"He uses statistics as a drunken man uses lamp posts — for support rather than illumination," as Scottish writer Andrew Lang said of (it is thought) Bernard Baruch.
The same might be said of the US economy amid the current welter of conflicting data.
Take, for example, the latest US retail sales figures, released by the Commerce Department. If these are to be believed, the nation's stores and autodealers experienced an all-singing, all dancing Christmas which contributed to a miraculous, forecast-busting 1.2% rise for December and a 8% increase for the year.
Yet dig a little deeper, strip out car sales, and retail sales rose by just 0.3% on November 2004 figures.
Just over week ago the International Council of Shopping Centers bemoaned that same stores sales for the holiday period failed to fulfil expectations, rising by 2.7%, well short of the hoped-for 3.5%. [WAMN: 10-Jan-2005].
But the National Retail Federation announced a 5.7% rise in holiday sales in 2004, beating both its own forecast of 4.5% growth and the 5.1% rise in 2003. Consumer sales for the year reached $4.06 trillion (€3.07tr, £2.15tr) - something, surely, to cheer.
But the Labor Department reported a surprising jump in the numbers of unemployed people. Those filing initial jobless claims rose to 367,000, the highest rate since September. The department is quick to point out, however, that longterm claims have fallen to their lowest levels since 2001.
Says Ken Perkins, head of RetailMetrics: "There was a lot of conflicting data out there, more so than I've seen in the last five years."
Ain't that the truth.
Data sourced from BBC Online; additional content by WARC staff