The setup seems ripe for M&A activity. Further bolstering this line of thinking is Salix was previously a target of both Actavis (ACT) and Allergan (AGN), but talks fell apart during due diligence around inventories. However, an audit committee composed of independent directors have been convened, and I believe that the audit committee will uncover more accounting problems. For those who still wish to wager on a Salix acquisition, I have outlined a strategy at the end of this article.
One of the most useful tools for detecting earning manipulation is the Beneish M-Score. Developed by Dr. Messod Beneish of the Kelly School of Business, and detailed in The Detection of Earnings Manipulation. The M-Score uses forensic accounting principles to produce a "Probability of Manipulation" or PMAN for any given company by comparing two 12 month periods.
I was first introduced to the Beneish M-Score in Quantitative Value: A Practitioner's Guide to Automating Intelligent Investment and Eliminating Behavioral Errors. This book is excellent and I wholeheartedly recommend it.
In the book, the authors relate a story about an MBA team from Cornell which used the M-Score to give a "sell" recommend on Enron, a full year before it's collapse.
So what does the M-Score and its Probability of Manipulation say about Salix Pharmaceuticals?
In Quantitative Value, they noted the M-Score gave Enron a 30.5% PMAN. For 1Q14 Salix has a 50.8% PMAN. That is what I call a red flag. Even it's 2Q14 score of 5.8% is high enough to be considered a weak signal.
Let's do some more digging; what is the M-Score recognizing in Salix to give it such a high PMAN?
The Beneish M-Score is a multifactor model with eight variables, each being a hot spot of earning manipulating activity. The eight variables are below:
- DSRI - Days' sales in receivable index: Detect revenue inflation activity such a channel stuffing.
- GMI - Gross margin index: Firms with declining margins may be more likely to manipulate earnings.
- AQI - Asset quality index: Rising intangible assets may point to deferred costs, boosting earnings.
- SGI - Sales growth index: High growth firms have incentives to manipulate earnings when they fall short.
- DEPI - Depreciation index: Slower deprecation can artificially boost earnings.
- SGAI - Sales and general and administrative expenses index: Can point out overcompensated execs, who have been associated with earning manipulation.
- LVGI - Leverage index: Firms with increasing leverage are incentivized to manipulate earnings when faced with a shortfall due to debt covenants.
- TATA - Total accruals to total assets: Firms with higher accruals are likely to be earning manipulators.
Lets look at the table below to see how Salix scored on these variables compared to peers:
- Salix has been manipulating their expenses and will have to restate at least one quarters earnings. The method used is either intangibles to defer costs (AQI), or hiding something in accruals (TATA).
- The lower margins will be disclosed to be part of the channel stuffing (GMI), as will the higher accounts payable (DSRI), as discussed previously.
- Salix will not restate revenues, but may lower revenue guidance once the full impact of channel stuffing is known.
In short, I believe investors should be very cautious on owning Salix shares until the audit committee releases their findings. An approach for those determined to make a bet on a Salix acquisition should consider a small starter position, with the intention to use the negative headlines from the audit committee as an opportunity to buy on weakness.
"Never is there just one cockroach in the kitchen" - Warren Buffett
I have no position in Salix Pharmaceutics as of this writing, but may establish one in the future.
- All my data is from Ycharts.com, as is the Beneish M-Score calculation.
- Peer Group: US Listed Biotechnology, Pharmaceutical, and Generic/Specialty Drugs Companies with Enterprise Value greater than $3.0B, annual Revenues greater than $100M, and annual EBITDA greater than $0. After removing any companies with incomplete data, 38 remained.