Election sentiment and the Markets

Every four years, October brings with it a rather unique risk: The US Presidential election. Each election cycle is different. In most cycles, the Democratic candidate rails against the evils of big business while the Republican candidate supports business interests. However, as even the most detached election observer knows, 2016 is different.

In order to understand just how different 2016 is, we've worked with our partner ICG to generate an Election Sentiment factor. ICG is able to extract election sentiment by analyzing numerous social networks.

We then turned these sentiment metrics into risk factors that can expose how vulnerable investment portfolios are to electoral outcomes. This sentiment measure is relatively correlated to election polls (>30% correlated to FiveThirtyEight forecasts), except it is released in real time, which allows it to be potentially used in conjunction with the poll forecasts.

This data goes back to December of 2015, but election movements become more relevant to the markets as the date of the election gets closer. The candidate’s promises in the primaries are often different from the promises made in the general election as they move back to the center. In our analysis of which market sectors are exposed to which candidates, we look at one sample from April to September and one period from September to October.

Here's a look at how the relative change of Trump and Clinton sentiment impacts specific sectors:

The conclusion is clear: those with exposure to financials have a lot at risk from a Trump presidency.

DJIA

Individual portfolios can also be analyzed. One such portfolio is the Dow Industrial Average - a curated index of 30 stocks that are price weighted and thus somewhat wonky.

Note: These values are displayed using the ranking from our Decile Mode where 1 = 10th percentile, 9 = 90th percentile, etc.

No matter how it's broken down, the DJIA’s sectoral exposures are pretty balanced relative to the 2000 largest US stocks.

BRK

When we look at Berkshire Hathaway’s portfolio in Decile Mode, we see something different:

Interestingly, when Trump does well the portfolio is a little bit better positioned than average. Everyone knows that Buffett is a big Clinton supporter, but our analysis suggests that his portfolio would outperform the market if the election doesn't turn out as he would prefer.

It’s important to note that this approach of extracting sentiment data and turning it into a risk factor generalizes to more than just the US election. An election tracker might prefer to use polls (which have a lag) or prediction markets (which have large bid/ask spreads), but this methodology can be applied to situations where there are no liquid prediction markets and it can be used as a supplement to professional forecasters who often only release their projections after the market has moved.

Regards,
Omer