Factbombing

The old joke has it that you can prove anything with statistics but the evidence from the political debate over recent years is that the opposite is true. There is indeed such a thing as too much information. Full Fact, the charity that tries to assess the claims and counterclaims made in the course of the UK public debate, has called the weapons grade use of statistics ‘Factbombing’ which is a highly evocative and entirely accurate description of the phenomenon.

I first became aware of the extent to which politicians throw out huge quantities of statistics during George Osborne’s budget speeches without being able to put my finger on what was happening. But it has been taken to new levels in the course of day-to-day parliamentary business, particularly in the wake of the Brexit referendum. Full Fact quote a case from prime ministers’ questions in January when Jeremy Corbyn made 13 separate factual claims in the course of one question, which then did not relate to the question he subsequently asked. Of course, this makes it impossible to give a sensible answer, which is precisely the intention: Much of the pointless trading of facts is designed to gain political advantage and not to enhance the quality of the debate. The flip side is that data overload makes it very difficult for the average voter to follow the twists and turns of the debate.

As for accuracy, fact bombers do tend to quote accurate statistics, although they are often cherry-picked and therefore quoted out of context which reduces their usefulness when the subject matter is put under further scrutiny. But by the time we get to that point, the debate has moved on and nobody cares whether “a Labour government properly funded the police force” or that the Conservatives are responsible for “rising crime, less safe streets.” Perhaps an even bigger problem is that politicians are generally not trusted, so it is questionable whether the electorate believes or cares what “facts” politicians quote in the course of debate which further undermines the case for evidence-based policy.

Of course, one of the reasons why politicians are not trusted is that when they do lie, they lie big. Think about the claim made in 2016 that the UK could save £350 million per week by leaving the EU. This is based on the gross contribution to the EU budget before rebates and the amount returned in the form of EU funding. It is true that over the period 2013 to 2016 the UK’s gross contribution does correspond to this figure, but the net figure is 45% lower (£190 million per week). To the extent that UK fiscal finances never work off the gross figure but off the net number, which means that such a large figure never enters the budget calculations, it is simply untrue to quote the £350 million figure, as the UK Statistics Authority Chairman pointed out to Boris Johnson in 2017. Not that British politicians have a monopoly on untruth: One of the masters of the dark arts is Donald Trump (see here for a rebuttal of some his more egregious comments) and let’s not get started on Watergate.

One of the great problems in the use and misuse of economic statistics is the lack of context. I have made the point previously (here)  that spending “millions” on government outlays is to miss the point that overall revenues are measured in billions and the size of the economy is in the trillions. For example, an outlay of £100 million is worth 4.7% of annual GDP today compared with 9.7% 20 years ago. “Record spending” on the NHS is misleading if it has not kept pace with the overall size of the economy, particularly when the population is growing older and demands on healthcare resources are rising.

It is not just politicians who are guilty of overegging it. The Department for International Trade made a similar howler recently by sending out a tweet pointing out that “exports in the 12 months to April 2019 grew by 4.0% to a record high of £645.8bn.” Exports at a record high in money terms, yes. But as a share of GDP they are still below the high recorded in Q1 2012. Indeed, it is pretty likely that the latest release of economic time series such as exports or GDP will always post record highs so long as they continue to show positive growth. In statistical terms, such series are non-stationary, meaning that their mean and variance are not constant over time. I would be much more impressed if growth (which is statistically stationary) reaches a record high. But in a world where we need to reach for the superlative, it is much easier to go with the lazy headline regarding all-time high levels of exports.

Perhaps the real problem with Factbombing is that it diminishes the quality of debate. We need to know which facts we can trust and which we can discount; which are germane to the debate at hand and which are irrelevant? In a world where we are bombarded with information on a daily basis, it is more important than ever to remember that the singular of data is not anecdote.