The Department for International Trade has been rightly criticised, including by me and colleagues at Cebr, Doug McWilliams and Cristian Niculescu-Marcu (see this report) for its economic modelling of some of the key trade agreements it is negotiating, especially the US-UK, UK-Australia (formally launched yesterday), UK-New Zealand and UK-Japan FTAs. We argued in that article that current models were far too static in nature, and did not properly weight non-geographic distance factors such as common law and common language. We also argued that the models did not take into account the potential gains from a reduction of market distortions inside the border, which could be significantly higher than a reduction of mere border barriers (see in particular the Cebr report on market distortions behind the border here and the IEA Plan A Plus here). Current modelling does not take into account the impact of pooled markets where, for example the UK is able to secure a financial services agreement with the US leading to a transatlantic financial services area or a similar common defence area. This is ironic, because treasury models appear to value pooled markets tremendously highly in the EU context, but ignore them completely in any other context.
We are therefore pleased and supportive of the DIT Modelling Review Academic Panel which is designed to inject more rigour and more realism into the modelling used. But there are some important principles that we must bear in mind with respect to economic modelling of trade agreements and trade policy.
Treasury departments constantly underestimate the benefits of trade agreements.
First, we must acknowledge that these are complex systems, akin to biological systems that are notoriously difficult to model. That is why the International Trade Commission (when the US was still a member) looking at the then Transpacific Partnership concluded that while tariff reductions could be relatively easily understood, the benefits of a reduction of behind the border barriers was much more difficult to model.
Secondly, we must also acknowledge that Treasury departments constantly underestimate the benefits of trade agreements. The NZ Treasury underestimated the impact of the China-New Zealand FTA by 500%, and the agreement reached the gains projected for 11 years in a mere 11 months.
Thirdly, this is more art and less hard science. Any practitioner of the physical sciences knows that economic modelling is by comparison educated and guided speculation. It can certainly be useful speculation, but I would not bet my life on it, however good the modelling is. It is therefore important to use multiple models to try and capture specific benefits.
Fourth, no group of trade negotiators in the world goes around with a laptop modelling potential agreements in order to decide where to negotiate. Trade negotiations support foreign policy objectives, they support geo-political realities, and often they benefit from opportunities that present themselves with narrow windows in which to act (such as the US-UK FTA for example).
Fifth, it is almost impossible to say anything useful about the economic impact of a trade agreement before it has actually been agreed and we know what it actually does. A US-UK FTA that dealt only with industrial goods tariffs would have limited benefit, but a much deeper comprehensive agreement that tackled the anti-competitive restrictions in both markets would have a much bigger impact and we cannot know which one we are modelling until we see the words on a page.
We give two cheers for the Modelling Review because it is very important that it is recognised that the modelling produced so far has been deeply flawed. However, it will remain flawed if the considerations we have set out here and in the documents referred to are not properly considered. It is to be hoped that the Academic Panel, and anyone involved in modelling aspects of the UK’s trade policy in the future will bear these factors in mind.
In his Ditchley speech, Michael Gove referred to the danger of groupthink, and referred specifically to modelling exercises. He called out the need to get proper challenge from qualified outsiders. As Gove implored, we must not allow this new Panel to become the preserve of the usual suspects, but we must be prepared to experiment and hear dissenting voices. In this, as in other areas we must be prepared to innovate, and challenge the status quo and received wisdom.
We give two cheers for today’s announcement, but will hold our third for the evidence that economic modelling in government has actually taken into account the many ways in which an external trade policy agenda will contribute to economic growth for this country, and for the wider world.