Studying innovation systems and dynamics using an empirical approach is challenging, as innovation comprises both tangible and intangible outputs

(e.g., new technologies, machines, products, patents but also ideas, process innovation, managerial, and organizational innovation). Following a large empirical innovation economics literature, we study the more tangible and measurable aspects of the innovation process. While these constitute only a part of the innovation output relevant for the energy system and sector, they nonetheless provide important insights that can complement those from qualitative and bottom-up case studies focusing on more intangible and less measureable aspects such as organizational innovation.

Here we focus on R&D expenditures and patent counts.[1] The former informs on the inputs of the innovation process, while the latter is a proxy of innovation outputs. Both indicators suffer from some specific shortcomings. R&D investments provide insights on innovation effort but not on innovation quality. Conversely, patent statistics provide a partial measure as not all innovations are patented, even though they can be weighted using information on several indicators to control for quality (for instance, claims or citations, see Griliches 1990). Furthermore, patents may increase due to changes in patent law or strategic reasons to signal in which companies to invest in, regardless of innovative activity (Mazzucato 2013).

In the case of energy innovation, matters are further complicated by the fact that it is unclear how clean innovation or even energy innovation are defined (Gallagher etal. 2011). A number of studies focus specifically on the energy supply sector (Salies 2010; Sterlacchini, 2012; Costa-Campi etal. 2014) but energy-saving R&D and innovation are pervasive. Energy is an input for nearly all sectors of the economy and the way in which energy is produced, transformed, and distributed depends on innovative activities well beyond those of the energy supply sector itself. All R&D expenditures are inputs into complex processes that ultimately lead to innovations that may or may not be clean. In order to proxy for industrial R&D investments in energy, we rely on the Analytical Business Enterprise Research and Development (ANBERD) database (OECD 2016), which provides information on the R&D expenditures at the sectoral level for 30 countries for the years 1990-2013.[2] We define energy R&D investments in two ways. First, we focus on R&D spending in the ‘Electricity, water and gas distribution industry’, which represents the downstream sector for energy production (power R&D). Second, we define energy investments as a combination of R&D expenditures from ‘Electricity, water and gas distribution industry’ and ‘Mining’, which capture the combined R&D effort in the upstream and downstream energy supply sector (energy R&D).

These measures arguably represent a lower-bound estimate of energy- related innovation (Upstill and Hall 2006), as they only include the R&D directly performed by the energy supply sectors. Indeed, non-energy sectors indirectly contribute to energy-related innovation. For instance, improvements in the manufacturing of chemicals and chemical products, and in computer and electronics, contribute to the development of energy system technologies, such as solar power or smart grids. These are ‘embedded’ in the capital that is supplied to the energy supply sector. The sum of the direct and ‘embedded’ R&D can be considered an estimate of the upper bound of industrial energy-related R&D in a given country. Input-output data can be used to provide an estimate of energy-related R&D expenditures including the research performed in other economic sectors that are embedded in the capital purchased by the electricity and the mining sectors. Dasgupta, De Cian, and Verdolini (2016) provide a detailed description and application of this method.

While providing insights on the extent of energy innovation efforts, the ANBERD statistics have some shortcomings. For instance, they report R&D expenditure by sector of performance expenditure, regardless of whether funds were provided by the private or by the public sector. This means that industrial R&D reported by ANBERD statistics might include a fraction of R&D expenditure funded by the government and therefore reported in the government budget outlays as well. For this reason, we refer to the R&D reported in the ANBERD statistics as industrial rather than private R&D.

Another widely used proxy for innovation is patent counts, which is an indicator of the output of the industrial R&D process (Griliches 1990).[3] The temporal and country coverage of patent data is often broader than that of R&D statistics and makes it an attractive empirical proxy. In the specific case of energy-related innovation, a further advantage of using patent data is the possibility of assigning patents to specific energy technology classes in the energy sector, which also include renewables (Johnstone, Ivan Hascic, and Popp 2010) and efficient fossil-based technologies for electricity generation (Lanzi, Verdolini, and Hascic 2011). We collect patent statistics from the Organization for Economic Co-operation and Development (OECD) Patent Statistics Database (OECD 2015b) and count applications through the Patent Cooperation Treaty (PCT) by the inventor country and priority date. The technologies included in our patent counts are the following:[4]

  • (1) Power Patents: related to energy generation, they include both energy generations from renewable and non-fossil sources and technologies improving the efficiency of fossil fuels, such as Integrated Gasification Combined Cycle and improved burners. Both renewable and fossil- efficient technologies have significant mitigation potential (IEA 2014).
  • (2) Green Patents: include power patents as well as the patents in the technology domains of general environmental management, technologies specific to climate change mitigation, energy efficiency in buildings and lighting, technologies with potential or indirect contribution to emissions mitigation, emissions abatement, and fuel efficiency in transportation.

  • [1] Arguably, R&D investments and patents represent only part of the full innovation process,as they somewhat disregard the issue of technology diffusion. Specifically, patent data is animperfect indicator of technology diffusion, but nonetheless widely used in the literature toproxy for the other, earlier stages of innovation (see, for instance, Hall and Rosenberg 2010).
  • [2] Our analysis focuses on the 20 countries between 1995 and 2010 for which both policy andinstitutional data are available.
  • [3] Indeed patents are positively correlated with power R&D (correlation coefficient for powerpatents is 0.41 and 0.62 for environmental patents) and with energy R&D (power patents 0.50and environmental patents 0.44).
  • [4] Please refer to Hascic and Migotto (2015) and OECD (2015a) for more details about thetechnologies included.
< Prev   CONTENTS   Source   Next >