Fact Check: do the Treasury’s Brexit numbers add up?

What the Treasury says. Matt Cardy

If we take as a central assumption that the UK would seek a negotiated bilateral agreement, like Canada has, the costs to Britain are clear. Based on the Treasury’s estimates, our GDP would be 6.2% lower, families would be £4,300 worse off and our tax receipts would face an annual £36 billion black hole. This is more than a third of the NHS budget and equivalent to 8p on the basic rate of income tax.

EU referendum: HM Treasury analysis

The UK Treasury has published a much anticipated analysis of the long-term impact of EU membership on the economy. My guiding questions were: how reliable are these estimates? How much can one trust the headline figure that households will be £4,300 worse off each year if we choose Brexit? Is the methodology appropriate, are the assumptions fair, and are the results credible? And above all, if the answer to all of these is positive, why are the Treasury’s figures so much bigger than the current consensus, which puts the decline at between 1% and 3% of GDP?

Overall, the Treasury report contains serious analytical work. It is comprehensive, clear and rigorous. The first and second sections lay down the facts about UK membership in the EU and the alternatives. The three alternatives it presents are joining the European Economic Area (like Norway); reaching a bilateral agreement with the EU (like Switzerland); and relying purely on the trading arrangements within the World Trade Organisation (WTO) (like Russia or Brazil). There are reasoned discussions on regulation, migration and future changes in the EU with and without Brexit; though ultimately the report argues that the economic effects of these are small in comparison to those from trade and foreign investment.

The three main scenarios are estimated using appropriate and sophisticated econometric techniques (in section three and the annexes). The report does a very careful job at comparing and relating each of its main findings to the existing empirical evidence. The message the alternative scenarios deliver is clear: in all three cases, Brexit will entail significant and substantial economic losses.

The smallest losses are for the EEA model and amount to a reduction in annual GDP by 2030 of 3.8%. The intermediary case corresponds to the “Swiss” model and amounts to a reduction of 6.2%, while the WTO option is estimated at a 7.5% drop. These figures are also presented per person and per household, as well as in losses of tax receipts. With the Swiss route, for example, the estimated range is between -4.6% and -7.8%. This corresponds to a loss of about £1,800 per person – about £4,300 per household. This is the also the source of the headline figure given by the Treasury and equates to a loss of UK tax receipts of £36 billion.

Why the more negative figures?

This report may not be the last word we hear on the matter, but it will surely set the bar higher for the economic argument. Yet why are these estimates so much higher than the majority of the previous estimates? I think there are at least two reasons.

In or out? Aaz Zys

The first is that when the Treasury considers how the change in trade and foreign investment will affect the UK’s productivity, it combines the annual effect over a number of years (the so-called “dynamic” or “rate” effects). The rest of the literature has tended to estimate a one-off step change in productivity (“static” or “level” effect) rather than this cumulative impact. The Treasury’s broader, more encompassing (and more realistic if you are studying the long-term) approach makes its estimates of the losses from Brexit to be bigger than the rest.

But there may be another reason that makes all those other estimates so much smaller. It has to do with regulation. Regulation may be burdensome and “imposed by Brussels”, yet providing serious and reliable estimates of its costs and benefits is no easy task. The larger your estimate, the better Brexit looks. The Treasury numbers include them but treat them as smaller than the other estimates do. One major problem is that the econometric evidence for very large regulation costs is non-existent.


In short, the Treasury estimates are credible. Yet they are conservative: there are at least two ways in which the estimated losses from Brexit could increase. The report could have factored in the EU reforms being pushed by the UK government into the main numbers. It could also have used a definition of productivity that didn’t flatter the UK (like GDP per hour worked rather than GDP per capita). These are both included in section three, part three of the report, if anyone wanted to look at them. But I think the decision was correct to downplay these, given the uncertainty.


*Jonathan Perraton, Senior Lecturer in Economics, University of Sheffield *

This piece notes that a key reason the Treasury study estimates a larger impact is that it models the impact of Brexit over the medium to long run. If the effect of Brexit leads to lower productivity through lower trade and foreign direct investment, this will lead to a cumulative relative loss of income. As the author notes, all such estimates are subject to major uncertainties, but the Treasury results are based on solid econometric analysis from standard approaches.

Another possible reason for the Treasury results is that they model both a larger impact of Brexit on trade and foreign investment and a larger negative from this impact on UK productivity growth. Overall, though, the author rightly notes that the Treasury has provided a systematic assessment of the possible effects.