Reuters Breakingviews delivers agenda-setting financial insight. Its global correspondents react to stories as they develop, delivering sharp and provocative commentary on big financial news as it breaks. Click here to read more international insights.
GDP is probably the single most important economic statistic around. In theory, it makes sense for the Bureau of Economic Analysis (BEA), which calculates this measure of output in the United States, to keep up with the statistical state of the art. In practice, the changes in the first comprehensive review of U.S. national accounts since 2009 are mostly counterproductive.
GDP compilers are torn between simplicity – just record what it costs – and economic accuracy – measure what it is worth. The simple treatment of research and development expenses, for example, is to expense them when they occur. The economically accurate way is to amortize them over the years that the R&D adds value. That slower recognition of costs leads to a larger value of current output.
The changes will sacrifice simplicity for the sake of what is supposed to be greater accuracy. Most notably, a move to amortizing R&D will increase current annual GDP by $300-billion (U.S.), or 2 per cent. Expenditures on films, long-lived TV programs, books and other artistic originals will also be capitalized and accrual accounting will be applied to defined benefit pension plans, some other types of deferred remuneration and the transaction costs of buying houses.
Theory is fine, but all these changes involve huge estimation problems. For example, the BEA will write off 3.8 per cent of the cost of movies each year. That implies a highly implausible average economic life of 26 years. The slow amortization also increases measured GDP. If this methodology were applied to Hollywood's 1930s "Golden Age", it might have made the Depression much less Great. Fortunately, it will be used only from 2007 onward.
Skeptics have long worried that the national accounts use too many dubious techniques, for example the "hedonic adjustments" which are supposed to recognize quality improvements, and the lax assumption that pay is the best measure of output for government employees. Both of those make it easier to report strong growth in real GDP.
The latest set of adjustments will make GDP less transparent, and easier for any politically minded statisticians to manipulate.