The heterogeneous impact of Brexit: Early indications from the FTSE

Abstract

The UK's decision to leave the EU is surrounded by several studies simulating its potential effects. Alternatively, we examine expectations embodied in stock returns using a two-part estimation process. While most firms' prices fell, there was considerable heterogeneity in their relative changes. We show that this heterogeneity can be explained by the firm's global value chain, with heavily European firms doing relatively worse. For firms with few imported intermediates, this was partially offset by a greater Sterling depreciation. These changes were primarily in the first two days and highly persistent. Understanding these movements gives a better understanding Brexit's potential effects.1 Norway has a free trade agreement with the EU but is not a member of the EU's customs union, so it faces the non-tariff barriers that apply to non-EU countries.2 Other studies in this vein, which cover various simulation exercises, include Head and Mayer (2015), PWC (2016), Fraser of Allander (2016) (who focus on Scotland), HM Treasury (2016),and OECD (2016). All of these find negative effects of various magnitudes whereas Minford, et al. (2016) finds the potential for positive impacts on the UK. It should be noted that Sampson, et al. (2016) argue that Minford, et al.'s optimism is based on implausible assumptions on trade barrier changes and import elasticities.3 Head and Mayer (2015) describe three possible disadvantages of Brexit for FDI. First, an increase in trade barriers makes production in the UK less attractive because it becomes more costly to ship to the rest of Europe. Second, supplying inputs and staff from brands headquarters becomes more difficult (higher co-ordination costs). Third, UK products become less attractive to EU consumer after Brexit.4 Strictly speaking, this is a return that is 14.4% worse than expectations; since in the long run a firm's return should equal the market, results in our comparison. See below for a detailed discussion of how to 11 Timmer, et al. (2014b) provide a recent overview of this literature. 12 Davies, et al. (forthcoming) find that tariff pass-through within a multinational is roughly half that of an arm's length transaction.13 See London Stock Exchange (2010) for details.

The Economic Effects of Brexit: Evidence from the Stock Market

“…By contrast, increasing the share of subsidiaries in the UK or the EU by 10 percentage points reduces abnormal returns by 0.9 percentage points and 0.4 percentage points, respectively. Contrary to the results reported in Davies and Studnicka (), however, the latter effect is not statistically significant . Our recession‐proof dummy is also positive and significant as expected, indicating that stocks in industries that perform better during downturns experienced abnormal returns that were 3.6 percentage points higher.…”

“…(8) Davies and Studnicka (2018), however, the latter effect is not statistically significant. 24 Our recession-proof dummy is also positive and significant as expected, indicating that stocks in industries that perform better during downturns experienced abnormal returns that were 3.6 percentage points higher.…”

“…The results in columns 2 and 3, where we include tariff barriers for goods-producing industries and non-tariff barriers for selected service industries, respectively, suggest that such expectations did not matter as both trade barrier proxies are insignificant. We caution that sample sizes are, of course, considerably smaller in both regressions, explaining why some of the 24 The insignificance of the EU share variable seems to be driven by the inclusion of regressors not used by Davies and Studnicka (2018). After dropping the recession-proof and trade status dummies, as well as using log(assets) instead of log(sales) as in Davies and Studnicka (2018), the coefficient on the EU share regressor becomes significant at the 1 per cent level.…”

“…They do not regress abnormal returns on explanatory variables, however, and thus cannot formally test hypotheses about differential sector-level impacts advanced in the prereferendum literature. Similar to our paper, Davies and Studnicka (2018) correlate abnormal returns with a number of explanatory variables, focusing on the role of global value chains. We study a wider range of determinants, however, and link our choice of explanatory variables more closely to forecasts made before the referendum.…”

“…The insignificance of the EU share variable seems to be driven by the inclusion of regressors not used by Davies and Studnicka (). After dropping the recession‐proof and trade status dummies, as well as using log(assets) instead of log(sales) as in Davies and Studnicka (), the coefficient on the EU share regressor becomes significant at the 1 per cent level.…”