A CGE analysis of Boris Johnson’s Brexit including Melitz, multinationals and unemployment effects

The role of multinationals in services sectors in the British economy has received little attention in the context of Brexit, even though foreign affiliates sales constitute the most important way of provision of services. We simulate Boris Johnson’s proposal for Brexit to illustrate the effects of barriers to the operations of foreign multinationals in services sectors (both EU multinationals in the UK and UK multinationals operating in the EU). These multinationals operate in a climate of monopolistic competition à la Krugman (1980), which is important to grasp scale economies and variety effects. We also include the impact of barriers to trade across goods and services. The majority of manufacturing sectors include a Melitz (2003) structure, which allows us to grasp the effects along the extensive margin of trade, together with productivity impacts. In addition, we include a specific characterization of unemployment, using a wage curve à la Blanchflower and Oswald (1994a; 1994b, 2005), with which we can estimate the effects of Brexit for employment, for which very few estimates exist. Our model uses the latest GTAP10 database, although we have conducted a forward calibration following Böhringer et al. (2009) to the year 2021, in which the Brexit shock would be implemented. While we offer several macroeconomic results for the UK, the Rest of the European Union, the US and the Rest of the World, we focus on the macro and microeconomic results for the Spanish economy. Beyond Boris Johnson’s proposal, we also analyze other UK alternatives for Brexit to try to illustrate the effects of different policy options. They include not only a soft and hard Brexit but also a Customs Union arrangement, together with an ambitious or modest agreement between the UK and the US, and between the entire EU (including the UK) and the US. Finally, our analysis covers the short run impact together with a steady state simulation following the approach of Francois et al. (2013).