This article was written prior to last weekend’s attempted military coup in Turkey.
2016 is shaping up to be a rather eventful year. Markets have been steadily working their way through a long list of worries; the pace of Fed tightening post Liftoff[1]; a hard landing for China’s economy; the Brexit referendum result; and, more recently, the Italian banking system (see last week’s post). Also bubbling away in the background, but set to come strongly to the fore in the next few months, is the US presidential race between Hilary Clinton and Donald Trump. It is not just the immediate economic implications that flow from such events that investors need concern themselves with. There are also significant political implications that can have profound effects on financial markets over the medium-to-long term [2].
Last month’s Brexit result is a classic example. Reflecting concern about the economic ramifications for the UK, financial market volatility immediately surged in the wake of the referendum result. However, the decision to rescind membership of the EU also served as a catalyst for a wave of resignations by leading British politicians the like of which we have never encountered before (we stopped counting after five!). Most people would agree that in light of such developments political risk in the UK has risen since June 23rd[3] but what exactly does that mean?
According to Wikipedia:
“The term political risk has had many different meanings over time. Broadly speaking, however, political risk refers to… any political change that alters the expected outcome and value of a given economic action by changing the probability of achieving business objectives.”
Historically, political analysis has tended to be more of an art than a science due to the lack of timely data (infrequent, or ad hoc, economic data, surveys or opinion polls) whose output is difficult to incorporate into an investment process because of its intangibility. In dual recognition of the importance of incorporating political risk to enhance investment decisions[4] and the difficulty in quantifying it, several approaches have been adopted to try and rectify the situation.
One approach is to conduct a comprehensive review of numerous economic, political and social factors to generate an aggregate measure of a nation’s stability, which as Ian Bremmer astutely observed over a decade ago represents the “unit of measure for political risk”[5].
There is a great deal of merit to this approach and it does provide a quantitative measure of political risk. However, because these measures rely on proxy variables and there is often significant human intervention in the collation, estimation and aggregation process the outputs tend not be published at a high frequency (annually in some cases). This may be acceptable for a corporate client examining alternative locations for a major investment project, but perhaps not for a hedge fund manager considering whether to take profit on an equity trade.
In order to overcome this deficiency some organizations have adopted a different approach to estimate political risk – one encompassing financial market prices such as sovereign credit spreads, credit default swaps and/or volatility etc. While this solves the timeliness issue, the drawback with using financial prices as opposed to social indicators as proxy variables is that movements can very easily be driven by factors completely unrelated to political risk. As outlined in Bekaert et al (2015)[6] sovereign bond spreads can also be impacted by international economic and financial risk conditions, local macroeconomic conditions and secondary market liquidity all of which contains no information pertinent to the political situation.
In addition, this approach also implicitly assumes that political risk is strongly correlated with sovereign credit risk, which again may be invalid. Credit risk represents a country’s ability to repay its obligations, whereas political risk is more closely aligned with a willingness to repay – a subtle but important distinction. Similarly using sovereign credit default swaps (CDS) as a proxy for political risk is undermined by the fact that in the EU at least naked CDS positions have been banned as part of the response to the region’s crisis. Finally, as is well documented, financial market prices have a tendency to be very noisy. Hence, the gain in timeliness comes at a cost of a higher probability of false signals.
A new alternative, which bridges the gap between the two approaches i.e. one based on social indicators but maintaining the timeliness of financial market based measures of political risk, is to use crowd sentiment; an approach that has legitimacy because politics is, ultimately, heavily influenced by human behaviour[7].
Applying the same automated processes used to extract sentiment on thousands of financial assets from millions of mainstream news and social media published daily, Amareos tracks sentiment on various macroeconomic and social topics. It is from this latter subset that we are able to derive a high frequency macro political risk indicator[8].
Given the overtly political tone to recent events we have felt compelled to include this political risk indicator in several blogs. At the time we provided a brief description as to what the indicator incorporates, but given space constraints we were a bit short on details.
Specifically, our composite measure of political risk is the equal-weighted sum of five sentiment metrics: debt default, financial sector instability, government anger, social inequality and social unrest. This choice reflects the fact that not all political risks, or in extremis crises, are created equal. They can have multiple sources or catalysts. To illustrate this consider a few examples, beginning with the UK Political Risk Indicator (PRI) in the run-up, and subsequent to, the Brexit vote.
At the time PM Cameron announced the date of the EU referendum vote[9], the UK’s PRI stood at zero – its long-run average[10] – but it rose notably over the following three week period as the public were subjected to the scare tactics of Project Fear[11], heightening concern about financial instability. While a brief lull was observed after the initial wave of warnings, concerns about financial sector instability quickly resurfaced when official campaigning began in mid-April, rising further in mid-May when the BoE issued its report warning of recession in the event of a No vote.
By the time of the vote the UK PRI had increased to its highest level since the widespread and violent street protests of summer 2011 when social unrest surged. That the index remains at such elevated levels is a clear signal that despite the UK being a stable and open economy – and hence likely to be considered a low political risk on some metrics – the environment contains substantial political risk. Certainly, it will be interesting to monitor how the UK PRI evolves in light of the unexpectedly early victory of Teresa May in the Tory leadership race leading to her replacing Cameron as British prime minister this week.
Exhibit 1. UK Political Risk Indicator
Source: www.amareos.com
Turning to another example, we came across the following humorous tweet posted after the Brexit vote by Tom Nuttal, a columnist for The Economist magazine:
As is often the case with humour it contains more than a grain of truth. Until the last week or so, Greece’s PRI was running at very modest levels indicative of low political risk. More recently, there is evidence of modest sentiment contagion from the Italian banking sector, which is raising financial instability concerns in Greece. Also, there has been an uptick in public perceptions of social inequality, most likely triggered by the recent passage of yet another austerity bill by the Greek parliament – see exhibit below.
Exhibit 2. Greek Political Risk Indicator
Source: www.amareos.com
That said, the present relatively low reading for the Greek PRI stands in marked contrast to the elevated levels seen during the EUR crisis, when public concern about the government’s fiscal position was high fuelling strong readings in sentiment towards debt default. Also noteworthy is the fact that last summer’s rejection by the electorate of more fiscal austerity[12], which fuelled fears that Greece would have to leave the Euro, did not trigger a big increase in the PRI. This was a useful signal considering the Syriza-led government, despite losing the referendum vote, accepted the austerity package just a few days later.
Finally, and more forward looking, what about the situation in the US in the run-up to November’s presidential election? The US PRI has been fairly low over the last year or two[13]. However, it jumped very sharply over the last few days reflecting rising social unrest and government anger – obviously events in Dallas have had a strong bearing.
Exhibit 3. US Political Risk Indicator
Source: www.amareos.com
Given the vitriol seen already between the two presidential candidates, and Trump’s sometimes less than diplomatic language, we strongly expect the US PRI to rise as we head in to the final months of campaigning. By how much US political risk rises over the election campaign is another question? We will have to see. What ever happens what is clear is that heightened uncertainty arising from increased political risk is both a challenge and an opportunity for investors. And, anything that can give an edge, like being able to monitor political risk in almost real time, is a valuable addition to any investment process.
*Sentiment Analytics are based on MarketPsych indices.
[1] That seems like an eternity ago doesn’t it?
[2] A conclusion we noted in an earlier blog post see: https://amareos.com/blog/politics-trumps-all/
[3] See: http://www.ft.com/cms/s/0/9cc65980-3dfc-11e6-8716-a4a71e8140b0.html#axzz4E5ZLm25i
[4] The focus being not just on alpha generation but risk mitigation.
[5] See: https://hbr.org/2005/06/managing-risk-in-an-unstable-world
[6] Bekaert, Harvey, Lundblad and Siegal (2015) “Political Risk and International Valuation” estimate that on average less than a third of the sovereign spread reflects political risk. See: https://www0.gsb.columbia.edu/mygsb/faculty/research/pubfiles/13995/Bekaert_Political_Risk.pdf
[7] This is even true for the likes of North Korea and other undemocratic states, albeit on a much longer – multigenerational – timescale.
[8] Macro meaning country level political risk as opposed to micro or project specific political risks.
[9] February 20, 2016
[10] The Political Risk Indicator is calculated in such a manner that it is both comparable over time and across countries.
[11] For our take at the time see: https://amareos.com/blog/2016/04/20/brexit-project-fear-is-working/
[12] A vote that came to nought as alluded to in the previous paragraph. So much for direct democracy!
[13] Optically, the extreme spike witnessed during the Wall Street protests dampens subsequent risk levels.
