[10/10/19] Can You Distinguish Between Highly-Valued and Overvalued Stocks? Hint: Look at Job Postings

Amazon has a price-earnings (PE) ratio of 72 (as of October 2019), while star performers Apple and Microsoft have “lowly” PEs of 19 and 27, respectively. Thus, amazon’s shareholders are willing to pay $72 per dollar of earnings whereas only $19 and $27 for Apple’s and Microsoft’s earnings. Amazon is undoubtedly a highly-valued company by investors, but is it overvalued? The distinction is crucial: highly-valued stocks will likely persist having elevated prices, whereas overvalued stocks are, by definition, bound to fall in price, otherwise they aren’t over-valued. Overvaluation means that investors’ elevated expectations of future growth will not be met by future company performance. Thus, as its earnings or sales disappoint, the overvalued stock plummets. General Electric’s stock, for example, traded at $27 during March 2017, but as its grim operating and financial situation became gradually clear to investors, GE’s stock fell 40% to $16 (today $8.50). GE at $27 in 2017 was obviously overvalued, relative to its weak fundamentals.

I know this is hindsight. Not very useful for investors in real time. So, once more: Is Amazon with a 72 PE ratio overvalued today, or just highly-overvalued? Safe to say; no one knows. And yet, the distinction between highly- and overvalued stocks is crucial to investors. What should investors do?

You are now the first to hear about my ongoing research (with Xi Wu, a Ph.D. candidate) on the ability of corporate online job postings to distinguish between highly-valued and overvalued stocks. Naturally, highly-valued companies with bright growth prospects will search for lots of employees, whereas overvalued enterprises with faltering operations and stressed financial condition will look for fewer employees, if any. Is this prediction borne out by the data? You bet.

Xi and I use a unique database (Burning Glass) which reports the near-universe of all online corporate job postings in the U.S. This allows us to observe the complete job posting profile—the labor demand—of each company, classified by types of jobs and employee skills. We focus on annual changes (growth rates) in the number of a company’s job posting, relative to previous year’s postings.

As a preliminary test, we examined the predictive ability of job posting changes with respect to 1-3 years ahead of sales and gross margin changes. After controlling for known performance predictors, such as R&D, capital expenditures, and the price-earnings ratio, we find that the annual changes in companies’ job posting numbers strongly predict future sales and gross margin growth, up to at least three years. Interestingly, the predictive power of job postings—a non-accounting indicator—is stronger than that of R&D, capital expenditures, or PE. This gave us hope for the distinction between highly-valued and overvalued stocks by employee demand. We performed various statistical tests which confirmed the ability of job posting to distinguish between highly-valued and overvalued stocks, and I will report here one, intuitively obvious test.

First, we determined for each year all the highly-valued and overvalued stocks. Following the well known studies of Gene Fama (Nobel Laureate) and Ken French, we defined highly-valued stocks as those ranked top 30% of all stocks by the market-to-book (MB) ratio (the ratio of the total stock market value of a company to its balance sheet book value, or net assets). We then classified all the top valued stocks (roughly 1,500) to five groups of increasing annual change in the number of total job postings. Finally, and this is crucial, for each of the five groups of stocks classified by job posting, we measured the percentage of companies that were in the top 30% MB ranking in a given year, but dropped from this group in the subsequent two years. Obviously, a company which is ranked in the top 30% of stock valuation in a given year, and drops down from this group during the next two years, was overvalued in the first year by investors.

The table below reports our findings. The top couple of lines refer to the stocks dropping from the top 30% during the next two years, and the bottom couple of lines refer to stocks dropping during the next three years. Look, please, at the top two lines. Group 1 (left column) are the highly-ranked companies with the lowest annual growth in job posting.  A full 34% (0.340) of these stocks dropped down in price in the subsequent two years, indicating that about a third of the highly-ranked stocks with the lowest growth in job postings were likely overvalued in the first year. Now move to the right of the table, to column 5. These are the companies with the highest annual growth in job postings (lots of employees demanded). Note that only 19.9% (0.199) of these stocks dropped down in valuation during the next two years.


Thus, a full third of the highly-valued companies with low job postings were likely overvalued, while only a fifth of the companies with high job postings were overvalued. A difference of 41% (34% vs. 19.9%) overvaluation between low and high job posting companies. Job postings—a new indicator—can thus distinguish between highly- and overvalued stocks.

Our additional tests show that the predictive power of job posting with respect to overvalued, soon to fall, stocks is stronger for high tech firms, and for jobs requiring high skills.

There are now various free websites reporting on companies’ job postings (e.g., Glassdoor.com) which can be used by investors as an additional, potentially powerful tool to identify overvalued stocks, soon to tank. Personally, given my very low appreciation of the usefulness of financial report (accounting) data to investors, it give me special pleasure to identify non-accounting indicators, like job postings, which can assist investors in their valuation decisions.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s