I am shocked – shocked – to hear that impeachment talk is going on in this market!
I recently attended a large quant conference in Hong Kong during which I overheard some of the participants – not me - discussing what the recent political noise out of the US could mean. And well, you know, some of them – not me, of course – were wondering if, in fact, this could all lead to an impeachment of President Trump, and if so, what it might mean for their portfolios. Some – not me – were wondering if a stress scenario could be built to model the impact of such an unlikely even, and if so, what this scenario might conceivably look like.
At this point, I will go on the record saying that political impeachments – presidential ones no less – do not lend themselves well to stress scenarios given the time frame over which these usually play out and the fact that when they come, they are no longer a surprise to anyone, and sometimes even come as a relief which leads to a market rebound. In fact, using the recent example of President Park’s impeachment in South Korea, talks of impeachment started in the fall of 2016, she was then first impeached by the national assembly in December, and eventually by the courts on March 10 of the following year. From November 2016 to the end of March 2017, the Kospi was one of the best performing markets.
Likewise, in the six months surrounding the last US impeachment (vote to impeach on December 18, 1998, Senate acquittal on Feb 12, 1999), during the period of Sept 1, 1998 – Feb 28, 1999, the S&P 500 was up about 26% (one day change on Dec 17 – Dec 18 was up 1.5%), while US BBB corporate spreads tightened from 169 to 156 bps over the same period. Of course this was the beginning of the dot com frenzy and plenty of other differences (probably few people thought the Senate would actually convict). Nevertheless, like a child with a scab, we find it hard to leave alone, and so we embark on this exercise in numerology and offer one such possible scenario, the timing of which would most likely coincide with the discovery of a smoking gun (if any) rather than the actual impeachment itself.
Despite succession being solidly established and the Republicans keeping their majority in congress, a reversal of the Trump rally, under the assumptions that it would cause delays in the passage of the pro-market legislations it was based on, seems like a good place to start. Next, the Japanese Yen and Gold, two safe-haven assets with a penchant for kneejerk reactions to political uncertainty, can both be expected to rise by 10%. Next, given that the US economy is the world’s engine of growth right now and any speed bumps can be expected to drag cyclical commodities down, we would add a 10% fall in the price of oil to our scenario. Finally, we modelled a volatility shock by giving a 20% rise to the Vix (US Equities), vStoxx (European Equities), and the JNIV (Japanese equities).
Since we are attempting to predict what is to be, it seems appropriate to reflect on what has already been, for everything that happens in the market at any time has a genuine resemblance to what happened in previous times. The results of Brexit having not yet left the collective consciousness, we used the previous two years as guide to calibrate the correlations between the constituents of the FTSE World Developed Index and each of the independent factors in our scenario.
We report the results as a percent change in the present value of the FTSE Word Developed Index for both a one-day (instantaneous) and a five-day holding period. The instantaneous one-day results implied a 7.8% fall in the index, while the five-day results showed a 9.7% fall. The US component of the FTSE World Developed index representing 58% of the index, contributed 6% and 6.1% of the decline in value in the one and five-day results respectively. To put that into perspective, at their worst, Brexit had caused about a 10% fall in the five major European stock markets before all (except Italy) started recovering. The US markets lost only 3.6% that day (with the benefit of a delayed reaction).
As stated above, designing a scenario to model the impact of political events is more art than science, and readers should therefore feel free to find any such results befuddling and ferociously disputable. The sudden appearance of this disclaimer being offered as an endnote, like a fly hovering over a dunghill, should indicate that we can never definitively reconstruct a true picture of how markets will react to such events; one might as well try to complete a half-broken Rubik’s Cube.
 Given the current historical low volatility levels, we do not expect those to contribute much to our results – a (much) bigger shock would be needed.
 It should be noted that we used the market prices and index constituents as of January 31st 2017 for this analysis.