Can News Sentiment Data Help Distinguish the Overvalued Stocks Most Likely to Fall?

Can News Sentiment Data Help Distinguish the Overvalued Stocks Most Likely to Fall?
geralt / Pixabay

Stocks that are considered overvalued can remain elevated for a long time before reverting to fair value and investors looking to short-sell such stocks can run the risk of entering the market too soon and getting stopped out. One of the eternal questions of investing is how to pick those stocks that are ripe and ready to fall versus those that still have higher to go. Timing, as ever, is key.

Get The Full Series in PDF

Get the entire 10-part series on Charlie Munger in PDF. Save it to your desktop, read it on your tablet, or email to your colleagues.

Q2 2020 hedge fund letters, conferences and more

News Sentiment Data Offer A More Reliable Stock Price Signal

Some investors look to price action for confirmation but according to analysts at Empirical Research Partners (ERP), a more reliable signal might be offered by news sentiment data, aggregated from myriad online news sources. Not surprisingly, news sentiment is usually pretty good for stocks that screen as expensive but when it turns sour, especially over a 3-month period, it is a strong sign the stock is ripe for shorting.

Q2 2022 Hedge Fund Letters Database Now Live!

Hedge funds HFMQ2 2022 hedge fund letters database is now up. See what stocks top hedge funds are selling, what they are buying, what positions they are hiring for, what their investment process is, their returns and much more! This page is updated frequently, VERY FREQUENTLY, daily, or sometimes multiple times a day. As we get new Read More

The U.S. stock market is trading just shy of all-time highs at the time of writing despite being in the midst of a global pandemic - a sign in itself valuations may be stretched - so it seems the perfect time to run a similar model to highlight stocks that are overvalued right now and vulnerable to downside.

The exact specifications of Empirical’s model remain a trade secret but we can pretty easily recreate a similar stock screener simply by combining our own overvaluation criteria with RavenPack news sentiment data.

Of course, there are likely as many ways to screen for stocks that are overvalued as to skin a cat. An initial attempt using an off the shelf screener provided by our data provider MarketInOut yielded too few candidates to generate a sufficient sample. Another problem with the model was that it threw up stocks like Amazon, which whilst trading at vaunted multiples, does not in this author’s view fit the description of ‘overvalued’ because it is a tech growth stock with the perfect business model to profit from the current COVID-19 crisis. In addition, the company’s levels of debt are relatively low.

In the end, we created our own model using three key metrics: a high P/E ratio, a higher-than-average level of debt-to-equity for its industry group, and a vulnerable PEG ratio, which is a gauge of future earnings. These three metrics encompass the three areas where companies can be vulnerable to reversals: an exceptionally high stock price, taking on too much debt, and some doubt as to the sustainability of future earnings. An initial screen threw up a large enough sample to provide us with a short-list without overly loosening the criteria for overvaluation.

For news sentiment, we used our own RavenPack news analytics platform and look for stocks from the overvaluation screener that are also in the bottom 20% ranked by sentiment.

Current Shortlist

The screener threw up three short-sell candidates that met both the overvaluation criteria and exhibited low news sentiment. Screens were conducted at weekly intervals from between June 5 and July 10 2020.

The first stock on the short-list was Brandywine Realty Trust, a real estate investment trust that invests in office buildings in Philadelphia, Washington, D.C., and Austin. No prizes for guessing why this stock is raising red flags - the commercial property market is highly vulnerable to both the post-COVID trend towards remote working and the economic downturn in general.

The second stock flagged up is Hurcos Co Inc, which is an industrial technology company, that “designs manufacturers and sells computerized machine tools, consisting primarily of vertical machining centers (mills) and turning centers (lathes), to companies in the metal cutting industry.” This is not as clear-cut a short-sell candidate as Brandywine, but it is showing signs of overvaluation and negative sentiment, so according to our model it’s vulnerable.

The third and final stock on our screener is Office Properties Income Trust, which is a “REIT, focused on owning, operating and leasing buildings primarily leased to single tenants and those with high credit quality characteristics like government entities.” This is a clearer cut sell candidate for the same reasons as Brandywine: trends towards remote working and exposure to the economic fallout from the coronavirus pandemic.

All three of these stocks are at heightened risk of decline over the next 12 months, as shown by the ERP research where overvalued stocks in the lowest 4 deciles of sentiment fell by on average at least 10% over the next 12-month period, although as already discussed the two models differ slightly.

Review From 12-Months Ago

We next thought it might be interesting to go back 12 months and see how well our model worked at selecting sell-candidates then. Had the stocks selected then gone on to decline as expected by the model? If so it would provide us with greater confidence in its efficacy.

Screening for stocks in 2019 gave us 6 short-sell candidates, listed in the table below, along with how far their share-price has fallen in the subsequent 12-month period.

We also include each stock’s ESG - or Environmental, Social, and Governance - rating as a further criterion, to see if that might be an optimizing factor.

ESG rankings

*ESG rankings are provided as percentage scores by CSRHUB.

Analysis of Results

The results show that out of the 6 stocks picked up by the screener between November 2018 and May 2019 only 3 went on to decline over the subsequent 12-month period as expected, although, aggregate declines  (-117.48%) substantially outdid aggregate gains (36.14%), suggesting the model may have efficacy at selecting stocks vulnerable to sell-off.

Impact of ESG Overlay

The addition of an ESG criterion does not seem to optimize the screener despite research showing that lower ESG rated companies tend to experience deeper drawdowns compared to their higher-rated peers.

Stocks that fell over the period had an average ESG rating of 46, which was actually higher than the average 43 rating of those that rose, suggesting in this case, at least, ESG would not have been an optimizing factor.

About RavenPack

RavenPack is the leading big data analytics provider for financial services. The company’s products allow clients to enhance returns, reduce risk and increase efficiency by systematically incorporating the effects of public information in their models or workflows. RavenPack’s clients include the most successful hedge funds, banks, and asset managers in the world.

Updated on

No posts to display