Use Red Flags to prevent losses

Introducing the red flag score

The term “red flags” is synonymous with “bad things are going to happen,”. For investors, they are a sign that you should likely stay away from that investment. All this talk of red flags being able to predict that a company is going to do poorly makes us wonder, can they actually?

To start let’s take the red flag score from the “Synopsis with Red Flags” tear sheet that contains 13 comprehensive red flags.

In the event you can’t read the photo clearly, the score looks for the following thirteen items:

  1. Bad Accruals
  2. Current Accrued Expenses are increasing faster than expenses
  3. Two bad days in a row with a 4.5% drop in share price each day
  4. Close has decreased over the past 20 days
  5. The company has a low Cash Flow score
  6. The company has a low Piotroski F-Score
  7. Outperformance while EPS 1Y is decreasing
  8. Average market cap is in the top ten companies
  9. RawClose is less than $2/share.
  10. EPS T12M has fallen by 25%
  11. Price is a third of its all-time high
  12. Cash is rapidly decreasing annually
  13. Cash is rapidly decreasing quarterly

If you decide to run this screen for yourself, it will take some time do to the complexity of the parameters we are looking for. Though there are 13 different parameters for the score, we may be able to get the results we need to prove the effectiveness of red flags, simply by looking for companies with a red flag score of greater than 3.

Since we pulled the red flag score from the Synopsis tear sheet, all of the red flags are automatically plotted in the results.

Let’s look at AXON, as an example. They are a large company worth over $2 billion dollars in the fascinating sector of biopharmaceuticals. What you wouldn’t usually know off the bat, is that AXON exhibits four of the thirteen red flags. That may not seem like a lot, but let’s see how well this screen performs over the past twenty years.

That chart is all over the place and is pretty much unreadable. Let’s go ahead and put it on a log scale to make the data more comprehensible.

That’s much better, and just what we hypothesized would happen at the beginning of the article. It may seem self-explanatory that “red flags” would result in poor performance, but it may not have been foreseen that a company with as few as four red flags could result in the loss of 99.85% of your initial portfolio. At the end of the day, a Red Flags score is just that, a red flag to tell you to stay away from something, or jump ship while you still can. Personally, to be on the safe side, I would overlay a plot panel(which has been prebuilt into the software, meaning all you have to do is select it under the trading rules tab) to the screen to quickly identify companies with a large number of red flags and shorten my analysis time even further.

One of the stocks you are holding may need attention.  It may need to be sold. Holding it may cause you to give up all the gains you’ve accumulated. What makes it a good time to sell a stock you believe in? We’ll talk about a few good (and not so good) alerts, and discuss what makes a good alert.

Down 10%?

How about if your stock has gone down suddenly (say 10% in a single day)?

Verdict: No — this is not a good alert because it is far too volatile and does not consistently under-perform the market.

Overvalued?

How about companies that are grossly overvalued, have no earnings, but do have positive estimated earnings?

Lets look at our -99.8% loss another way: year by year.

It went down almost every year, making it a wonderful short screen, but a terrible long one!

Verdict: Definitely sell.

Sudden Gain?

How about a sudden gain, with no change in earnings (estimated or announced) or sales (we’re rebalancing daily)?

What if we break it down by year?

Verdict: If you can’t figure out why the stock went up, sell.

What are your favorite reasons to sell?

Here are some of my personal favorite reasons to sell:

Went up 40% in one day

What should you do if a stock you are holding goes up 50% in one day? For example, yesterday night at 5:30 I found myself contemplating this happy question on ARTX (Aerotech, Corporation), which had gone up 50% on news of a large order from the U.S. Department of Defense. I did what I usually do when confronted with such questions — I used Equities Lab to get an answer.  

How to ask the question in Equities lab, though? I chose the most naive query possible: “Find me all stocks that have gone up at least 40% in just one day”, and then backtested it, rebalancing weekly on Mondays? 

That doesn’t look good!  Maybe I have micro cap stocks (even smaller than ARTX’s 80 million) messing up the results. Or perhaps penny stocks are causing the problem…  Either way, I realized I was going to have to investigate further — a possibility at a 200% gain (.5 to 1.5 in 2010) did not make up for the loss of everything thereafter.  Given that my rebalance period was only a week, such losses seemed to come promptly.  So I put in rules to rule out closing price below 2 dollars, and market cap less than 40 million, and got the following:

Decidedly not an improvement!  It was 5:40pm now — if I didn’t come to a decision soon, I would not be able to unload my ARTX, even in the after hours market.  Could crossing $5 dollars have made the difference, though?  What about before 2010?  Lets take a look at what I saw next:

Wow! I could sell short with this! No major updrafts, at least not when compared with the relentless downward trend. Going back in time we have:

Fine!  Let’s try the entire time period:

Thanks to ARTX, Equities Lab now has a short screener that can turn a dollar into less than a penny in 10 years. Shorting is a dangerous strategy, but I had all the data I needed to go ahead and unload my shares of ARTX.

UPDATE: I only managed to unload 90% of them — but that should shield me from the worst of the imagined future declines. I still think the company has huge potential, and may well buy back in after it goes back down some from its spike yesterday (more backtesting to do).  And I still think that the “Estimates Fun Top 10” screen is a fine screen. It is not its fault that ARTX had the bad manners to go up 50% in one day.  (OK, maybe it is…)

High standard deviation of closing price

Here is something interesting I found: stocks with highly variable closing prices do not do well. Not at all.

One would not think this to be the case, since all the finance textbooks talk about a risk-reward frontier. The idea being that as you add risk, you get more reward. Somehow, that’s not the case.  Yes, the screen sometimes goes up, but it goes down a lot more than it goes up.  Also the standard deviation on a monthly basis is a lot higher (surprise, surprise!). We rebalance quarterly: every quarter the portfolios are updated with new, equal weighted collections of matching stocks.  What if we tried different rebalance periods, such as weekly, monthly, yearly, etc.  We have a collection of rebalance periods that use different months, different parts of the month, etc, and we’re not afraid to use ‘em.  Monte Carlo, for the win!  All Monte Carlo means is that we run a collection of different rebalance periods, and show the one in the bottom 10%, the bottom 30%, the median, the one in the top 30%, and the top 10%. This gives you an idea about how much rebalancing matters.

Moral of the story: Avoid stocks whose prices bounce around like ping-pong balls.

Another UPDATE: I got curious what happened at each standard deviation slice. I was worried that one slice was terrible, and the others OK.

It turns out all the standard deviation slices beyond 2 are terrible. So stay away.

How negatively does a bad Beneish score affect your portfolio?

With all the recent articles regarding red and green flags, it’s time to review flags within the score. One such flag, the bad Beneish score, is a key component of the red flags score.

The Formula

Beneish score, Beneish formula

The line item, “Bad Beneish Score” within the red flags score is calculated using the above editor.

  1. The Beneish M Score over the trailing 12 months (The most recent calculation) must be greater than 0.
  2. The Beneish M Score over this past year (Two quarters behind the T12M calculation) must also be greater than 0.
  3. The company must not be classified as a Depositary Receipt.

The Results

Returning a total of 133 matches – some of whom are household names – there are a fair number of companies that you should be wary of when looking at new potential investment opportunities.

Beneish Score

Now, we don’t expect you to take our word for it, especially since there are recognizable names in that list, like $TSLA. So let’s go ahead and see how well this strategy performed over a period of 22 years:

There was a slight spike during the dot-com bubble, but those gains were quickly eliminated. You then spend the next 17 years hugging the breakeven line without actually seeing a profit.

Just losing money over the past twenty years doesn’t automatically classify something as a red flag. There are plenty of losing strategies that just caught a bad break. We want to ensure that this screener is consistently selecting companies with a high probability of dropping in price. As a standard, we want to ensure that of all positions taken on by the strategy over the past 22 years, at least 50% of them have lost money. Here our results are a little worse.

To add insult to injury, this red flag is fairly accurate at predicting companies that will eventually be delisted – and not because they are doing wonderful share buyback programs, but because their companies have become almost completely worthless.