On 10th August 2018, the popular political website ‘Guido Fawkes’ published a graphical article entitled ‘Brexit Britain Beating Eurozone Growth’.
The graph is misleading in more ways than one.
Truncated y-axis: Britain’s stated quarterly growth rate is 33% larger, but represented by a bar five times the size of the Eurozone.
The data is provisional: Unexplained by the article, the Eurozone quarterly GDP growth rate was a ‘flash estimate’, which has since been revised upwards.
Selective beatings: Even if British growth exceeded that of the Euro area countries in Q2 2018, that would only be second time in the last 10 quarters.
Edward Tufte’s first principle of graphical integrity — given in The Visual Display of Quantitative Information — is:
The representation of numbers, as physically measured on the surface of the graphic itself, should be directly proportional to the numerical quantities measured.
Bar charts should be proportionate. A quarterly growth rate of 0.4 points is 33% larger than 0.3 points, but is represented by a bar five times the size.
Neither the chart nor the accompanying article indicate that the data is provisional and preliminary.
The provisional nature of the 0.3% quarterly growth rate in real GDP for the Euro area countries is clear in Eurostat’s reports:
The preliminary flash estimate of GDP growth in the second quarter of 2018 will be released on 31 July 2018 and will be updated on 14 August 2018. The next regular estimates of European aggregates are due on 7 September 2018 (GDP estimate) and 12 September 2018 (employment estimate).
As it happens, the ‘flash estimate’ for the Euro area countries was revised upwards, to 0.4%.
The quarterly GDP growth estimates in Q2 2018 is now 0.4% for both Britain and the Euro area.
Even if British quarterly GDP growth exceeded that of the Eurozone in Q2 2018, that would only be second time in the last 10 quarters.
In the recent past, the Eurozone has generally been growing faster than the British economy.
Sometimes, strange axes are only the start of a graph’s problems.