[Note to readers: This project is by now a little bit dated as the Brexit vote has been four years ago, but I leave it on here as a nice example of using electricity as a proxy for economic activity. It has also provided pointers during the current COVID crisis].
We are developing an empirical model that allows us to assess the economic impact of Brexit nearly in real-time. For this we use the electricity demand in England and Wales as a proxy of economic activity. We control for a large variety of external influencing factors like varies weather indicators, calendar and time variables, as well as public holidays and other events (e.g. Olympics). The resulting residuals allow us to make inferences on the economic, and here in particular manufacturing activity, in the country.
To do that we created a model using pre-Brexit data and used to predict energy demand. This way we obtained the following errors that show how there is a change around the Brexit vote.
To get impression of the economic uncertainty generated by the Brexit vote, we show in the Graph below the monthly variance of these errors. We observe an explosion of the error after Brexit.