Photo: Wikimedia Commons
In its latest FX Pulse, Morgan Stanley looks at the idea of save-haven currencies, the three most note-worthy typically being The Swiss Franc (CHF), the Yen (JPY), and yes, today’s huge winner, the Dollar (USD).The firm looked at 18 outbreaks of war since 1990.
What it found, specifically, was that the dollar does best at the onset of a crisis. Following that, however, the Swiss Franc outperforms.
The full note:
In the full sample, the return versus USD on the announcement day was negative for both the safe haven CHF and JPY, as well as the higher-beta currencies such as AUD. In fact, on
the day of geopolitical announcements, it is typical for USD to rally across the board.
The expected divergent results are strongest in the week and month after an announced event (see Exhibit 3). The only reactions that were statistically significant are a stronger JPY
a month after. However, we note that the AUD response is only negative a week after and the oil response is statistically significant in the week after. These would be consistent with
responses to geopolitical tensions. As such, we would focus on the weekly response, in which CHF outperforms JPY.
Given the low level of significance, we split the periods into decades, looking at the increased amount and speed of information flow in the 2000s to see if that made a difference. In the full sample, the return versus USD on the announcement day was negative for both the safe haven CHF and JPY, as
well as the higher-beta currencies such as AUD. In fact, on the day of geopolitical announcements, it is typical for USD to rally across the board.
By looking at the events in the 1990s versus the 2000s, there was a rise in significant moves in the 2000s. However, the 2000s sample includes the only event to have occurred in a
G10 country – the terrorist attack on the US in 2001. Including this event generates statistically significant moves, and in the expected direction, in CHF across all timeframes. JPY was
not significant. Given the extremely large influence of this event, we re-tested the 2000s without this event.
The 2000s decade saw an increase in the Sharpe ratio of being long CHF and JPY in the event of geopolitical tension, compared to the full sample. It pays to be long dollars on the announcement day, but in the period after, the best outcome is to be long CHF and then JPY . However, the results were not statistically significant.
In almost all cases, it is more rewarding to treat CHF as the safe haven rather than JPY.
It is surprising to see USD generally weaker across the board (after the initial day). This pattern may represent the broadly negative environment for the USD through the decade, it may also suggest that the influence of these non-G10 events fades relatively quickly. The sample size is also much smaller forthis period. As such, we would not place any great emphasis on the specific numbers. What we would take from this is thatin searching for a safe haven post-geopolitical tensions, we would prefer being long CHF rather than JPY.