Facts are stubborn, but statistics are more pliable

31st January 2012

Consumer confidence figures from market researcher GfK NOP issued today show a four point improvement over the past month. It now stands at its highest level since last June as consumers became more cheerful over the economy and their own money.

So is this polling statistic – complete with accepted margins of error – good or bad news? 

The Guardian goes for gloom. Heading the story Consumer Confidence "Seriously Depressed", it concentrates on the -29 figure.  It says "the rise came as a surprise given the worrying state of the economy after a 0.2% contraction in GDP in the final quarter of 2011 fuelled fears of another recession."

But the Daily Telegraph takes the opposite line. From the same underlying figures, it stresses the rebound to the highest score since early last summer as proof that consumers are becoming more cheerful.

Same numbers, vastly different interpretations. It's a glass half full or half empty question – and like so many with investment implications, it all depends on what those writing the narrative want to put over.

Goebbels often said the big lie, repeated very often, was the key to people's minds. While the Nazi propaganda minister was evil beyond compare, his idea of wearing down opposition by coming up with the same story over and over again, finds some resonance with today's economic and political spin doctors. But as well as lies, there are also damned lies, and then, in the final circle of hell, statistics.

With figures, investors have to look not so much for the outright lies but for the partial truths – and what is left out. It is not a question of 99% of all statistics are made up, but a very large proportion (it is impossible to find an accurate statistic) are selected or interpreted to suit a particular agenda. Mine down the consumer confidence figures for proof.

Official statistics – primarily from the government – are intended to be trusted. They should be more robust than an opinion poll.

But as the monthly inflation figures show, there are your rising prices and the Bank of England's rising prices.

Inflation figures will seem a Bank of England success

The figures due in February will look good – or at least a lot better than those released in January.  And that is because the January 2011 VAT rise will fall out of the calculation. Consumers may still feel the pain of a 20% VAT rate and its effect on consumption, but as it is now unchanged from last year, it is no longer inflationary.

That's not the only problems in the general distrust of the figures. The Office of National Statistics works on what it considers to be a typical buying pattern. But no single person or household is exactly typical. The buying pattern for retired people is different from those in their twenties – the former apparently spend on National Trust memberships while the latter are reputed to go clubbing.

We tend to notice items that rise in price – perhaps bread – but ignore those that fall, milk recently.  And because the official figures contain big ticket items such as laptops which, on a same specification basis are falling in price, we pay little attention because we buy these comparatively rarely.  Many think items such as TV sets are rising in price (they are not) because they upgrade from a 20 inch screen to a 37 inch machine with built in internet and 3D.

Those with unusual purchasing patterns – some ethnic groups that only eat certain foods – may find they are not catered for.

Changing the parameters

But the biggest assault on inflation figures – and other official statistics  – comes when the government alters the basis.  Inflation used to be measured by the Retail Prices Index which includes housing costs. The now favoured Consumer Prices Index (CPI)  tends to show a lower rate of increase because it ignores housing costs so it is now preferred for annual pension updates.  There is even a move towards CPIY – a measure which ignores the effect of indirect taxes such as VAT and duty on alcohol.

Unemployment statistics are regularly criticised because it is all too easy to move the some of the out of work from one benefit category to another – from jobseekers allowance to a disability benefit, for instance.

Official statistics on crime and other social issues may not be reliable due to the "statistical iceberg".   Frequently, figures on sensitive issues may remain hidden because people either do not come forward to report matters such as alcohol problems.  And many give bland or expected answers to researchers – the financial services survey which finds that a huge proportion of respondents support teaching teenagers about pension matters even if they do not really think it possible. No one knocks motherhood or apple pie.

US criticism of government statistics

The financial world waits each month for inflation and jobless figures from the world's largest economy – they move bond, currency and equity markets.

But Shadow Statistics regularly throws cold water on them. It believes they are selectively released to suit the Washington agenda.  The site "exposes and analyzes flaws in current U.S. government economic data and reporting, as well as in certain private-sector numbers, and provides an assessment of underlying economic and financial conditions, net of financial-market and political hype."

The site is run by John Williams, who describes himself as an independent consultant. He can be apocalyptic – he believes that the US government is deliberately creating hyper- inflation by devaluing the dollar.

Williams has his detractors as this site and many others show.  But what they all prove is that you can create more than one narrative from the same set of figures. 

Econobrowser deconstructs many of his claims – but those posting on the site are sceptical of both arguments.

What about the UK?

Anyone from industry bodies to popular newspapers uses figures to back up their arguments. And if official figures don't support the thesis, throw doubts on them. The big guns of the building industry did not agree with ONS figures last spring. It cited special reasons such as the weather in the same way that the Bank of England tends to cite one-offs in inflation figures – yes, they won't be repeated but some other unforeseen circumstance could easily take their place.

More progressive forces in the UK also use statistics. The Radical Statistics website uses numbers for a left of centre programme, challenging the par
tial use of figures by others. It believes statistics should inform, not drive, political and social change.

At least, they probably won't have to spend too much time on this Richard Littlejohn offering. He admits that his conclusion that Britain is still overwhelmingly white, Christian and robustly heterosexual comes from the "National Office of Sex, Lies and Statistics."


More from Mindful Money:

Lies, damned lies and investment statistics

What announcement would prompt a severe market shock?

The problem with predictions

Sign up for our free email newsletter here, for your chance to win an Amazon Kindle 3G Wifi. 

Leave a Reply

Your email address will not be published. Required fields are marked *