With the GeoQuant platform/technology behind us, our political scientists have successfully and persistently predicted ~76% of commercially-relevant political events since June 2017 — a track record of which we’re rather proud. That said, we often pay more attention to the predictions we got wrong. By explaining why we forecast events incorrectly, we not only help improve our ‘quandamental’ predictive methodology going forward, but also help unpack the analytical benefits of a systematic, transparent, and model-based approach to quantifying and forecasting political risk.
With that in mind, below we tackle this weekend’s surprise election in Australia, followed by updates on our ongoing analyses on India and geopolitics in the Middle East.
We incorrectly predicted the Labour party would displace the incumbent Liberals back in December 2018, when both opinion polls and GeoQuant data were much more favorable to the challengers. Over the past six months, both types of data tightened, with the Liberals polling within two percentage points of Labour over the past month and GeoQuant’s Government Instability Risk indicator declining rapidly heading into election day. Yet the ending level of both the polls and our Government Instability Risk score — shown below relative to the past two general elections in Australia — still suggested a Labour victory come Saturday.
To wit, note below that Risk, although clearly declining, was still higher than (i) during the July 2016 snap election, in which the incumbent Liberal coalition barely prevailed; and (ii) before the 2013 general election, when the Liberal coalition defeated an incumbent Labour government to take power. This analysis supported a Labor victory, with higher risk suggesting that the current incumbents — the Liberal coalition and PM Scott Morisson — would likewise be displaced, and government turnover would occur.
Ultimately, however, the trend in risk — particularly from late March/early April, when the election campaign officially kicked off — was more important than the level. Indeed, our data suggests a similar dynamic/pathway for Australia’s Mass Support Risk indicator, which is a daily, inverse measure of mass support for the incumbent government. Mass Support Risk is a trial component of Government Instability Risk and related predictions of election outcomes.
Despite the lack of government alternation, the policy implications of this election may prove significant. The Liberal coalition had a majority after the 2016 election but scandal-driven resignations and lost by-elections had reduced the coalition to minority government status. As a result, the coalition could not pass bills independently and was stalling until the end of its three-year term because the probability of losing power was high. Now PM Scott Morrison’s government will be able to move forward on tax break campaign promises, including tax cuts and credits. Higher social spending plans by Labour will probably be shelved, but a weaker fiscal balance will nonetheless keep Sovereign Risk elevated.
Meanwhile, other expected policy change will be derailed. The election results will bolster the conservative coaltion’s increasingly restrictive immigration policy, which a Labour government would have loosened. Also, a Labour government was expected to address climate change aggressively, but the Liberal coalition has been, and likely will remain, the party of coal.
Modi on track for repeat victory
Meanwhile, another (and, ironically, more counter-consensus) election prediction we made last December looks to be bearing out, with Indian PM Modi set to retain the premiership with a majority of seats for the governing NDA coalition — including Modi’s BJP — in the Lok Sabha, per this past weekend’s exit polls. Although still the largest party in parliament, it remains unclear if the BJP will be able to form a governing majority without coalition partners;. Our Mass Support Risk indicator tells the story quite clearly, with Risk falling quite consistently over the course of Modi’s first term in office.
Heading into term two, it’s all eyes on the economy – particularly farm income and job creation more broadly, both of which have become major flashpoints in the run-up to the campaign period, and (per our Socio-Economic Risk indicator on the app) have seen Risk plateau at a somewhat higher level relative to that observed in the run-up to GE2014.
Middle Eastern geopolitics and Saudi oil risk
Iran’s alleged role in last week’s “sabotage” attacks on Saudi oil tankers and subsequent Houthi drone attacks on pumping stations serving the East West pipeline comes amid growing tensions between the US and Iran in the Persian Gulf, potentially exacerbated by yesterday’s (relatively minor/no casualties) Katuysha rocket attack on Baghdad’s ‘Green Zone, which houses the now partially-evacuated US embassy.
As analyzed in more detail last week, these incidents have accelerated our Infrastructure Risk indicator for Saudi Arabia, which per below (yellow line) has been building since January of this year and is now set to peak again in mid-June. The dashed red line overlays our Oil Price Risk (Production) indicator, a multi-country/multi-factor model of political risks to oil production that has also been increasing going into the attacks.
Yet as the graph helps illustrate, Houthi attacks on Saudi oil infrastructure have been an ongoing facet of Saudi involvement in the Yemen war, and Saudi Infrastructure Risk is still projected to remain just below previous flare ups in Houthi attacks, indicating relatively contained risks to oil production. To be sure, our Oil Price Risk (Production) indicator — which per the figure below, correlates with the daily Brent Crude closing price at r= 0.65 since 2013 — looks elevated through at least July. But it remains well under its recent June 2018 peak, surrounding the conterminous OPEC meeting at which the cartel agreed (albeit loosely) to cut production. As such, while geopolitical risks to oil are clearly elevated, our indicators do not anticipate the kind of risk spike that would come with a wider US-Iran conflagration.
Implications for US-Iran tensions
Of course, these events and this analysis come amid growing tensions between the US and Iran in the Persian Gulf, a geopolitical dynamic in which the Saudi-Houthi conflict — with Iran backing the Houthis to proxy fight the Saudis, themselves core allies of the U.S. administration — fits most explicitly. Indeed, these dynamics are so linked that our best proxy for forecasting the extent of current US-Iranian tensions is our daily measure/projection for Geopolitical Risk for Saudi Arabia, displayed below. Note that this figure looks similar to the Infrastructure/Oil Production Risk graph above: while Saudi Geopolitical Risk is projected to accelerate through year end, it will remain below recent peaks, suggesting a significant military exchange between the US and Iran and/or Saudi and Iran is unlikely.