Is screening completely worthless?

Screening for investment ideas is universally derided.  Value investors hate it, growth investors hate it, it seems everyone hates screening.  We're told screens never turn up good ideas, and screens are a one way ticket to value trap heaven, or a million other things.

The unanimous view that screens are worthless has led to a predictable result, almost no investors screens for stocks anymore.  In the past few years I haven't spoken to anyone who proudly admits to screening, myself excluded.  A few investors will mention in passing they use it as part of their strategy, but the importance is downplayed.

I believe that screening is just another tool in an investors toolbox.  Each tool has a role, and if we try to use a tool incorrectly we'll end up frustrated or with incorrect results.  A drill is great to bore holes, or twist a screw, but I've never tried to use the butt of my drill to hammer in a nail.  Yet that's exactly what investors try to do with screens, and inevitably they come up frustrated.

Imagine I gave you a giant bucket full of a variety of coins.  I told you "find me the most valuable coins in the bucket."  How would you proceed?  Because the bucket is big, and no one has endless time you'd probably try to create a search strategy.  The first step might be to divide the coins into respective groupings based on denomination.  Quarters in one pile, pennies in another.  From there you'd want to focus your efforts on the pile with the most potential.  If pennies from before 1920 are valuable you'd probably disregard the nickels and quarters and start looking at each penny.  But even within the pennies you wouldn't look at every one, you'd just scan the dates for anything before 1920 and put those into a pile.

If you found pennies minted before 1920 then you'd probably search through that much smaller pile for the ones in the best condition.  But what if you didn't find any pennies from before 1920?  Maybe you'd start to look through the dime or nickel pile.  This process would be repeated until you found the most valuable coins in the bucket.

Screening for stocks is the same as searching for valuable coins in a bucket.  The problem is people expect it to be something different.  They expect a stock screener to be like a magical tool that will sift through all the coins in the bucket and find the most valuable ones automatically.  

At the most basic level a screener is a tool that's used to limit the universe of stocks that an investor needs to search.  It doesn't eliminate searching or analysis, but it helps reduce the work load.

In our society there's a macho factor to doing tasks by yourself.  I get this, I'm a DIY-er, but it's not for the macho factor, but because I like to save money.  For many though it's a source of pride.  "You bought your Christmas Tree at a nursery? Ha. I walked through the snow in the woods to cut it down myself!"

There is a benefit to knowing how to do something yourself.  There's a cost savings usually associated with it as well.  But there's also a time component.  While it might be cool to rebuild an engine in the garage it also takes a lot of time.  If engine repair is a hobby and it's enjoyable then it's probably a good use of time.  But it doesn't make sense in most cases for a highly paid professional to take off a week and try to learn how to rebuild their engine when someone else can do it for them allowing them to focus on their work.

Investing like most things in life is about results.  We congratulate our kids for how hard they worked on the field, or how hard they worked in class, but if they're working hard and failing something is broken. Likewise hard work can lead to good performance, but it's no guarantee.  If a company is cheap the stock doesn't reward the investor who read every 10-K for the last 30 years any more than the guy who coat-tailed and purchased given they buy at the same price.  That's what's great about investing, but also frustrating.  One investor can put in loads of work, while another barely scans the financials on Bloomberg and if both invest at the same price and the stock rises 50% both investors gain 50% regardless of the work effort before committing capital.

Outside of investing screening/filtering/searching is hailed as a great innovation.  Instead of having to leaf through volumes of the encyclopedia one can type a few words into Google and in seconds find the answer they were looking for.  The amount of data and availability of it is greater now than it's ever been.  And ironically investors are now ignoring tools that help them sort through this data easier preferring a tedious manual approach.

There are a few man criticisms of screening that I want to address.

Garbage in garbage out

The biggest weakness of a screen is the underlying data.  If the underlying data is crummy then any tool built on top of it will be equally crummy.  This is intuitive.  How could a tool with underlying bad data produce good results?

The challenge is when investors want to run a screen across an investment opportunity set where little good data exists.  In this case the best research is picking up the phone and calling anyone who will talk.  I'd say this sort of scenario exists in maybe 1-3% of all investment situations, and if you find yourself in that situation a screen is not the right tool for the job.

Before running a screen make sure the data source used for the screening tool is sound.

Screens are historical not forward looking

A lot of investors are out searching for investments at inflection points, or for investments with catalysts that will catapult shares to the moon.  If you're looking for stocks where something abnormal needs to happen for value to be realized or created then no tool will be of use.  Investors in these situations rely on past patterns and gut feelings.  Will Sears retail turn around?  Past results indicate it's unlikely, but Lampert and Berkowitz have stuck their fingers in the wind and decided otherwise.

Thankfully for most traded companies the future will resemble the past.  A bad management team that has destroyed value will most likely continue to destroy value going forward.  Likewise a good management team that has grown a company will likely continue to grow it in the future.  The future is uncertain but in the business world no one likes change.  Changing suppliers is difficult, changing processes is difficult, changing a business is difficult.  Most business is biased towards sameness.  

Screens uncover bad investments

When a tool is used incorrectly it shouldn't be surprising if the results are incorrect.  The purpose of a screener is to limit the universe of potential investment opportunities one needs to research.  It's not to find a good investment.  The more strict a screen the more likely it will turn up garbage.

I have tried to craft extremely complicated screens in the past and the result is always the same.  I'll end up with three companies that match, and none of those three companies are worth investing in.  

What screeners are good for

Screeners work very well when one is looking to narrow down the investable universe.  For example, if I know I want companies with revenue why would I waste my time evaluating zero revenue companies? Rather it'd be easier to run a screen and ask for all companies with revenue greater than zero.  Is it possible I'll miss some company that the market thinks has no revenue but really does?  Sure it's possible, but I recognize and have accepted the fact that I will miss thousands of good investments because I don't have time to track 66,000 stocks worldwide.

I've had a lot of success pulling lists of stocks with broad criteria such as "stocks trading for less than book value." Or "stocks with insider ownership that are profitable."  These are not complicated screens, but they reduce my time weeding through companies I'd never consider investing in in the first place.

Screens are also paradoxically good at finding a very specific match.  When I built CompleteBankData one of my goals was to provide a search mechanism that would allow a user to search over 1,700 pieces of industry specific bank metrics.  While investors scoff at the ability to find something quickly non-investors cherish this functionality.  Bankers can find direct competitors in a few minutes, users can find potential clients that match their ideal customer profile, and researchers can find institutions that have certain characteristics.  The ability to search through an enormous pile of data and find an exact match is eye-opening.  Our users no longer have to open the metaphorical encyclopedia, they now have something similar to Google.

My parting thoughts on screening are varied.  I'd encourage investors to look at screens because no one else is anymore.  If you are looking for bargains it's helpful to look where no one else is.  Everyone is looking at VIC, Seeking Alpha, SumZero, hedge fund holdings and message boards.  But no one is screening to source original ideas.  There is value in going where the crowd isn't.

On another level some people like to work for the sake of working.  Maybe you like to read annual reports in companies you'll never invest in.  Or maybe you just like to analyze companies to pass the time.  Or maybe you have a boss that requires it.  In these cases screening doesn't make sense.  It's like the person who spends an entire Saturday and $50 changing their oil when it could have been done in 15 minutes for $20.  If you enjoy the process then why not?

For myself I want the results, and I want shortcuts.  I'm busy enough as it is, and if investing didn't have shortcuts then I'm not sure how I'd have time for other activities.  I'm a firm believer in the "work smarter not work harder" school of thought.  I've built my business and my life around this.  That with intelligence and proper tools I can get more and better things done.  I believe that screening to reduce the universe of potential investments makes sense for me.  If you're interested in reducing your workload, or spending more time analyzing rather than searching then I'd consider looking at screens again.

For the naysayers I'll leave with this last thought.  I've had a 3-bagger and a 10-bagger both discovered with screens.  They've had value for me at least, and I'm content to be digging where others aren't, especially with a power tools when the masses prefer to dig with their hands.











7 comments:

  1. Several of your posts hit on this more general phenomenon, that some consider a tool or technique totally worthless or, in opposition, the only one to which they appeal. It's telling of the average (ie. the majority) that many take such hardline stances, whereas a practically minded individual will use what's appropriate or necessary per the circumstance. And to that latter point, throwing in some philosophy, Cicero argues we might do what's befitting to the given situation.

    I think it's mildly comical that so many will blame a methodology when their investment turns south, for whatever reason, instead of reflecting on their myopic reasoning. But I think therein lies the opportunity to exploit: if some never use screens or, likewise, always use sum-of-the-parts models, then we might take this as information in our own evaluation efforts. Obviously, this was one of the points you were trying to make.

    My view aligns with the investment-as-art disposition. We can't afford to treat analysis in this discretized manner because a particular technique will only work so long until it doesn't, as the saying goes. We must take a more continuous approach, availing of as many relevant methods as possible, using a range of possibilities as opposed to two or three point estimators. Moreover, with Klarman in mind, we shouldn't necessarily bind ourselves to one particular niche category either. Value is like an ebb and flow; it sloshes back and forth amongst the asset classes as time progresses. I'm a proponent of nano cap equity, but it's not the only aspect of the cap spectrum in which I invest.

    I guess what I'm getting at is a contradiction to what I originally stated, the only hardline we should take in investing is that there are no hard lines (perhaps other than "value," which is vague anyhow as you've mentioned in your skilift story). We leave ourselves open to serious error when we continually fit all our theses into the same narrative box which, to me, doesn't fully reflect reality. Again dispensing with the philosophy, Nietzsche offers, "a man with a hammer sees everything as a nail." Again, it's an art NOT a science, and we should act accordingly, lest we deceive ourselves.

    As always, keep it up, you do great work. Hope I made sense.

    ReplyDelete
  2. anyone who says they don't screen is full of it. what they probably mean is: "I screened before it became popular" :)

    like anything its how you use it. think creatively and it can be a really small investment in time to test a theory or compile lists of companies that are similar only for reasons of your own imagination.

    i recall for example back in july 2014 (http://goo.gl/WsMLyS), OTCM shared growth, cash flow, valuation and dividend characteristics with just two companies in the investment universe: AAPL and QCOR. that alone doesn't tell you everything, but with investing as in life, the company you keep can be a strong signal.

    i'll close with a plug for screener; it's good value and while there's a bit of learning curve i've found the cust support very helpful in scaling it.

    ReplyDelete
  3. Good post, Nate. Any favorite screens, apart from the ones mentioned, and screening software?

    ReplyDelete
  4. I guess machine learning is going to change the way we do screening. We just need to wait watch. The technology is already there and it is getting better everyday.

    ReplyDelete
  5. I forgot to reply to this post but I've been meaning to ask you - what screeners do you like? Do you know of any good free ones?

    My favorites are ft.com and my library's subscription to value line. I used to have a subscription to stockscreen123 which was nice, as it allowed you to take their data, perform calculations with it, and then screen using your own formulas. I'm not sure why more screeners don't allow this since the data is already there. Maybe they don't want people plugging up their servers with all these computations or something.

    Also, have you ever given any thought to developing your own screener?

    ReplyDelete
    Replies
    1. Matt,

      I like FT.com. I use CompleteBankData.com to screen for banks, so I did develop my own. It's just more of a niche. Outside of that I have a Bloomberg that I've used to screen from time to time. It's nice. Not mind blowing. Screener.co has some of the functionality for a fraction of the price.

      Delete
    2. I use ft.com for screening and i never thought it's worthless

      Delete