After the EU referendum vote last month, the British Pound suffered significant depreciation vs. the US dollar. The GBP/USD rate slumped by almost 8% in just a single day (June 24, 2016), which is an all-time record for this currency pair. On the other hand, the Japanese Yen appreciated vs. the US Dollar and the USD/JPY rate decreased by 3.4%. This is a significant joint extreme for these two currency pairs.
Figure 1: Pre- and Post- Brexit correlations
* The rolling daily correlation is estimated using a historical time window of 1 year daily observations and an EWMA decay factor, which provides good out-of-sample estimates.
It is clear that there are significant changes in correlations for three currency pairs: the three main European currencies vs the Japanese yen. The other pairs are not as affected. Figure two provides comparisons within Cognity of the correlation matrixes for the two dates: June 23, and June 24, 2016.
Figure 2: Cognity correlations comparison report – Pre- and Post- Brexit correlations
With Cognity we had the option to highlight large differences. In this case only differences larger than 15% are highlighted. We see the same three pairs – EURUSD vs USDJPY, GBPUSD vs USDJPY and USDCHF vs USDJPY. The most significant change is in the GBPUSD vs USDJPY pair.
Actually this is the most significant correlation change since the beginning of the century. The following figure provides the daily correlation changes in the GBPUSD vs USDJPY currency pair.
Figure 3: Daily Change in the historical correlation between GBPUSD and USDJPY rates.
The same holds also for the following EURUSD vs USDJPY pair.
Figure 4: Daily Change in the historical correlation between EURUSD and USDJPY rates.
Clearly the major impact and change in the FX markets correlations is driven by the British Pound's large depreciation vs. the US Dollar. And on the opposite site – the Japanese Yen appreciation vs. the US dollar. It should be noted that similar correlation changes were observed after the Lehman collapse in 2008.
Figure 5: GBPUSD, USDJPY rates and their rolling daily correlation.
Looking at the data above, the long-term daily correlation for the period January 2001 – July 2016 between those two FX rates is slightly negative (approx. -11%) with periods of significant negative and significant positive correlation. The interesting observation here is that the Brexit event entirely flipped the correlation from a relatively large period of negative correlation to a period of high positive correlation comparable to the levels of post Lehman collapse in December 2008.
BISAM Cognity® is a comprehensive, multi-asset class market risk solution for multi-factor modeling, risk budgeting & portfolio construction, stress-testing and real world risk measurement.