>Global: House of Mirrors
>
>Stephen Roach (New York)
>
>Lest I be accused of piling on, read no further if you're looking for
>the next WorldCom. I don't have a clue. But I do know that Corporate
>America is not alone in cooking its books. Washington statisticians seem
>poised to join the restatement sweepstakes with a stunning rewrite of
>the recent performance of the US economy. So much for the boom!
>
>Each July, when many of us head to the beach, the guys with the green
>eyeshades are hard at work in Washington. They are compiling the
>so-called benchmark revision of the national economic statistics -- an
>annual restatement of recent economic history based largely on more
>complete (and presumably more accurate) samples of underlying activity.
>This particular benchmark revision is slated to be released on 31 July.
>Mark that day on your calendar.
Well the numbers are out, and the boom was less boomy than we thought. Some excerpts from the BEA release follow; the full product is at <http://www.bea.gov/bea/newsrel/gdpnewsrelease.htm>. The last couple of paragraphs deal with the arcane matter of the statistical discrepancy, which I normally wouldn't bring up, except that in this case it embarrasses Alan Greenspan. The full name of the GDP exercise is the national income and product accounts (NIPAs). There are two sets of estimates - one of product (expenditures on consumption, investment, etc.) and one of incomes (wages, profits, etc.). In theory, the two accounts are supposed to match; every dollar of income is earned in production. In practice, they don't quite match, and the difference is called the statistical discrepancy. The statistical discrepancy had been unusually large in the last few years, for reasons no one quite understood; income was running well ahead of product. The BEA says the product number is more reliable, and they feature it in their statistical releases (e.g., many people have heard of GDP [gross domestic product], but almost no one has heard of GDI [gross domestic income]). But New Economy optimists, chief among them Greenspan, focused on the income figure, saying that it might be the more reliable number. They were wrong. The income numbers have been revised down significantly, and the statistical discrepancy has returned to more normal levels. But since Greenspan is the greatest central banker of all time, this bit of irrational exuberance won't be held against him.
Doug
----
>Summary of major revisions
>
>The revised data modify the quarterly pattern of real GDP for
>1999-2001. Both the previously published and the revised estimates
>show GDP growth peaking in the fourth quarter of 1999 and slowing
>substantially during the quarters of 2000. However, the revised
>estimates show declines in GDP for each of the first three quarters
>of 2001, whereas the previously published estimates showed positive
>but decelerating growth in the first half of 2001 and a decline in
>the third quarter. Both sets of estimates show GDP growth resuming
>in the fourth quarter of 2001.
>
>The most important differences between the revised and the
>previously published estimates for 1999-2001 are the following:
>
>The annual rate of growth of real GDP from 1998 to 2001 was revised
>down from 3.1 percent to 2.7 percent. The annual rate of growth of
>real GDP from 1998:IV to 2002:I was revised down from 2.8 percent to
>2.4 percent. The largest contributors to the downward revisions were
>downward revisions to the growth of personal consumption
>expenditures (PCE) and of nonresidential fixed investment.
>
>For 2001, the revised estimates show real GDP growth of 0.3 percent;
>the previous estimate was 1.2 percent. Lower growth of PCE and
>larger declines in nonresidential fixed investment and in change in
>private inventories accounted for most of the revision.
>
>As described above, the revised estimates show a longer downturn in
>real GDP than the previously published estimates. The percent change
>at an annual rate in real GDP was revised down from 1.3 percent to
>-0.6 percent for the first quarter of 2001, was revised down from
>0.3 percent to -1.6 percent for the second quarter of 2001, and was
>revised up from -1.3 percent to -0.3 percent for the third quarter
>of 2001.
>
>For 2001, personal income was revised down 0.4 percent. Wages and
>salaries was revised down 2.9 percent, and personal interest income
>was revised up 9.8 percent.
>
>[...]
>
>The statistical discrepancy is current-dollar GDP less
>current-dollar gross domestic income (GDI). It arises because most
>components of GDP and of GDI are estimated independently. GDP
>measures final expenditures -- the sum of consumer spending, private
>investment, net exports, and government spending. GDI measures the
>incomes earned in the production of GDP. In concept, GDP is equal to
>GDI. In practice, they differ because they are estimated using less
>than perfectly consistent source data.
>
>As a result of the annual revision, the statistical discrepancy as a
>percentage of GDP was revised from -0.8 percent to -0.4 percent for
>1999, was revised less than 0.1 percentage point at -1.3 percent for
>2000, and was revised from -1.5 percent to -1.2 percent for 2001.
>The revision to the discrepancy for 1999 primarily reflected a
>downward revision to GDI. For 2000 and 2001, downward revisions to
>GDI were partly offset by downward revisions to GDP.
>