Alexopoulos (2006) discusses, a potential advantage of using book indicators is that new books on technology (e.g., manuals) are likely to be written when the idea or product is first being utilized because publishers want to maximize their profits and recoup their costs associated with the publication. As a result, the timing between changes in technology captured by my indicators and economic activity should be much smaller than the corresponding lag when technological change is measured using research and development expenditures or patent indicators. ’ Next, I use these measures in vector autoregressions to explore how the economy responds to a technology shock. My approach for exploring the impact of technology shocks is closely related to the one used by Shea (1998). However, instead of using data on patents (or R&D), I create a new measure based on previously unstudied information on book titles in the fields of technology used in the U.S. economy. My findings indicate that changes in telecommunications and computer technologies did significantly affect U.S. productivity and GDP growth in both the short and long run. In particular, I find that, in response to a positive ICT technology shock, employment, total factor productivity and capital all increase at the aggregate level. However, not all sectors of the economy respond in the same way. Shocks to computer technology significantly affect GDP in the largest number of sectors, including manufacturing, retail and wholesale trade, telecommunications, transportation, and services. The increase following a positive shock is mostly due to the positive affect of the shock on Total Factor Productivity in the sector. Employment increases and capital increases appear to play a smaller role. Telecommunications shocks affect GDP in a smaller number of sectors (the Retail trade, Service, Transportation, Manufacturing and Construction Industries). Again, I find that the majority of the increase following a positive shock is attributable to the shock’s affect on the industry’s total factor productivity.