Innovations, Patents and Cash Flow

Studies of the innovative activity of firms which focus on R&D spending generally conclude that relatively few firms engage in R&D, but that those who do display a relatively stable pattern of spending on R&D over time. Differences between firms in R&D spending (or in spending intensity) are typically much more important than variations in spending within firms over time. As a consequence, the interesting research question to be addressed in these studies is: ‘Which firms do R&D?’ By contrast, studies of innovative output using patents or counts of major innovations generally show that many of the firms who do produce an innovation do so only sporadically. Few firms put together multi-year spells of sustained patent or ‘major’ innovation production, and the timing of their innovative activities is episodic, idiosyncratic and relatively hard to predict. This adds a second interesting question to the research agenda, namely: ‘When (if ever) do firms innovate?’ At a purely statistical level, answering this second question means finding exogenous variables that display the same kinds of variation as patents or ‘major’ innovation counts; that is, finding independent variables that display a high ratio of within to between variation, plus a tendency towards irregular bursts of sub- or supernormal activity. For economists interested in the determinants of innovative activity, this is likely to rule out factors such as ‘technological opportunity’, ‘conditions of appropriability’ or market structure (all of which tend to be different for different firms or industries, but stable over time). However, potential determinants of innovation, such as demand or the financial state of firms, are both interesting on theoretical grounds and potential candidates on purely statistical grounds.