How to Use Big Data for Fundamental Investing: a Biotech Use Case

Recently, Empirical Research Partners showed how news and social media sentiment can improve the timing of stock picking, not only in the US but also internationally. It turns out it is particularly useful in biotech.

This new Empirical Research Partners’ study clearly shows that RavenPack News Sentiment is not only for quantitative investors, but surely also has applications within a fundamental investment framework. The webcast specifically addressed:

• Big Data sources for fundamental investors
• News sentiment signals vs. traditional signals such as price momentum
• Sentiment around earnings expectations vs. sell-side analysts' earnings revision
• Sentiment for timing entry into value stocks and as an input into failure modelling
• Event detection and sentiment to augment fundamental understanding of what drives long-term performance in a given industry

Peter Hafez, our Chief Data Scientist, summarized the study below:

Biotech Investing with Big Data

According to Empirical Research Partners (ERP), the biotech sector has significantly matured over the years, leaving mega-caps mostly indistinguishable from their pharma parents. Where they do still have a bit of an edge is in their ability to convert research into cash flows. However, they have been confronted with recent growth roadblocks that have significantly reduced their growth rate advantage over traditional pharma, taking them from a double-digit lead in past decades to a more recent +7 percentage point advantage. With patent expiries in the coming years, and considering their existing drug development pipeline, ERP argues that the outlook of biotech looks very similar to pharma.

Downward Pressure on Drug Prices

ERP highlighted another topic driving the convergence between big biotech and pharma: their shared exposure to downward pressure on drug prices. Using RavenPack sentiment data extracted from news stories about product pricing*, Figure 1 illustrates how the relative forward-P/E ratio of biotech to pharma stocks has converged to parity as the optimism surrounding biotech pricing power has waned.

*For example, a positive-sentiment pricing story would be something like “Eli Lilly raises prices for nine drugs, CNBC reports”, while “Pfizer to release new generic Viagra, slashes price by half” would be a negative-sentiment pricing story.

Given the intriguing relationship between drug price sentiment and valuations depicted in Figure 1, ERP took a closer look at whether media sentiment might add value to their modeling efforts in the biopharma space. In particular, they focused on the vitality of the product pipeline, a critical conduit of long-term success or failure in the industry.

Pharma & Biotech Product-Pipeline Sentiment Index

Using RavenPack’s database of news events, ERP created a product-pipeline sentiment index for pharma and biotech stocks, see Figure 2. This was done by hand-selecting 36 event types (out of about 3,000 in the whole database) that are related to upcoming drugs. These include stories about patent grants, ongoing clinical trials, FDA priority review grants, and so on. The line in each chart overlays the relative price-to-free cash flow multiple in each industry. Pharma in particular stands out: historically, the valuation of the sector has mostly tracked product-pipeline sentiment, except for the last 18 months where the two have been decoupled. In the report, ERP argued that this could be due to pricing concerns overwhelming everything else, something that would indicate potential optionality on offer, should the glide path of drug prices be less bad than market expectations.

Patent & FDA Priority Reviews

ERP also considered whether particular types of news have long-term implications for stock returns. For example, they highlighted that on average, stocks with positive patent news have outperformed their industry over the following year by more than +6 percentage points, while those with negative patent news have lagged by a point or two, see Figure 3. Similarly, stocks with news confirming that the FDA will grant priority review to a company’s product have outperformed by +7 percentage points in the following year, see Figure 4.

Product Releases & Product Pricing

Another useful event highlighted by ERP is the launch of new drugs, see Figure 5. Here, stories about successful new product launches generally carry positive sentiment, and such stocks have outperformed by about +2.5 percentage points per year. Negative sentiment stocks have lagged by a similar amount.

Finally, news around product pricing indicates that, in the long term, being all over the news because of big price increases is a counterproductive strategy as far as equity returns are concerned, see Figure 6.

To Summarize...

Putting it all together, Figure 7 summarizes the product- and pricing-related news events that have had the biggest impact on one-year forward returns. Given these results, it is no surprise that the Empirical Research Partners’ Media Sentiment Indicator has been quite effective in both biotech and pharma, see Figure 8.