“Sterling at lowest level since records began” – “another volatile day” – “unprecedented systemic shock”

Just a few examples of the powerful headlines blazoned across the media in the weeks and months that have followed the UK’s EU referendum result. But what can we really infer from the exchange rate movements and what does this mean for market risk? We dig down into the data behind the headlines.

Daily movements

This graph shows the distribution of daily GBP to USD exchange rate movements, since records began in 1975, using the following key:

We’ve taken a look at the exchange rate movements observed on key dates in 2016, where an increase in exchange rates corresponds to a decrease in the value of GBP compared to USD.

On 20^th February, David Cameron announced the date of the UK’s EU referendum, but markets didn’t react markedly and the largest daily movements in the month were towards the beginning and the end.

On the 5^th May, a number of UK elections and 27^th May marked the start of a four-week purdah before the EU referendum vote. Whilst the value of GBP decreased, this was not significantly out of line with market volatility previously experienced around elections.

The 23^rd June saw GBP rise as markets and forecasters predicted a Remain vote. But upon announcement of the Leave result, GBP crashed. The headlines are true – the 8.64% movement was the largest daily increase in the exchange rate since records began, almost double the longstanding record of 4.51% during November 1978. This illustrates the importance of making allowance in capital models for “events not in data”.

When Teresa May announced that Article 50 would be triggered in 2017, another increase was experienced – the markets do not expect the UK to get a good deal in EU exit negotiations.

Monthly movements

Exchange rates are clearly volatile. But whilst these daily movements lie outside the distribution’s interquartile range, monthly movements have been somewhat more modest:

Only the exchange rate movement over the month of June lies outside the distribution’s interquartile range. What this graph does show is the continual weakening of GBP. Over the coming year, more weakening is forecast and what may have once been a 1 in 200 exchange rate increase is now considered to be more likely.

Market risk capital

From an insurer’s perspective, the primary consideration in relation to FX risk is asset and liability matching. We are already in an extremely low interest rate environment, so it is unlikely that the most successful insurers are relying on investment returns. Therefore the experience that really matters will be quarterly and annual movements.

Moody’s is just one of the ESG providers that responded quickly to the referendum result; revising its forward-looking assumptions and therefore the ESG output simulations. Insurers should continue to monitor economic forecasts against internal positions and ensure that capital can withstand extreme shocks.

We are already working with clients to understand the potential implications of Brexit on capital – the Brexit result and resulting market movements have provided insurers with more severe historical scenarios with which to back-test their models.

According to Steven Denn, “You can never make the same mistake twice because the second time you make it, it’s not a mistake, it’s a choice.” This famous quote made more sense to me when I first moved into the world of data analytics. Since the major goal for most businesses is to maximise profits, analysis of historical data can be crucial in making decisions on how to reduce costs and increase the revenue thereby avoiding the concept of “making the same mistake twice”.

Data analytics allows a business to make decisions based on historical information. This can be achieved by identifying appropriate patterns, classifications and correlations within the historical data. One of the industries where data analytics is appreciated is in the insurance industry. Since the profits emerging from insurance policies are unknown until they are fully written off the books, most insurance companies use the analysis of historical data to estimate liabilities and future costs, thereby reducing future uncertainty.

In other businesses, decisions have to be made as to how to reduce costs and how the products need be sold to maximise revenue. Data analytics can play a crucial role in decision making. This implies that data analytics is not only useful in situations where there is uncertainty with regards to liabilities and expenses, but if applied efficiently it can also be used to maximise profits in any organisation.

The amount of historical data in terms of volumes and variety has an effect on the accuracy of the results. However, as data increases in size, more complex methods and tools will be required to analyse and extract useful information from the data. I will give a very simple example where one might need to analyse the historical data for three companies A, B and C with 4, 5000 and 5000000 records respectively. When dealing with 4 records only, one might think of using a simple calculator, whereas an excel spread sheet might be more useful on 5000 records, but these tools won’t be helpful when dealing with 500 0000 records.

Although the complexity of the tools used in analytics are proven to increase as the dataset size increases, I have also realised that with innovation in the IT world it is possible to simplify the analysis of big historical datasets. One example is Dynamo Analytics’ Psicle software which consists of various gadgets which performs different types of routine and non-routine calculations.

I was proud to be part of the team representing Dynamo at the recent GIRO 2016 conference in Dublin, at which we publicly launched our actuarial and financial modelling platform, ‘Psicle’.

GIRO (General Insurance Research Organising) Conference, considered the premier conference for general insurance actuaries in the UK, is not only a great way to keep abreast of recent developments, but a great way to meet and catch-up with peers. As always, there was a diverse mix of presentations covering the key topics of pricing, reserving, capital modelling and the future of the profession, with the role of actuaries in data science (and data analytics in insurance) keenly discussed. Between the plenary sessions and workshops, we thoroughly enjoyed meeting people at our exhibition stand and our Psicle director, Shil Patel, was kept very busy demonstrating the functionality of Psicle. Social events included a visit to the Guinness Storehouse and the usual gala conference dinner, with some very original entertainment.

Thank-you to everyone who came to see us at the Dynamo Analytics exhibition stand. If you missed us at GIRO and would like to see Psicle in action, please get in touch.

And we look forward to seeing everyone again at the next GIRO…Edinburgh October 2017.