23 March 2016

Our view remains as laid out in our Brexit report from last year – the impact of Brexit is not predetermined but depends on choices the UK takes afterwards. The long term economic impact is likely somewhere between -2.2% of GDP or +1.6%, with a more realistic range between -0.8% and +0.6%. The key factors in determining the cost will be in our view: whether the UK strikes a free trade agreement (FTA) with the EU, the impact of leaving the customs union and imposing a border between the UK and the EU, how successful the UK is in opening up to trade with the rest of the world, the choices it takes on migration, the savings from a reduced EU budget contribution and potential deregulation. So how do these latest studies compare to our assessment and where do they differ?

The two studies I will focus on here will be the PwC report commissioned by the CBI and Oxford Economics own assessment.

There are plenty of questions one can ask about the underlying motives and views which lie behind these reports, but there has been plenty of playing the man not the ball in this debate so far. As such, I feel it more important that I focus on the content of the reports.

CBI/PwC report on Brexit

I won’t rehash all the findings of the report here but the key findings are that Brexit would lead to a decline in GDP of between -3.1% and -5.5% (-3% and -5.4% in GDP per capita terms) in the short term (up to 2020). In the longer term the decline falls to between -1.2% and -3.5% of GDP (-0.8% and -2.7% in GDP per capita terms) by 2030.

Pwc 1 itemprop=

  • Uncertainty/short term impact: a big difference to our report and many others in this space is that this report attempts to incorporate the impact of short term uncertainty, hence the much larger negative impacts up to 2020. While I think most people would agree there will be some short term uncertainty in the event of Brexit, it is an incredibly tough thing to credibly model. Ultimately, PwC take the view that Brexit would increase the cost of capital (cost of borrowing for government and firms). The estimates for the increases in cost of capital are based off of increases seen during the height of the Eurozone crisis. Clearly, it is hard to find any precedent for how this might change post Brexit, but it’s not obvious why this is seen as the most valid estimate/comparison. As such, I would largely look past the short term estimates which I do not think tell us much.
  • Decline in investment: linked to the above is the fact that there would also be a significant decline in investment off the back of the uncertainty and the broader trade effects. The investment impact is determined by the inner workings of the model so it is hard to judge if decline of between 16% and 25% up to 2020 is fair and whether it would last for between five and nine years. There is an additional question here around the nature of investment – whether a steep decline in absolute amount is likely (suggesting investment crowds-in other investment) or whether it is more likely that total investment could remain high but it would  be more costly and less efficient (suggesting investment crowds-out other investment). In any case, these are transitory effects, it seems likely to me that once we know what relationship the UK is pursuing the uncertainty will begin to subside and investment/pricing decision will be based on the bigger issues (trade/migration for example).
  • Significant impact from migration: as the chart above shows, migration is seen as a significant factor in terms of the costs of Brexit. As we highlighted in our own report, limiting migration is likely to come with economic costs in terms of GDP and may lead to skills shortages in certain sectors (though these already exist, they may be shifted as the mix of immigration changes). It is worth noting though that the report does acknowledge that while the lower migration will harm GDP as a whole the impact on GDP per capita will be more limited.
  • Gains from trade: while the report does acknowledge the potential gains from striking FTAs with other countries outside the EU, it only incorporates the gains from an FTA with the US which are seen to be very small. If, for example, you incorporated the gains from unilateral trade which we found in our model (0.7%), a significant chunk of the long term cost would be removed and actually the economic impact would be around zero.
  • Gains from regulation: the report does incorporate these issues and builds substantially from the Open Europe database on regulation. They apply our view of a politically feasible regulatory saving of around £12.6bn per annum, though they take a slightly different view of the potential savings on financial regulation (believing that a dual rule book would increase costs offsetting any gains). In GDP terms the saving comes out at lower levels, 0.3% vs 0.7% in our model. It’s not entirely clear why this is, but it may be due to the assumptions about GDP levels/growth in the baseline of the model and/or what PwC sees as lost benefits from some of these regulations.
  • Fiscal gains: broadly the analysis agrees that the UK contribution of around 0.5% of GDP per annum could be saved, though it assumes half would go on capital investment and half on paying down debt, in line with the current government priorities. Only the first half would deliver gains, but even so, the fiscal gains of 0.1% are lower than expected. Why this is, is not entirely clear, but it seems due to the second round fiscal effects in the model of reduced revenue or increased restrictions on government spending due to the broader economic impacts. Our model didn’t incorporate such dynamic fiscal effects and while this is a consideration it will also depend strongly on the choices made on government expenditure post Brexit which are hard to second guess.
  • Services access: the assumption here seems to be that in no scenario will the UK gain services access of any kind to the EU single market. While we have highlighted the trade-offs in this area ourselves and that gaining access will be a complicated political negotiation, it seems a bit overly pessimistic to assume that no access would be achieved and that barriers on this area would continue to grow over time. It should also be reflected that the opportunity cost in many services sectors is quite low given that the single market in services in the EU remains severely underdeveloped. Furthermore, given the UK will be ‘equivalent’ in terms of its financial services regulation, at least to start with, the possibility of getting some access via this path should also be considered.
  • Most likely impact, not actually that bad: ultimately, the trade effects found by this study are actually quite low, -0.5% of GDP in the FTA scenario. More broadly, the study essentially concludes in a scenario where the UK secures and FTA with the EU (and not on unrealistic terms), keeps fairly open migration and attempts some deregulation (but not including potential trade gains) the impact in the long term would be -0.8% of GDP. When all is said and done, that is not a huge cost.
  • Undermines previous CBI figures: as the chart below shows, the impact on households in the longer term is quite different from previous CBI suggestions. As we have noted before, the CBI has claimed that UK households benefit from EU membership to the tune of £3000 each. While it was never directly claimed that this would all be lost, they failed to discuss any kind of alternative, thereby implying it was under threat. In the longer term and in the likely scenario where the UK has an FTA, the impact on households is -£600. I’ve highlighted reasons above why this could eventually be even lower depending on policy choices, but clearly if the EU does deliver gains of £3,000 per household – according to this analysis, much of this can be maintained outside.

Pwc 2 itemprop=

So what can we conclude about this report? Well, it’s certainly far more serious than the CBI’s previous interventions thanks to the work of PwC. Nevertheless, the short term estimates are necessarily very speculative and unfortunately received the most attention. Ultimately, the report actually shows that the impact in a likely scenario where the UK strikes an FTA would be fairly small at -0.8% of GDP. Furthermore this could be boosted by potential gains to trade outside the EU and by less pessimistic assumptions on fiscal/regulatory savings. All told though, if I were looking at this from the Leave camp, the core findings of this report aren’t all that bad particularly once you tally with non-economic arguments for leaving. Of course, it does once again raise the fundamental question of what approach they want to take outside the EU.

Oxford Economics report on Brexit

This report does not examine the short term impacts but looks through to the longer term, finding the economic impact of Brexit would be somewhere between -0.1% of GDP to -3.9% of GDP.

  • Huge number of scenarios: My first thought on the study, which identifies 9 different scenarios for life outside, is that it slightly overcomplicates things. Part of the role of these sorts of reports is to try to narrow down the potential scenarios and paint those which look most realistic. Looking at the scenarios I would think the more realistic range is somewhere between the -0.6% GDP and -3.1% (or in GDP per capita terms -0.6% to -2%). For example, the most optimistic scenario involves the UK staying inside the customs union – this seems very unlikely as it would not even allow the UK to regain control of its own trade policy and (in economic terms) would essentially amount to staying in the EU. Equally, the very pessimistic scenario of falling back on WTO rules and then taking a very protectionist approach also seems rather unlikely.
  • Depends on choices outside: The above notwithstanding, the report is comprehensive and further teases out a point we made in our paper – there will be a big difference depending on what the UK chooses to do outside. The report convincingly shows that a free trade, open and liberal approach outside the EU is far better than a more protectionist approach, although it doesn’t take this to its full conclusions (see next point).
  • Gains from trade: Nevertheless, as with the CBI/PwC report, it fails to fully capture the potential gains to trade. To be fair, it takes a fairly neutral view in that the UK will match the EU’s approach to FTAs outside (more pessimistic approaches could easily have been applied). However, given the multitude of scenarios modelled, it seems strange that one where the UK strikes a number of FTAs with other countries was not included. As such, once potential trade gains are factored in, the range for the long term impact begins to look fairly neutral.
  • Migration matters, but trade matters more: Again, as with the CBI/PwC  report, migration plays an important role in determining the outcomes in different scenarios. But there are a couple of points to note here. First, the trade impacts still dominate the migration effect as do the capital and productivity impacts in terms of the supply side effects. Secondly, the report itself caveats the effects by saying, “It is important to note that whilst the migration assumptions are important in driving differences in the overall impact on long-run GDP they are much less significant as a source of difference on long-run real GDP per head between the scenarios.”
  • Fiscal impact: Oxford Economics are particularly pessimistic on the fiscal impact believing that the second round fiscal impacts (from decreased migration and broader economic impact) would have quite negative effects and would significantly outweigh the gains from the UK’s EU budget contribution. As noted above, this is very tough to estimate and will depend on a number of policy choices post Brexit. Nevertheless the consideration of the longer term fiscal impact has been absent from the debate so far.
  • Again, outcomes aren’t too bad: Overall though, as with the CBI/PwC, the outcomes aren’t all that negative as they first appear. The range finds that in the best case there would be a rise of £40 income per person and in the worst case a decline of £1,000 per person. But in what is probably the most likely scenario, where the UK strikes some form of FTA with the EU, the impact in GDP terms would be around -0.8% (or -0.6% in GDP per capita terms).

What to make of these reports and their conclusions?

Clearly, economic modelling (including our own) should always be viewed cautiously and the outcomes should be judged on the credibility of the assumptions and inputs –I have tried to assess these above. What we can see is that there are some assumptions which could be questioned especially around the short term effects and some of the long term gains.

Nevertheless, the fundamental conclusion I draw from these reports is that in all likelihood the impact of Brexit in a scenario where the UK strikes an FTA with the EU will be around than -1% of GDP in the longer term. This impact might be even lower in per capita terms and does not include some of the potential upside – though realising said upside is not a given and would not come easily. I think the reports also highlight that taking a more open and liberal approach helps reduce some of the potential negative shock. Much of this fits with our own analysis. There will also undoubtedly be some short term uncertainty post-Brexit and people will try to factor this in, but these analyses confirm my view that it is incredibly difficult to estimate. Finally, they also drive home that we still don’t know exactly what life might look like outside, or even what the broad policy preferences are – this means it’s hard to know when any potential benefits might accrue and how significant they might be since the trading relationship we have with the EU will matter for the economic impact.

A couple of questions arise from all of this. Firstly, why was the coverage of these reports so heavily on the negative side? Well, I think the reports themselves include a huge amount of info which makes it hard to get to the crux of the findings, but their summaries also focus on some of the more negative parts. Secondly, is this really that bad for the Leave side? I’m a bit surprised by the Leave response as in the end the results show a pretty minimal economic impact from Brexit in some of the more likely core scenarios.