Contributing Editor
Gordon Reikard is a statistician at Leap Wireless. His research interests include the economics of growth and application of statistical methods in the physical sciences.

As the economy moves from initial recovery to more sustained expansion, one of the key policy issues is whether more investment in technology will boost the growth rate. The issue takes on increased importance because the expansion has been anemic so far. Looser fiscal and monetary policies successfully stimulated a recovery from the 2008–2009 recession, but an increase in the rate of technological advance is needed to propel the economy into a more lasting expansionary phase.

It is only recently that policymakers have recognized the relationship between technology and growth; President Obama’s State of the Union speech explicitly mentioned it. Historically, many of the policy debates in Washington have been about taxes and regulation, and these debates have often been phrased in terms of the implications for growth. Unfortunately, these debates have tended to sidestep the more fundamental issue. Removing regulatory barriers and cutting taxes will, in general, result in a one-time increase in the level of gross domestic product (GDP), but not necessarily influence its long-term rate of change, which depends critically on the rate of technological progress.

One of the most significant findings in the last 50 years is that a large share of economic growth—more than one-third—is driven by technological advance. This originated with a seminal 1957 paper by Robert Solow titled “Technical Change and the Aggregate Production Function” that was published in Review of Economics and Statistics. Solow demonstrated that capital and labor accounted for less than two-thirds of growth. The remainder was technology.

Recent estimates indicate that, since the late 1940s, about two-fifths of growth can be attributed to technology. The standard, broad measure of technological advance is total factor productivity (TFP), which is the residual calculated by subtracting the contributions of labor and physical capital from GDP. In the short run, TFP is volatile, but the fluctuations average out over long periods of time. At lower frequencies or over longer periods, the trend in TFP measures the rate of technological advance. Using the Bureau of Labor Statistics’ estimates, technical advance has contributed to 38% of growth since 1948.

But there were obvious problems with the Solow model. TFP is, by construction, a residual, rather than a direct measure of technology. In principle, it can encompass everything from new products or better products to process improvements, reallocation of resources, and increases in efficiency. So, since the 1950s, there has been a great deal of empirical work—much of it by economic statisticians—linking TFP to observed measures of technology. These studies have ascertained that TFP is highly correlated with patents and indexes of scientific knowledge.

There was a major new theoretical development in the 1980s, when Paul Romer of Stanford argued that technological advance depended on investments in research, or human capital. Romer’s model implied that rates of technical advance could fluctuate over longer periods, which was confirmed by Romer and statisticians George Evans and Seppo Honkapohja in “Growth Cycles,” published in American Economic Review. Decades such as the 1960s and 1990s were prosperous because they enjoyed faster rates of technological progress.

The strongest empirical evidence, however, was for research and development (R&D), although even the highest estimates indicate R&D can account for only a fractional ratio of technological advance. More specifically, R&D contributed as much as 75% of technical advance from roughly 1948–60, and much of this was propelled by government-funded defense research. By the 1960s, however, process improvements in the private sector had largely displaced R&D as the major source of technical advance. During the second wave of high productivity, beginning in the 1990s, R&D contributed only about one-fourth of technological advance and private sector R&D was increasingly more important than government-funded research.

The estimated elasticities of R&D in the production function support this. Over the last decade, the elasticity of industry-funded R&D has risen to 0.064, while the elasticity of government-funded R&D has declined to 0.049. So while the combined elasticity for all R&D works out to 0.112, industry-funded research is having a greater impact. Given this, governments can realistically expect to raise long-term growth rates through policies such as tax credits for R&D.

In the early 1990s, many economists were pessimistic about the ability to raise growth, but by the middle of the decade, productivity was achieving gains not seen since the 1960s. Clearly, technological advance was
speeding up. The causes had to do with computers and software.

The reason computers and software had such a powerful influence was that their effect was not limited to a single industry. Information technology (IT) could generate substantial spillover effects into other sectors. Examples include local area networks, computer-aided design (CAD-CAM), electronic banking, Internet retailing, statistical quality control, computerized inventory control, and faster communication of ideas. Industrial firms could use computers to reduce cycle times, achieve fewer defects, control inventory, and do specialized production runs tailoring manufacturing to demand. Computers not only made industrial processes more efficient, they made research itself more efficient, since R&D could now be performed with advanced software, leading to faster development and better products.

In sum, the role of technology in growth is large—about two-fifths—and will continue to increase in coming years. While R&D has historically comprised a significant share of technical advance, an increasing share is now coming from computers and software. Therefore, a two-pronged approach is needed. The government should adopt a permanent and, preferably larger, tax credit for R&D. It also should allow first-year expensing of all computers and software and exempt both the R&D credit and depreciation deductions from the alternative minimum tax.

In essence, the current period of slow growth can be overcome. If the country commits itself to investing in technology, the coming decade could be much like the 1990s.