Timely article on why GDP data is not an accurate guide

Scepticism is essential when considering a lot of economic discussion; this article on GDP reveals we should not believe slavishly it tells us everything we need to know. If anyone finds themselves interested in the points under discussion, I have a book looking at the history of GDP (it is really quite small), which covers some of the stranger anomalies, and will make you wonder why this one data set can have such a big impact on so many things.

 The Times

Why our economy is one big guessing game

GDP and other so-called definitive statistics are really no more than large-scale surveys

Ed Conway, 9th September 2016

Do you know what GDP is? No, I don’t mean what it stands for: gross domestic product. Nor do I mean what it’s measuring: total economic output/income/spending in the country.

No. What I’m asking is: When you scrape away everything else, where does GDP really come from?

The question feels pertinent because we’re at one of those junctures where there is more focus than usual on economic statistics. Take the latest run of data over the past week or so — the purchasing managers’ indices of the main sectors of the economy and GfK’s consumer confidence barometer. On the one hand, they provide us with tantalising hints that the long-feared post-Brexit slump may not happen. On the other, economists and policymakers warn us to be wary of so-called “survey data”. Far better, they say, to wait until the definitive, official GDP numbers arrive.

Which brings us back to that question: What is GDP?

I only discovered the answer a few years ago when I travelled to the Office for National Statistics’ headquarters in Newport to make a film about how the big number is put together. Half expecting to uncover a giant supercomputer in the basement churning through electronic receipts, instead I found myself inside a mailroom where eager administrators opened envelopes and fed handwritten forms into a scanner.

For even GDP — the greatest, most definitive statistic of all — is really just a survey too. Forms are posted out to a selection of companies where they are filled in and sent or faxed back (“Informed estimates are sufficient for our needs”, says the bumf). Those figures go into a model and soon enough they become GDP. The size of Britain’s economy, the question of whether we are in or out of recession, the fate of governments — ultimately it all hangs on a questionnaire.

It remains one of the great open secrets of economics. The vast majority of what we call official statistics is really built on surveys. Unemployment? Retail sales? Inflation? All survey data. In the case of inflation, a team of 300 people are dispatched around the country with electronic clipboards to collect prices from corner shops and supermarkets. For unemployment, researchers must chase subjects around the country asking them whether they have a job or not.

Of course, the scale of these surveys is far greater than most private companies could manage. The sample base for the purchasing managers’ index survey of manufacturers is 600 firms; the ONS equivalent covers 6,000. Moreover, as time goes on and more data comes in from other parts of government, those initial surveys are bolstered by “real” data from company accounts.

All the same, it is pretty odd. Here we are, decades into the computer age, with all the data we need to give us a comprehensive picture of the state of the economy sitting in Whitehall — VAT returns and HMRC data on income tax. Every day companies collect more and more data about sales and performance which, combined, could provide us with an instant proxy for economic growth. And yet our most important statistics are collected in pretty much the same way they were in 1945.

The smartest thing we could all do is to treat our economic statistics with a little more scepticism

Britain is a laggard on this front. A host of countries, from Canada and New Zealand to most of Scandinavia, have been using such tax data to measure economic growth for decades. Ninety-six per cent of Finland’s statistics come from these so-called administrative sources. The upshot is that businesses spend less time filling in unnecessary questionnaires, the statistical offices save millions in costs and the figures themselves, which are based on the entire economy’s activity rather than a sample of companies, are more reliable.

But here in the UK, thanks in part to a cumbersome legal framework for data protection, we are only now starting to move in the right direction. Come next year the ONS, which still sends out one-and-a-half million survey forms a year, will start incorporating VAT data into its GDP numbers. In future, it might be able to provide on-the-day estimates of economic growth or contraction. Data scraping services can already collect enough prices from across the internet to give a pretty reliable estimate of inflation — and one that’s updated daily, not monthly.

Then again, when it comes to economic data, more is not always better. The big challenge will no longer be how to account for the gaps in the surveys but how to digest billions of datapoints while telling a story that is still accurate.

Even a measure of every pound that Britons spend can only tell you so much. When it was invented in the 1930s, GDP seemed a pretty good yardstick for economic performance. But in today’s economy, dominated by services and, increasingly, the shared economy, it seems less appropriate. How to account for the fact that you can rent your home out as an Airbnb hotel? Does a new, updated iPhone boost productivity or will all that money spent on new wireless headphones represent a waste of economic output?

The smartest thing we could all do is to treat our economic statistics with a little more scepticism. In the words of Charlie Bean, former deputy governor of the Bank of England who wrote a report on the subject, they are more impressionist paintings than old masters, let alone photographs. Something to chew over as we await those “definitive” figures on the Brexit effect.

Ed Conway is economics editor of Sky News