BBC News


Page last updated at 08:50 GMT, Tuesday, 12 August 2008 09:50 UK

The myth of counting

Sale signs

The economy is big news at the moment, but it isn't always easy to see whether a particular statistic is good or bad. Wouldn't it be good to have the mental agility to separate the wheat from the chaff? In his second lesson of a weekly series, author Michael Blastland gives some hints for comprehending growth.

Michael Blastland
Counts and measurements swamp the news - of everything from the growth (or slump) of the economy to the extent of yobbish behaviour.

They're often reported as if straightforward child's play, like counting your toes. That couldn't be more wrong - the world is a massive mess; the task more like counting an ocean of mushy peas.

Lesson Two: Counting

The story: The economy is on the edge. GDP growth sank to a lowly 0.2% last quarter. Slip below zero (ie the economy starts shrinking), stay there six months, and we're officially in a recession. Cue doom all round.

The flaw: GDP is not counted. Most things about the economy or society, in fact, are not counted. And though you would never know from the fanfare that often accompanies GDP figures, or the way we're hanging on the next set, they are usually wrong.

And not only wrong, but wrong in the same direction - too low, typically - often badly. The first "count" of GDP has sometimes been so misleading as to be worse than useless.

The lesson: We tend to talk about big numbers describing the economy or society as if there were legions of bureaucrats with clipboards, standing outside factory gates or peering into our living rooms, counting. Better still, some kind of computer to do it for them.

What we actually do is sample, make some guesses, and extrapolate. Sampling is a brilliant solution to an intractable problem - you can't count everything, there's just too much. But it can go badly wrong. How wrong you can see from the chart.

GDP estimates

The red line shows the GDP figure that gets the fuss - the first estimate. The green zone around it shows the range of updated estimates as more data comes in. The blue line shows the latest (and best) estimate. The differences are sometimes huge.

In Quarter 1 of 2000, for example, the number everyone seized on was quarterly growth of 0.4%. The latest best estimate is that it was really 1.1% (and at one time was believed to be 1.4%). That is, the economy was growing nearly three times faster than first reported.

the mountains made of the point when we enter "recession" are made of sand

This is a big error - apparently repeated (though not always on that scale) most of the time. Add up these revisions over a whole year and you have an object lesson in mass-delusion. Growth is usually - though not always - significantly higher than the first impression given by media concentration on that initial, ropey estimate.

The numbers are judged more reliable only after about three years, when we have tax returns and other more complete data to check. So although the lines are closer together in recent years, they may yet move further apart. If they do, don't expect to read much about it.

Piggy bank smashed
Times are predicted to get hard, but how certain are we?

This also turns upside down another story - how our economy compares with the US. Because the US is unusual in its tendency to overestimate growth, then revise down.

So first appearances have often suggested the US economy was racing ahead, and we have filled newsprint with arguments about how to close the gap. Sometimes it really is growing faster (though mainly because its population is also growing fast). Often, after revision, there is no gap.

Junk rating: Two out of five. The first numbers are not reliable, but people don't count GDP this way because they are stupid. Counting output is an awful lot harder than it looks (though some other countries seem to do better). It's still the best data we've got.

And this doesn't invalidate the general conclusion that the economy is growing more slowly than before. It's unlikely that the numbers are that wrong, this time. But the mountains made of the point when we enter "recession" are made of sand.

In truth, we could be in a recession already, or might avoid one altogether, even as the figures say otherwise. Watch for the headline: "Recession, it's official!" then blow off the froth of media excitement, and wonder what makes us want to appear so certain about an uncertain world.

As for those trying to steer the economy, like the Bank of England setting interest rates, or the chancellor of the exchequer, they might know where they want to get to, but not even they really know where they are.

Next week, Lesson Three: Percentages

Michael Blastland is the author, with Andrew Dilnot, of The Tiger That Isn't.

A selection of your comments appears below.

Great - finally an article that talks some sense. It is so refreshing to see someone write an article that is not one-sided sensationalism. Myself and so many people I know would really like to find a news source such as this for all stories which tries to be objective and tell it how it is instead of either "doom and gloom" or "over optimism".
Nick T, London

Great - this is long overdue for highlighting - pity most people will still not understand. Makes you wonder about 'the environment'... ?
Marshal Jim Duncan, Dundee

Excellent article. It's wonderful to see an item like this as most people aren't prepared to read a book like "Freakonomics". I'm also glad to see it isn't just folks in the US that are so easily mislead by poor reporting and even poorer numbers. :)Spoonman, Upstate, NY USA

House prices are a good one. Apparently they fell 4% last month. Well add up all the houses sold and divide by the number and you get the average? Well no, house prices themselves may be completely unchanged, only the cheap ones sold. Larger houses sat there with sellers refusing to drop prices. The is proved by the "record low number of houses sold". Then this month "house prices more realistic" shows they finally HAVE dropped. The most telling piece was the fact we went from record high mortgage borrowing to record lows in a couple of months. But these facts were hidden on tiny articles, its HOUSE PRICES TO CRASH that is the headline. This distorts the truth. House prices are not falling, no-one will drop their value and no-one will pay to buy, the market is frozen.
Nich Hill, Portsmouth UK

Has China's housing bubble burst?
How the world's oldest clove tree defied an empire
Why Royal Ballet principal Sergei Polunin quit


Americas Africa Europe Middle East South Asia Asia Pacific