Summary, March SGE Monthly Luncheon
Summary of Remarks by Carol Robbins
"How Should We Treat R & D Expenditures in National Accounts? New Thinking from the BEA"
By Andrew Felton, FDIC
At the March luncheon, Carol Robbins of the Bureau of Economic Analysis (BEA) presented an overview of the BEA’s current work on integrating research and development (R&D) into the national accounts.
Why is this important? First of all, economic theory says that we should. R&D is critically important to economic growth. Second of all, R&D is a large and important sector of the economy. Some economists have estimated that businesses spend more than $1 trillion on intangibles per year (Corrado, Hulten, and Sichel (2004)).
The BEA has established a Research and Development Satellite Account. Satellite accounts are not officially part of the national accounts but present a framework that can provide a different perspective on economic activity that is consistent with the other accounts. For example, there is a Travel and Tourism Satellite Account (TTSA). They are experimenting with environmental and health innovation satellite accounts as well. BEA is also investigating the possibility of developing a satellite account that would include copywrited material such as books and movies.
Some types of R&D are more amenable to measurement than others. BEA is focusing on technology, pharmaceuticals, software, trade secrets, patents, and other quantifiable things. There are three conceptual challenges:
- Defining the unit of R&D
- Developing an R&D price index
- Accounting for depreciation
One way to measure the value of R&D produced is by measuring its impact on corporate profits. However, this is extremely difficult to do, so BEA is using expenditure data – specifically, a National Science Foundation survey on R&D spending. This also helps deal with the fact that R&D is in some ways a public good and its value would not be fully captured by its impact on corporate profits.
Even more difficult in many ways is developing a price index. Using the aggregate GDP deflator is unappealing because it covers many sectors other than R&D. Another way to construct the index is by measuring the change in prices of R&D inputs. This actually tracks the GDP deflator fairly closely, and would be similar to the way that BEA currently measures price changes for government services. However, without measuring output, we cannot measure productivity gains in R&D. A third way is to use prices of R&D output, based on the Producer Price Index as produced by the Bureau of Labor Statistics. However, this index has fallen dramatically since the early 1990s and would therefore imply improbably high depreciation of R&D.
This brings us to the third difficulty: depreciation. Right now, BEA is using estimates from the academic literature, ranging from 11 percent per year for chemical manufacturing to 18 percent per year for transportation equipment manufacturing.
Finally, R&D has some aspects of public good presents measurement problems. For example, suppose a firm engages in R&D in the United States and then “exports” that R&D to its factories in another country. Should that count as an export and therefore a decrease in the stock of R&D in the United States? For the regional accounts within the United States, this is even more of a problem, because a firm’s R&D activity may be difficult to confine to a single state.
All these caveats aside, the latest BEA estimate is that R&D investment was $317 billion in 2004. This raised the level of business investment by 10.6 percent and the national saving rate rose by 2.7 percentage points. Recognizing R&D as an investment raises the level of GDP by an average of 2.9 percent per year between 1959 and 2004 and raises the average real GDP growth rate from 3.2 percent to 3.3 percent between 1995 and 2004. R&D accounted for 7 percent of real GDP growth from 1995 to 2004, about the same as the investment in computer technology.
In 1987, the motor vehicle industry was the top R&D investor, while in 2004; the pharmaceutical industry was, accounting for 20 percent of total R&D investment. R&D has increased the level of value added by the pharmaceutical industry by 38 percent between 1987 and 2004; other industries dependent on R&D for growth include computer manufacturing (30 percent increase in level of value added) and semiconductor manufacturing (26 percent increase). But R&D is concentrated in a few industries: it has only added 0.7 percent to the level of value added to the economy outside of the 13 most R&D-intensive sectors. Since the 1960s, there has been a shift from the government doing most R&D investment to businesses doing most of the investment.
More information is available at http://www.bea.gov/industry/index.htm#satellite.
References:
Corrado, C, C. Hulten and D. Sichel. (2004). "Measuring capital and technology: an expanded framework," Finance and Economics Discussion Series 2004-65, Board of Governors of the Federal Reserve System.