Academic journal article Research-Technology Management

New Hope for Measuring R&D Effectiveness

Academic journal article Research-Technology Management

New Hope for Measuring R&D Effectiveness

Article excerpt

There is no question that R&D is important both to firms and to the economy (1). For U.S. firms in the 25 most R&D-intensive industries, R&D expenditures comprise 5.8% of annual firm expenditures and the associated intangible assets comprise 5.6% of firm market value (2). Moreover, R&D is believed responsible for 7% of real U.S. GDP growth (3). Given its importance, it is rather surprising that firms and policy makers have to make R&D investment decisions using "visual flight rules"; because firms lack good instruments, they must use their intuition to assess the effectiveness of R&D and to set R&D budgets.

The questions of R&D effectiveness and the appropriate R&D budget are actually two sides of the same coin. In an ideal world, firms would invest at the level of R&D where the marginal benefit of a $1.00 investment equaled $1.00. While identifying this level is simple in principle, it has been difficult in practice due to limitations of existing measures. For example, a recent survey of the Industrial Research Institute members (4) indicated that 51.9% of firms were using R&D Investment/Sales (an input measure rather than an effectiveness measure), and for those using measures of financial returns, there were problems of uniformity, gaming and transparency to shareholders. Even the best of these returns measures, when calculable, are average rather than marginal measures. The firm would not enjoy the same return if it doubled its R&D.

Without a good measure of effectiveness, firms must guess the optimal amount of R&D. A recent study shows that in general they tend to guess incorrectly (5). Figure 1 plots actual R&D spending (y axis) versus the optimal amount of R&D (x axis) for publicly traded U.S. firms in the top 25 R&D-intensive industries. If firms were investing optimally, the data points would be clustered around the diagonal. Instead, approximately 3/4 of firms over-invest, while 1/4 of firms under-invest.

What is perhaps more interesting is that the stock market penalizes firms for over-investment. Each 10% increase in R&D above the optimum decreases market value 6%. In essence, the stock market seems to have a better sense of optimal R&D than firms themselves.

Until now, academics have been unable to help firms assess R&D effectiveness or choose optimal investment because their primary measure of R&D effectiveness (patent counts) wasn't superior to those used by firms. Patents are neither universal (fewer than 50% of firms engaged in R&D file patents in any given year), uniform (there is high variance in patents' value, e.g., the patent for copying DNA versus the 97% that are never commercialized), nor reliable (increasing patent counts has no impact on the stock price).

The IQ Solution

Recent advances in statistical software have enabled the creation of a new measure of R&D effectiveness that is universal, uniform and reliable. Moreover, it can be calculated from firms' financial data. I call the measure IQ because it is normally distributed across firms and because it captures firms' technical problem-solving capabilities in much the same way that individual IQ captures individual analytical problem-solving capability: those with higher IQ solve more problems per unit of input (dollars for firms, minutes for individuals) than those with lower IQ.

What Is IQ?

For those familiar with economics, IQ is the output elasticity of R&D holding all other inputs (capital, labor and advertising) fixed. The output elasticity defines a curve that relates R&D spending on the x axis to revenues on the y axis. The way to interpret IQ is a firm's ability to generate revenue from its R&D investment. Thus, a firm can have high IQ either by generating a large number of innovations and being reasonably effective exploiting them, or by generating a reasonable number of innovations and being extremely effective exploiting them. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.