Dumbing down the data

If macroeconomics isn't your main thing, you may have missed out on the big debate that's going on about the degree of slack in the American economy.
Come back, come back - it's important! If there's lots of slack still around, then there's no issue of inflation on the horizon, and the Fed's probably going to stay with its ultra-supportive monetary policy for some time yet. On the other hand, if the output gap is closing up fast, or has even closed up such that we're already somewhere near the NAIRU, then rising inflation is a real possibility, and the Fed will take back some stimulus faster than the financial markets currently expect. Which would have very large consequences for global financial markets: many asset classes have been carried up to expensive levels by the Fed's high tide of liquidity. The stratospheric levels of the NZ$ (in particular) and the A$ (to some degree) are also artefacts of the Fed's money printing (increased supply of US$, lower levels for the US$), so we've got a dog in the race, too.
There's a good summary of various views in this blog roundup from Bruegel, the European think tank, if you'd like to get up with the play on the substantive issue.
I've got a different point I'd like to talk about, and that's the role of the statistical agencies in helping us understand where economies are at.
By way of background, the US GDP figures, which are obviously the key data in all of this, come from the Bureau of Economic Analysis, which has three goes at it - an advance estimate not long after the quarter, a second estimate roughly two months after the end of a quarter, and a third estimate around a month after that. The BEA, by the way, reports US GDP as a (seasonally adjusted) annualised rate, rather than the quarter on quarter change we use in these parts. Enough background, on to the action.
On April 30, the BEA released the advance estimate for March quarter GDP. It showed that GDP was marginally higher: growth in the March quarter was +0.1% at an annualised rate, which was outside the range of advance forecasters' estimates, which ran from +0.5% to +2.0%. On May 29, the BEA released the second estimate, which was an annualised decline of -1.0%; the surprise number was again outside the range of advance estimates (-0.8% to +0.2%). And on June 25 it followed up with its bombshell third estimate, which made the decline a much more substantial -2.9%, and once again outside the range of advance guesses (-2.4% to -1.0%).
Three surprise numbers in a row on probably the most globally important macro statistic, would, you'd think, warrant some sort of explanation from the BEA.
What did we get?
Sweet FA is what we got. The only thing that's remotely relevant is the (boilerplate) statement that the second estimate "is based on more complete source data than were available for the "advance" estimate" and that the third estimate - wait for it - "is based on more complete source data than were available for the "second" estimate issued". Bear in mind that the underlying reason for the string of wildly unexpected numbers was the incredibly awful northern hemisphere winter: was there any analytical attempt to figure out how much that might have been responsible? No. Almost unbelievably, there wasn't even any mention of the weather at all.
You'll have your own favourite expression for exasperated incredulity: where I grew up, Holy Mother of God! was quite popular.
Now, I know that statistics in the States is a political battlefield. Nutters claimed that the official unemployment statistics were being manipulated in the run-up to President Obama's re-election, for example, and even how the US Census is done has drawn political flak (because if done properly it might affect the size and ethic composition of electoral districts). So if I were the BEA confronted by the Looney Tunes that is the present-day US Congress, I might veer away from offering interpretive analysis, too.
But even if the BEA didn't have the dingbats to face down, a lot of statistical agencies don't believe that providing interpretation or analysis is properly within scope for an official national statistics office. Their releases basically amount to reading out the numbers in the tables. It's not just the BEA that's down the read the numbers end: if you want to see an almost comically typical example, read this Australian Bureau of Statistics release on March 2014 corporate profits.
Frankly, in my view, this won't do any more. There are better ways, and I'm pleased to say our own Stats folks are among the more enlightened. Here, for example, is a little paragraph that was included in this week's CPI release (it was in this bit of the release):

Data influencers

Price changes may be influenced by specific events. Factors that affected the June 2014 quarter CPI are listed below. 

• The annual increase of 10.2 percent in cigarette and tobacco prices was influenced by an 11.28 percent rise in excise duty from 1 January 2014.
• The strong New Zealand dollar has had a downward influence on the retail prices of internationally traded goods including cars and appliances.
• The rise in the price of vegetables was influenced by seasonally higher prices for tomatoes, lettuce, and cucumber.
• The fall in the price of fruit was influenced by seasonally lower prices for kiwifruit and apples

See - that wasn't so hard, was it? So why don't more agencies do the same thing?