UK government has recently attempted to place the blame for weak UK productivity growth and the problematic fiscal situation on Brexit. But there is no convincing evidence for such a link. The data shows no slowdown in aggregate productivity since 2016, and no sign of larger effects in sectors most exposed to Brexit. Claims of a link between economic ‘openness’ and productivity are highly questionable, and in any case the UK has not become less economically open since 2016. The evidence also does not support claims that Brexit has hit productivity by reducing investment.
The UK government has recently tried to place the blame for its fiscal problems on Brexit, claiming that weak UK growth, and by extension government revenues, are the result of a slowdown in productivity growth resulting from Brexit.
This claim is not new. Indeed, it dates all the way back to the earliest studies of the potential impact of Brexit on the UK economy which relied heavily on an assumed hit to productivity to generate their big negative effects on GDP. An average of a motley collection of these studies forms the basis for the well-known claim by the UK Office of Budget Responsibility (OBR) that Brexit could cut UK productivity by 4% in the long run. The driving factor behind this is a reduction in economic ‘openness’, or trade intensity, caused by higher trade barriers.
We have been highly critical of this OBR claim of a 4% productivity hit on many occasions over recent years (see here, here, here, and here). Our objections can be summarised as follows:
- Many of the studies that underlie the OBR’s claim of a 4% hit incorporate highly questionable assumptions
- The economic literature does not support a strong causal link between changes in trade and productivity growth – indeed, it is more likely that the direction of causation runs the other way, from productivity to trade
- The post-Brexit empirical evidence for the UK shows no sign of a significant slowdown in productivity relative to its recent trend
Oddly enough, the OBR used to share our doubts. In a publication from 2018, it did an admirable job outlining the problems there are with claiming a strong link from changes in trade intensity to productivity. Among other things, the OBR correctly noted that:
- The empirical evidence for the impact of openness on productivity is mostly drawn from cross-country studies where much of the information in the data derives from rising trade intensity in developing countries, with that experience perhaps not being relevant to the UK
- There are significant problems with the econometric techniques used in such studies
- The relationship might be asymmetric. Developing countries may see productivity gains from ‘learning by doing’ as trade rises but it is unlikely that developed economies would lose learning already gained by trading less
- That any productivity effects might only manifest in the very long term
All these points remain highly relevant. For the first point, we note that looking at changes in trade volume and changes in GDP per head are statistically unrelated for advanced economies for the 1980-2023 period while there is a link for developing countries (see Chart 1).
Source: OECD
But the biggest problem for the claims of a link between Brexit and UK productivity growth over recent years is that the actual productivity data show no sign of such a link.
It certainly true that UK productivity growth (here measured as output per hour worked) has been slow over recent years, but this trend long predates the Brexit referendum of 2016 or the point at which the UK exited the EU’s customs union and single market (the start of 2021).
A look at productivity from 1995-2025 shows that it grew at a trend pace of 2.2% per year until the global financial crisis and since then has grown at a much slower pace of around 0.6% per year. But there is no indication of any significant further slowdown in the trend since 2016 (see Chart 2) – productivity growth has remained close the trend pace visible since 2008.
Source: ONS
Perhaps this aggregate data masks sectoral differences, with the sectors most exposed to Brexit doing worst? Again, there is no evidence for this.
In a ‘first pass’ at the data, Fernald and Inklaar divided UK industries into more and less Brexit-exposed groups based on export shares – and found that more exposed industries had actually shown stronger productivity growth than the less-exposed group.
This pattern seems to have remained in place. Manufacturing is arguably the most exposed sector given its high export orientation but has shown a much stronger productivity performance than the economy as a whole since 2016, with productivity rising some 17%. The biggest drag on the UK’s productivity performance has instead come from the public sector where productivity has actually declined (see Chart 3). This sector has very low exposure to Brexit – its problems derive from very different factors.
Source: ONS
Perhaps the UK has done very badly in international comparison? Again, the evidence does not support this. For the period 2016-2024, total UK productivity growth certainly lagged that of the USA. But it was the second highest among the G7 countries, slightly better than for Japan and Germany and significantly better than for France and Italy. For manufacturing only, the UK’s productivity performance was the strongest in the G7 (see Chart 4).
Source: National statistics, OECD
* Output per hour except France and Italy, output per person
Taking a step back, we also find that the supposed drivers of a post-Brexit productivity slowdown in the UK are not supported by the actual data.
Looking first at ‘economic openness’ or trade intensity, there is no evidence that the UK has become significantly less economically open since 2016. The most commonly used measure of this is the share of exports and imports in GDP, in constant prices. This share was 63% in 2016 Q2, the quarter of the referendum, and 66% in 2019 Q4 – the last quarter before the UK left the customs union and single market in which the data were not badly affected by front-loading of trade related to Brexit deadlines. The ratio in 2025 Q3 was 66%, so higher than in 2016 and unchanged from the end of 2019. There is no reduction in economic ‘openness’ visible here.
An alternative measure of openness is the share of value-added exports in output. This has some advantages as traditional trade data sometimes overestimate trade intensity by double-counting intermediate goods or even counting as ‘exports’ goods which may have passed through a country but with minimal or no value-added created in that country. This is an issue for the UK around Brexit because goods like clothing were often imported into the UK pre-Brexit from third (largely non-EU) countries before being distributed on to other EU countries.
Unfortunately, the value-added export data for the UK only runs to 2022 but again it shows no sign of a decline in economic openness. The share of value-added exports in output in 2016 was 23%, in 2019 it was 24%, and after dipping slightly in 2020-2021 it rose to 26% in 2022 (see Chart 5).
A final point here relates to relatively weak recent goods exports to the EU. Some of this reflects a shrinkage of the ‘extensive margin’, i.e. the number of firms engaging in exporting. But this is likely to have very small effects on productivity. Most of the firms that have withdrawn from exporting will be smaller and lower productivity firms. Nor will the potential reallocation of resources from such firms to domestically-focused firms have much impact as a firm-level study for the ONS shows that the ‘productivity premium’ of firms exporting to the EU relative to non-trading firms was very low.
Source: ONS, OECD
Another line of argument has been that Brexit hurt UK productivity by causing a decline in investment. Whether this link makes sense is open to question – some analyses argue that capital formation trends as a consequence of productivity and output growth trends, not their cause. But in any case, yet again, the actual data do not support the notion that weak investment has hurt UK productivity growth – at least if properly interrogated.
Attempts to claim a big hit to investment after Brexit have had serious flaws. Haskel and Martin rely on outdated investment data that was subsequently revised up significantly and base their claim on an extrapolation of a very strong short-term trend in business investment prior to 2016 – implausibly assuming this would have lasted forever if not for the Brexit referendum. Springford similarly uses the outdated data and bizarrely compares UK investment to a synthetic counterfactual index in which almost none of the individual country time series for business investment are statistically significantly correlated to the UK.
Moreover, once again the actual data are at variance with these claims. If we look at UK business investment (the correct aggregate if we are thinking about impacts on productivity) we can see that investment has cycled around a fairly steady trend since the mid-1980s, and that currently business investment is modestly above the trend level. Investment was above trend in 2016 but after a brief decline during the pandemic it picked up again and is now some 11% above the level seen at the time of the Brexit referendum. Moreover, this occurred despite a sharp fall in investment in North Sea oil and gas related to the long-term decline in that industry.
If we instead look at business investment as a share of GDP, we can see that it has been on an upward trend since the 1980s and reached a high of just under 11% in 2016. It is at the same level now. So, far from Brexit depressing UK business investment, business capital spending has stayed at historically elevated levels since 2016 (Chart 6).
The evidence above shows that claims of a big hit to UK productivity due to Brexit are not consistent with either the economic literature or the actual performance of key UK productivity, trade, and business investment data since 2016. Attempts to show such an effect rely on studies with dubious assumptions, outdated or cherry-picked data or in some cases (i.e. the use of synthetic counterfactuals) data that is essentially invented. We may also note that successive vintages of the OBR’s forecasts of UK productivity have been hopelessly wrong – they have not demonstrated any expertise in this area.
Britain certainly has a productivity growth problem, although it is far from alone in that. All the advanced economies have seen a notable slowdown in productivity growth since the global financial crisis with the reasons for this still debated. In addition, it is also true that the level of productivity in the UK lags that in the US in particular. The reasons for this are varied – among other factors poor education and skills, low levels of R&D and capital stock in hi-tech industries, misallocation of capital, regulatory constraints, an unfavourable structural pattern in the economy and statistical measurement problems have all been offered up as explanations. But these are very long-standing issues and nothing to do with Brexit.
