Does anyone have a screening rule or ranking system that screens out companies that use very aggressive accounting methods?
Valeant Pharmaceuticals (VRX) practiced terribly aggressive accounting methods and look what just happened to them. I’d love to be able to screen out companies like these.
I’ve been using Beneish M-Score as a screen, which is great. But VRX’s M-Score is -2.33, which is perfectly acceptable, while NXPI’s M-Score is -1.15–way too high–and it doesn’t seem to be doing anything wrong.
GMI tagged VRX as “very aggressive” for years based in part on ratios like intangible assets to assets, goodwill to assets, ppe to assets, accounts receivable to sales, operating revenue to operating expense, and inventory to cost of goods sold, all of which are measured in comparison to the rest of the industry, and all of which raise red flags if they’re too high. Unfortunately, since its acquisition by MSCI, GMI no longer produces AGR reports, and I’d love to be able to screen out companies like VRX. I tried setting up a ranking system based on factors like the ones GMI uses, but have no idea if its results are valid. Anyone have any ideas along this line?
I haven’t explored this very much yet, but two things I came up with seem very useful. One is to exclude companies whose net income exceeds their cash flow from operations for two years in a row. The other is to rank companies based on the following rule with lower ranking being better: abs(recvblchgttm/salesttm). Neither of these is going to exclude VRX, though.
Finally figured out a way to exclude VRX, BFCF, and similar very aggressive companies without hurting my bottom line. Exclude 1) high Beneish scores; 2) companies whose net income exceeds operating cash flow for two years running; 3) companies with high accruals. There seem to be two very different ways to measure accruals using the cash flow statement, one including investment cash flow and one not, so I use both. I now exclude the following: (netincbxorttm-opercashflttm)/asttotttm > 0.025 (this is almost the same as MScoreTATA) and (netincbxorttm-(opercashflttm+cashfrinvestttm))/asttotttm > 0.25. And like I said in my last post, it also helps to use in the ranking system the following: abs(recvblchgttm)/salesttm, with lower rankings being better.
I created a screen to visualize the companies who had a high proportion of annual EPS reports where their EPS was lower when including extraordinary items versus excluding the extraordinary items. I was thinking that one example of a quality measurement for a company is when they only have a small number of years (as a percentage of the years reported) where there is a variance. If they did it as a high proportion to the number of years that they report, then it might show they have ongoing cover-up of GAAP versus pro forma. At least that is how I am thinking about it. Comments are welcome.
I created this screen to help visualize it. It uses the following formula to count the number of occurrences when EPSIncl is less than EPSExcl and divides by the number of years reporting (so that younger companies are counted correctly).
(LoopSum(“EPSInclXor(CTR,ANN)<EPSExclXor(CTR,ANN)”,10))/(LoopSum(“EPSInclXor(CTR,ANN)!=na and EPSExclXor(CTR,ANN)!=na”,10))
Sort on the column @InclessthanExcl2YrsRpt. Lower is better.
Most companies don’t have a problem. The Median seems to be zero by eyeballing it. Notice that there are seven, though (including Beazer Homes) that have a bad habit of Inc<Excl.
I have not done any testing with this yet but if the median is 0, then it might show as a helpful ranking attribute. We’ll see.
I hate to tell you this but I would recommend abandoning the screen. Extraordinary items has been limited to the point of meaninglessness by accounting regulators. I’m not sure why they don’t come straight out and abolish the item, perhaps the year it will make them look bad, but that would be proper because they did bad. I suggest that all p123 members remove any InclXOR item form any model because some day, it will probably vanish due to long overdue regulatory action. And if they start using it, that will make things much more messy because . . .
Everything you think you are looking for in XOR is actually located in SpcItems. It’s good in that we at least know where it is. But it’s bad because it’s a pretax item and we do not know its impact on taxes, net and eps. (Once again, thank you FASB for letting companies off the hook when it comes to reporting; some report in the text portions of their statements but almost none report in numerical fields).
And by the way, none of this has anything to do with earnings quality. It tells you more about company/management/strategic consistency and it is something very important to be aware of. But nothing is distorted assuming users are willing to dig into the numbers and specify models appropriately.
The topic of Earnings Quality involves lawfully and accurately reported numbers that still suggest conclusions that may not be warranted by the underlying economics of the situation. That’s much harder to address since, in contrast to the situation with SpcItems, even a well specified model can get caught.
As mentioned, the Earnings Quality portion of the Strategy Design seminar is in progress. It’ll take some time to finish. For those who want to jump ahead on their own, check the book recommendations page I put in the Help section. One of the selections is an excellent book on earnings quality.
I’m trying to screen out a new breed of public company that has been on the rise lately. I have some rather blunt and crude ways to screen, but I was wanting to see if someone can come up with a more eloquent solution. The Special Purpose Acquisition Company (SPAC) or “Blank Check” IPO. You can read more here and here. https://techcrunch.com/2017/11/22/the-rise-of-spacs/
To quickly sum up, a SPAC is essentially an IPO that is basically a big bucket of pooled of cash for the purpose of fueling future mergers/acquisitions. An experienced executive is hired to consult and provide expertise on these acquisitions, usually with expertise in one particular desired potential taget industry. If this sounds a lot like the Valeant business model, you would be right, which is why I’m putting this in this thread. But there are a lot of holding companies that are perfectly legit business, and many of these SPACs not only slip through but do very well in quality and value based screeners and rankings because of their structuring (particularly EV based). And most of these companies have very compelling stories attatched to them that will pass the smell test of cursory homework. I’ve found, unfortunately the hard way a few times, that these are generally not good businesses to own. The SPACs have an incentive to aquire anything, so they do (and often get a big bump in temporary vol/price momentum). The “smart money” funds the pre IPO stage in exchange for warrants that they can sell after an announced acquisition pop, leaving new investors holding the bag. The investment banks get 10% of the IPO off the top regardless. They’re often acquiring private companies who have long wanted to go public but couldn’t pass IPO scrutiny but are cheap and have some sort of great turnaround story to them. The management has conflicted, or very short term, incentives (many of these “consultant CEO’s are only signed to 24-month contracts and are looking to pump and dump the price short term). All this to say, things generally look great until the quarterly earnings announcements when the rug is pulled back and you see what crawls out.
Since these are new (or newer) IPOs it would be easy to screen out something like close(365)!=NA, but some of these SPACs are several years old, and it seems a blunt way to go about it. Particuarly in dealing with small micro cap universes where there are a lot of companies who are not very old. I was wondering if anyone has a more elegant methold. If nothing else, maybe this will be a word of warning. I know based on the “Popular with our Users” panel on the front page that some of these have been popular holdings in the P123 community.
Tickers who fall into this category would be
NRCG ROSE DSKE RMNI YTRA LMB PRPL LEA PLYA HCAC AGFS YTRA LAZY PANL AMR AXRA ROIQ
Actually, as I recall, we did some testing a while back and Xor, in practice in the S&P database, is dominated by discontinued operations. If you’re SPECIFICALLY looking at discontinued operations for some reason then it might make sense to include it.
I should point out that I can’t imagine exactly why you would do that in the large aggregate that we tend to focus on. By definition, these are individual cases, and the easiest way to “eliminate” them is to just use ExclXor. But I thought that I should point it out. (I am otherwise in full agreement with Marc.)
I had a similar experience a few years ago with Chinese-based reverse mergers/IPOs.
The only thing I could think of was to create a curated list. But that is hardly elegant, and definitely no good for backtesting point-in-time fundamentals.
Compustat used to have a data handle for Initial Public Offering Date (IPODTE). P123 might be kind enough to add this. According to Compustat, “This item is useful in historical price
analyses and in screening a population.”
You could also screen out companies with major pending or past corporate actions, e.g.
PastCorpAct(#MANDA,#STATUSOK,365,#TRUEFALSE) = FALSE AND PendingCorpAct(#MANDA,#TRUEFALSE ) = FALSE
Oh, btw,
Using the type-param “#MANDAEXSPIN” causes an error.
A few ideas. These companies have very low or negative free cash flow if you define that as the sum of the operating cash flow and the cash flow from investments. They also tend to have low profit margins and extremely low projected profit margins (current year’s earnings estimate divided by sales estimate). They tend to have very unstable sales if you measure the difference from one quarter to another. Almost all of them increase the number of shares on the market rather than buying back shares. They have terrible momentum numbers by almost any measure, and extremely high price volatility. While by some measures of value, they’re pretty good buys, their forward earnings yields are very bad. They also tend to have pretty large long-term debts (though that’s not necessarily a bad thing). And while they might look good if you look at price to book ratio, their price to tangible book ratio is terrible. Hope that helps.
Yes. I will refine my rankings accordingly. The cash flow calculations are a very good idea in particular. That in particular does not fit the model of companies I want to own. The rest of the suggestions are good too, as usual. Thanks.