Innovation, Diffusion, and Regulation in Energy Technologies

The innovation and diffusion of new technologies is one of the central concerns of economics. New inventions or technological combinations do not spring fully formed into the world; as firms encounter and learn about new technologies they experiment, refine, and learn about them, improving productivity (and sometimes earning economic rents). Understanding the processes by which firms learn, and how these processes interact with regulations, is fundamental to understanding the emergence of new technologies, their contribution to growth, and the interaction of innovation and regulation.

This dissertation addresses how firms learn and respond to regulations in the context of emerging technologies. Within this framework, I address several questions. When production inputs are socially controversial, do firms respond to disclosure laws by voluntarily constraining their inputs? Do these public disclosure laws facilitate knowledge transmission across firms, and if so, what are the implications for public welfare - for instance, do the gains from trade outweigh any effects of reduced incentives for innovation? I study these questions in the context of hydraulic fracturing, though the results offer insight for more general settings. Panning out to a much broader view, I also explore how energy-related technologies - in both generation and consumption - diffuse across national boundaries over time, and whether innovation and diffusion of energy-efficient technologies has led to more or less energy-efficient economic growth.

In my first paper, I contribute to improved understanding of the conditions in which information-based regulations, which are increasingly common in multiple policy domains, decrease externalities such as environmental pollution. Specifically, I test whether information disclosure regulations applied to hydraulic fracturing chemicals caused firms to decrease their use of toxic inputs. Prior to these mandatory disclosure laws, some operators voluntarily disclosed fluid components for some or all of their wells. I compare the chemical mixtures used prior to the mandatory disclosure laws to those used after the laws took effect, using a difference-in-differences method motivated by the difference in timing of state-level disclosure laws. I use voluntary disclosures to measure the toxicity of fluids prior to mandated disclosure, and thus observe a composite effect of both full reporting and disclosure pressure. These effects likely have opposite signs; I employ several methods to tease them apart so that I can separately identify the effect of disclosure pressure. My analysis, which covers over 70,000 wells in seven states, suggests that state disclosure regulations resulted in a large and persistent decrease in the use of toxic and regulated chemicals in fracturing fluids. This is not the first paper to find that disclosure regulations can change firms' behavior, but it demonstrates such an effect in a setting in which consumer or market pressure is minimal or nonexistent: firms that produce undifferentiated products for an intermediate market, and disclosure policies that do not generate readily accessible or interpretable information.

The second paper tests whether disclosure laws facilitated the transmission of useful knowledge across companies. It is well established that economic agents learn about new technologies in part from other adopters, though even sophisticated firms may not take full advantage of social learning. With my co-authors, I examine whether firms took advantage of environmentally-focused disclosure laws to learn from competitors and improve productivity. We find evidence that they did: following mandatory disclosure we observe convergence in productivity per well, in production inputs, and strong evidence of a link between the two. To our knowledge this is the first study to examine this pathway for social learning in an emerging technology. This could also be interpreted as a form of technology diffusion facilitated by environmental regulation.

In my third paper, I address a broader scale of technology change, looking for evidence that improved technologies for energy generation and consumption have allowed less energy-intensive or pollution-intensive growth in developing countries. I analyze panel data on Gross Domestic Product (GDP) and national energy consumption to look for evidence of technology "leapfrogging" (i.e., decreased intensity of energy consumption for a given level of economic growth). I combine 1960-2014 data on energy consumption from the International Energy Agency with historical data that extends back to 1861 for several countries on energy consumption and fuel source, as well as GDP. I compare countries at the same income level and test whether energy consumption and energy intensity are different for today's less-developed countries compared to today's industrialized countries when they had similar income levels. Compared to prior analysis, my much longer time series allows me to test for leapfrogging over a scale appropriate to the pace of widespread technological change.