The effect of the Brexit referendum on the financial integration of the EU and the UK

Abstract

Britain’s decision to leave the European Union (a process known as Brexit) had and will have a wide effect on many spheres. The effects were particularly visible on the financial markets in the short run. The process could lead to political, legal and economic disintegration, a process in which the United Kingdom would move further from the European Union. We decided to analyze the change in financial integration that the referendum could have triggered. There are a large number of definitions, along with many measures for the concept of financial integration. We decided to follow a portfolio manager’s point of view and to define the integration as the process leading to higher correlation between returns of companies from different countries. Our measures are therefore based on the correlations between the returns generated in the United Kingdom and those generated by European Union firms. We use three measures: a comparison between pre-referendum and post-referendum correlations; a 6-month rolling-window correlation; a DCC-GARCH model. We show theoretically that the DCC-GARCH model has some advantages over the rolling-window correlation. We carry our analysis at the country and sector level. We also analyze the change in correlation of UK-centred and foreign-centred firms. We find that the referendum led to a decreased of the correlations between all indices, at the country and sector level, often only temporarily. We found that most correlations regained their pre-referendum level at the end of the sample. The effects varied depending on the sector or on the geographic orientation of the firm. Our robustness checks (mean equation, order of the GARCH, dataset) confirm our results. In a further step, we decided to estimate the relationship between conditional correlation and conditional volatility (the square root of the conditional variance) by running a regression. We show that for most sectors, there is a positive relationship between the conditional correlation and the conditional volatility. This is an undesirable feature for diversification purposes as it implies that the correlation between indices is higher when the volatility in one of the two economies is higher.