In Data We Trust

12/10/2012 at 5:01 pm

The big economic news in the US of late has concerned jobs. Last Friday it was announced that the unemployment rate dropped to 7.8% in September from 8.1% the month before, the lowest level since January 2009. Some thought this was too good to be true. Jack Welch, legendary former chairman of GE, was suspicious that such positive news should come two days after a presidential debate which the incumbent was widely seen to have lost, accusing “the Chicago boys” of having “changed the numbers”.

Inevitably, Mr Welch was accused of cracking up. But it is true that governments do fiddle the figures, or perhaps just as worryingly, have to fight hard to make sure they’re not being exaggerated.

In Argentina for example, inflation was running at an uncomfortably high level a few years ago. The former president took decisive action.

In January 2007, key staff in the statistics office started to be replaced.

A year later, inflation had fallen from 9.7% to 8.2% – proof that the strategy worked. Except, that is, according to those who kept an unofficial eye on consumer prices and whispered that the true rate was nearer to 20%.

Sometimes it’s the statisticians who need policing. At around the same time as the former President Kirchner was improving the quality of Argentine price measurement, the Chinese government set about toughening the laws on falsifying data:

Data falsification has long been a problem among Chinese officials, who seek to meet government targets to qualify for promotions. In 2007, nearly 20,000 violations “were uncovered,” China Daily reported in April. In one case, the NPC [National People’s Congress] reportedly discovered that officials in Chongqing municipality added a zero to the production figure of an enterprise to boost its output data tenfold. In 2004, the NBS found that local economic reports exceeded the national GDP total by 3.9 percent …

Of course, in the West we have the rule of law, and self-advancement is not so dependent on flattering the targets of the state. Nonetheless, it is amusing to note that British officials are currently looking at ways to reduce the quoted rate of inflation here, something which would coincidentally cut the government’s cost of borrowing: it was the cost of servicing Argentina’s inflation-linked bonds that so irked the late President Kirchner half a decade ago.

But we must not be too cynical. Mr Welch – if not cracked, exactly – is very likely to be wrong.

US unemployment has stayed high for an unusually long time since the end of the recession. In the early 1980s it took less than two years for the peak in unemployment to fall back down to its pre-recessionary level. In the 1990s it took a little longer, as it did following the collapse of the stock market ten years later. Today we are exactly three years from the 10% high of October 2009 – and yet the rate has only fallen back to 7.8%. That’s still a long way from the 4.6% average seen in 2007. If anything, we should be surprised that the rate of job creation has taken so long to increase.

There has also been some corroborating evidence to support the release of the unemployment data since last week. Only yesterday, the number of weekly initial jobless claims was reported at 339,000 – the lowest level since January 2008. (And this is an actual number, unlike the labour market data which is based on business and household surveys.)

It is right to be sceptical of data, and folly to impart too much significance to an individual release. And it is an indisputable matter of record that politicians with the necessary degree of motivation and cojones (as they say in Argentina) can corrupt economic numbers if they so choose. Jack Welch’s incredulity is understandable, up to a point. But the best advice is surely to follow patterns of behaviour over time; to build up a big picture. Once we have that in mind, it becomes easier to spot what might genuinely be out of place.


Entry filed under: Posts. Tags: , , , , , , , , .

Tail Risk Drawing A Line

Recent Posts