How Brexit warped the minds of otherwise sane economists

How Brexit warped the minds of otherwise sane economists

Recently, the forecasting record of the Economists for Free Trade has come under fire in the FT from Chris Giles (‘Growth and Brexit- four lessons’), with the assertion that it has been worse than that of the consensus of economists; and also no better absolutely than the Treasury before the referendum. We were ‘badly overoptimistic’, while the Treasury was ‘badly over-pessimistic’ but ‘we both had a referendum to win’. Furthermore ‘our record had got steadily worse over 2016 and 2017’, on the basis of our forecasts immediately after the referendum in July 2016.

Since the FT has not had the good grace even to publish so much as a morsel of my response to their criticism in their letters column, , I shall set it out in detail here. This, of course is consistent with the FT having refused to publish a single OpEd from us since the beginning of the referendum campaign.

Chris Giles’ line of argument is complete and indeed irresponsible nonsense, even on provisional ONS estimates, which look as if they will be revised upwards. The ONS recently has admitted to underestimating telecoms productivity growth and also the growth of exports. The latest GDP estimate for the first quarter has been bedevilled by terrible weather. We shall see how big the revisions are when the statistical fog clears in a year or three. What is undeniable and worrying is the flawed and downward-biased pessimism of the official and corporate economist consensus, which Chris Giles for some extraordinary reason is determined to defend. It really does seem that Brexit has warped the minds of many otherwise sane economists.

Take, first of all, the pre-referendum forecasts of the Treasury and ourselves for the Brexit case (during the referendum, we were called, ‘Economists for Brexit’).

It can be seen that our error in 2016 was a small overestimate of 0.2 per cent, against a Treasury underestimate of 1.1-2.0 per cent; and the accumulated error between Q2 2016 and Q4 2017 was a 1.2 per cent overestimate for us versus a huge 2.9-4.9 per cent underestimate by the Treasury, nearly three to five times our absolute error.

 

The Treasury’s forecast has of course rightly become nationally notorious whereas we were a bit on the upside, which given the usual problems with this data is within the customary margins.

Now, if we turn to the post-referendum comparison of EFT with the consensus we find:

We cannot compare what is happening in 2018 because the consensus did not forecast 2018 in July 2016. However, the first quarter of 2018 is in any case heavily affected by weather, with construction dropping sharply, probably related both to the actual weather and the weather forecast, so that work was simply pushed to later in the year. We must judge our and the consensus forecasts made in 2017 for 2018 when we have some more 2018 data.

As can be seen, the accumulated error of Liverpool/EFT by the end of 2017 was about half that of the consensus, while the error for 2016 was a fifth of it. While not on the Treasury super-scale, the consensus error is massive – basically they missed a whole year’s growth.

The OBR and the Bank made no comparable forecasts, though the FT includes them on the pretence they did. The OBR made no forecast until November by which time the economy’s post-Brexit stability was well apparent. The Bank forecast was published on August 4, when it announced a four-point monetary stimulus, including large QE and an interest rate cut- thus having a massive information advantage compared with private forecasts in July. Both organisations therefore were far more optimistic than the consensus in July 2016, with good reason. Even so, they produced forecasts for 2017 that were much too gloomy (their forecasts are appended below).

Chris Giles in the FT goes on from these forecasts to assert that ‘Brexit has reduced growth’, but forecasts have no bearing on this and this assertion is transparently false. We are currently at full employment with wages accelerating; without Brexit there could not have been ‘fuller’ employment. Brexit has in fact changed the shape of the economy via a sharp devaluation, causing a shift from consumption to net exports and traded sector profits, a necessary correction when we were running an unsustainable current account deficit of nearly 7 per cent of GDP. But Brexit cannot have changed the overall level of employment – or output, as it is too early for any effect on productivity.

In fact, what forecasters are now converging on is the limit on pre-Brexit productivity growth as measured (even after all obvious revisions) by the ONS. Even though 80 per cent of our economy is now services, the ONS is making no serious effort to measure their productivity growth accurately, which requires proper assessment of their growing quality. Yet it is plain that the quality of services has been and continues to grow, from the simplest observations of how our everyday lives are assisted by the services of mobile phones, WiFi and internet-based shopping (for more on this see here). So it is that we are faced by ‘GDP growth rates’ stubbornly stuck below 2 per cent, simply because we refuse to measure them properly. This can hardly be dubbed a Remainer triumph!

As Chris Giles at the FT well knows, the real question about the effects of Brexit is how the supply-side capability of the economy will react to it – namely how productivity will respond to free trade and other post-Brexit policies. We have estimated these effects on the basis of actual government policy as generating around 7 per cent more on GDP over the long term. Now that the Civil Service has jettisoned the Treasury’s methods in favour of the GTAP Computable General Equilibrium model from Purdue University, their estimates of the trade effects on the same assumptions ought to be rather similar to ours.

Getting these assumptions right is where economists should now be turning their attention.

Economists should focus on getting things right, not bashing Brexit
Brexit should generate around 7 per cent more GDP long term