Svetlana Borovkova: Factor investing and sentiment - a winning combination?

Svetlana Borovkova: Factor investing and sentiment - a winning combination?

Svetlana Borovkova (Probability) - 980x600.jpg

By Svetlana Borovkova, Head of Quant Modelling at Probability & Partners

A few months ago, here on the Financial Investigator website, I published two columns about the role of news and social media in investing. Today this topic is as actual as ever.

The events surrounding Reddit and GameStop having underlined again that, although the fundamental value of a company and the social media sentiment surrounding it do always go hand in hand, in some cases it is the sentiment – rather than value – that drives the stock price.

The case of GameStop is quite exceptional and, moreover, it was not so much the sentiment, but the “buzz”, i.e., the volume of talk, that drove the stock price to sky-high levels: in the few days after January 27, 2021, the amount of media buzz surrounding GameStop (and its close contender AMC) was five times higher than that for the two other newsworthy companies: Tesla and Astra Zeneca.

An exception to the rule?

Such extreme examples, however, are rare, and most of the time the media and its sentiment give quite an accurate – and early – indication of what is going on with a company, commodity, or an entire sector of the economy. This early warning indicator property of the sentiment gives all the more reason to incorporate sentiment in quant trading strategies.

The success of factor investing

The most popular quant investment strategies of the past decade are those based on “factors”:  specific drivers of stock returns, which have shown to yield good results in the past. The most common investment factors are size (exploiting the fact that smaller firms generated better risk-adjusted returns than bigger ones), growth (value stocks generating better returns than growth stocks) and momentum (firms that performed well in recent past are expected to also perform better again).

Investment strategies based on all these factors have been quite profitable. In the last decade, such factor investors generated approximately 2.5% per year over the market benchmark. So, a natural question is how media sentiment would affect or even replace such factor-based strategies? The answer to this question is not trivial and a lot of empirical work is involved. However, our recent research has shed light onto this question.

Sentiment as a factor

First of all, it turns out that, if sentiment is used as a single factor (this would involve buying stocks for which media sentiment is improving and selling those with declining sentiment), then such a strategy would generate returns that are very similar to the full-blown multifactor investing, so more than 2% per year over the benchmark.

If sentiment is added as a new factor to the multifactor strategy, that would improve returns by a further 1% per year. This comes from the fact that sentiment, maybe surprisingly, is almost unrelated to the traditional factors: returns on stock portfolios constructed  based on e.g., momentum, size or value are not correlated to those constructed purely on the basis of sentiment. 

Sentiment as a risk indicator

Sentiment can also be used in investing in a myriad of other ways. For example, since it is an excellent risk indicator (as I argued in my previous sentiment-related column), we can use it as a risk overlay in traditional multifactor strategies.

This would amount to monitoring sentiment for all stocks in which we can invest and only include in our multifactor portfolios those stocks whose sentiment is not declining (which would indicate risky investments at that moment). Such a risk overlay results into further 1% per year improvement in return, while decreasing portfolio volatility – a perfect recipe for improving Sharpe-ratio of your portfolio.

We can also use sentiment to screen entire sectors (and exclude those with declining sentiment), to rotate sectors in your portfolio, or can be combined (for example, as a risk indicator) with any other quant investment strategy – the range of possibilities is endless. One thing is certain: since sentiment is a reliable and timely risk indicator and it is not correlated to other investment factors, it will not harm, but most likely improve your portfolio performance.

Probability & Partners is a Risk Advisory Firm offering integrated risk management and quantitative modelling solutions to the financial sector and data-driven enterprises.