The probability of a hard Brexit is very much higher now than it was immediately after the referendum, despite a rumoured eleventh hour softening of May’s tone ahead of today’s invocation of Article 50. 
As we said just after the referendum, a soft Brexit by way of continued membership of the European Economic Area or a special bilateral deal is only likely if either the UK is prepared to cede some control of immigration or if the EU were to lessen the prerequisite nature of the 'four freedoms'. Statements from both sides have now all but formally ruled this out.
This leaves us with two main potential outcomes (both potentially hard): a free trade agreement (FTA) or a reversion to World Trade Organisation (WTO) rules. We ascribe a 50-50 probability.
The relative economic impact of the FTA or WTO outcomes depends on the extent to which an FTA would cover non-tariff costs to trade. Non-tariff costs to trade relate to regulatory compliance and proofs, other customs-related paperwork, competition policy, e-commerce policies, labour standards, intellectual property rights, government procurement policies and so on.
The biggest gains to membership of the EU have stemmed from the reduction of non-tariff costs to trade, although many still remain. We do not believe tariffs, which only apply to goods and not services, are particularly significant under WTO rules. The trade-weighted average tariff that would be applied to UK goods exports to the EU is just 3.3 per cent. This has already been paid three times over by the fall in the value of the pound, and, regardless, most British exports do not operate business models that compete aggressively on price.
A reversion to WTO rules (which barely legislate for non-tariff costs) or an FTA light on the reduction of non-tariff costs would exert a significant drag on the productivity of the UK's tradable sectors, deterring foreign direct investment and lowering the equilibrium exchange rate implied by the UK's long-term economic fundamentals.
On day one of Brexit, the regulatory regimes would presumably be very similar, so the costs of proving regulatory compliance may not be too large. But as the regimes diverge - which they presumably would, given the fervour around 'sovereignty' during the Brexit campaign - the costs would increase over time.
An FTA that does reduce non-tariff costs would significantly improve the UK's prospects outside of the EU, particularly if the UK can also go on to agree similar FTAs with other faster growing economies. However, given that much of the benefit of being a member of the EU stems from these non-tariff costs to trade, it does not seem to be in the EU's interest to offer up much in any resultant EU-UK FTA.
However, we cannot miss the fact that the ultimate cost or benefit of Brexit depends in part on what happens to the EU after the UK leaves too.
If the EU completes the single market without the UK, bringing down the remaining non-tariff costs to trade, then it is much more likely that investment will be reallocated from the UK towards the mainland than if further progress in the EU project stalls.
Even full-time academics in the political sciences disagree as to the likely future shape of the EU. We can offer a couple of observations. Recently started initiatives, such as the Capital Markets Union (which would open up access to non-bank financing across borders for EU firms) indicate that the EU is still travelling in the right direction. The experience of the telecommunications and air transport markets following their recent opening up should also help to convince countries that the gains in terms of price reduction, increase in R&D spending and increase in market size far outweighed the losses incurred by incumbent firms.
Thinking about the exchange rate, our proprietary analysis of the long-term value of currencies relative to their comparative economic fundamentals suggests that the pound is extremely undervalued versus the US dollar, euro and Japanese yen (less so versus emerging market currencies).
Although the long-run economic value of the pound would shift lower in a hard Brexit scenario, primarily due to the impact on productivity, the actual exchange rate is so far below the economic equilibrium value that we expect the pound to rise (on a long-term basis) in any scenario. It is really just a questions of speed. That said, such long-term analysis does not help us forecast currencies on a 6-12 month view, and the newspaper headlines generated by the negotiations could well drive exchange rate volatility. That said, the short-term technical positioning data suggest that sterling 'shorts' are near all-time highs.
Most of the negotiations will occur between September 2017 and September 2018. This is because they are unlikely to get going until after the German election, and the European parliament is likely to want six months before the official two year negotiating period is up in order to get any resultant deal ratified by each of the national parliaments of the EU member states.
That said, the scale of the logistical planning required to make Brexit a reality – on both sides, but particularly for the UK – means that some form of interim deal is likely, extending the uncertainty into 2019 and maybe even beyond.
Ed Smith is asset allocation strategist at Rathbones. The views expressed above are his own and should not be taken as investment advice.