Economic statistics – Odd numbers (zum US BIP)
Heute Nachmittag erwarten wir die Revision der ersten Schätzung für das US-Bruttoinlandsproduktes für das vierte Quartal 2007. Ein Grund mehr sich noch einmal Reaktionen auf die erste Veröffentlichung Ende Januar anzuschauen.
Selbst dem ‘Economist’ kam die Publikation suspekt vor, dass von ‘odd numbers’ geschrieben wurde; am 31.1.2008:
-
Economic statistics – Odd numbers
Why America’s advance GDP figures do not paint the whole picture
AS THE old joke goes, statistics are like a bikini—what they reveal is suggestive, but what they conceal is vital. America’s advance GDP figures tend to be more of a bathing suit than a bikini: they are a bit outdated and hide as much as they show. Nevertheless, on January 30th it wasn’t just hardened statisticians whose eyes were out on stalks. Fourth-quarter numbers, showing an annualised growth rate of 0.6% (or less than 0.2% in quarterly terms), suggested that America’s economy was barely growing at the end of last year. It was lower than economists had expected.
Financial markets were initially thrown by the data. Toss in a revision or two and it is conceivable that they might have marked the start of a recession—defined as two consecutive quarters of decline.
That could easily happen. Indicators for October and November were much stronger than for December.
Many of the source numbers for December were not available for the advance estimate, so statisticians had to make an informed guess, in some places using the stronger months as a guide.
That said, revisions could go the other way. A large fall in inventories subtracted 1.25 percentage points from fourth-quarter growth, making a far bigger dent than usual in the figures. Changes in inventories, however, are often subject to large revisions. They might well be adjusted upwards later.
The first estimate of America’s GDP is notoriously imprecise, and is probably more so during times of wrenching economic change. According to the Bureau of Economic Analysis, which computes the GDP figures, around 25% of the required source data are not available for the first estimate and only partial information (for example two out of three months’ data) is on hand for a further 30% of sources.
Between 1994 and 2004—years for which figures are no longer likely to be much updated—the average (annualised) revision to the growth rate between the advance and the latest figure was 1.3 percentage points. Recently revisions have tended to be downwards. In the past five years 60% of initial estimates were later restated at a lower rate.
The picture will become clearer on February 28th, when the second estimate of GDP is released. This iteration tends to include sharp revisions, because almost all of the final month’s numbers are available.
Some statisticians argue that there is a case for postponing the GDP data, as countries such as Canada and Australia do, until it is comprehensive. But given how jumpy the markets are at present, a gradual disrobing may be gentler on everyone’s nerves.
‘Merkwürdige Daten’ wollte das Bureau of Economic Analysis nicht auf sich sitzen lassen; der Direktor des BEA sah sich zu einer Antwort veranlasst:
15.2.2008
Statistical proof
SIR – Your article on measuring America’s economic growth (“Odd numbers”, February 2nd) argues that the “advance” estimate of GDP reveals too little about the American economy, stating that it is “more of a bathing suit than a bikini”. I would like to clarify the purpose and accuracy of early GDP estimates.
The advance estimate of GDP provides a quick reading on the economy. According to the Federal Reserve and OECD, this early snapshot is among the timeliest and most accurate estimates of GDP in the world.
The Bureau of Economic Analysis uses a well-established process that incorporates more complete and accurate data as they become available. We regularly publish revisions and the record shows our early estimates may be amended up or down. However, they typically tell the same story of high or low growth, an accelerating or decelerating economy and sector-specific growth contributions.
All economic statistics face a trade-off between timeliness and accuracy. America’s GDP figures strike a responsible balance.Steven Landefeld
Director
Bureau of Economic Analysis
Wir dürfen zur Kenntnis nehmen, dass die Veröffentlichung des US BIP nicht nur zu den zeitnahesten, sondern auch zu den genauesten der Welt zählen soll.
Im Chartarchiv etwas gewühlt, ergab folgendes interessantes Ergebnis, die erste Revision des BIP für das dritte Quartal 2003 ergab eine ‘Neubewertung’ der Lagerbestände, sodass letztendlich ein Wachstum von +8,2% annualisiert zum Vorquartal, statt der vormals +7,2% ausgewiesenen, gemeldet wurden. Leider nur unzureichend dokumentiert, konnten wir uns in der Vergangenheit häufig genug die Augen reiben, ob dieser abweichenden Einschätzungen.















