What do you use for stock screening and why? Why do you feel that the information you are looking at has not been automatically scrubbed and incorporated into securities prices by other investors or by computers running algos?
Curious to know if anyone is willing to share. Almost all of my best longs have come from things that screen poorly but have strong / reasonable economic value propositions at good / reasonable prices, almost all of which have occurred on the small cap side (less competitive).
e.g., BBSI at $10, AEPI at $22, LBY at $10-11, SGRP at $0.40, AM at $13, etc.
Best shorts have come from stuff that screens okay / well and has high investor interest.
e.g., BEAT, AMSC, QCOR, ZIP, LPHI, etc.
Crux of the question is, what exactly are people targeting on their screens (conventional screening wisdom), and why does / doesn’t that work?
Any book recs on screening? Trying to get to nail down the “conventional approach” for different investor bases – value, growth, momentum.
Assuming you really want an answer (your last reply to me was: “It would be better if you read my posts twice and then consider not responding at all, thanks”)…
speaking for myself only, I target value stocks with something similar to GARP. Low P/E, low P/B, low D/E, some dividends, some growth over the last 5 and 10 years.
This works most of the times, it worked for BKE and TUP a couple of years ago.
It doesn’t work for companies with significant buyback programs, since they have high P/B and high D/E. It doesn’t work for the most famous value stock, BRK/A because it doesn’t pay dividends.
BOLT was a good buy, I think we picked up a small amount in early 2009. Did not hold all the way to $14.
I also think screening is largely unsuccessful, but so many people swear by it, so I’m curious to know why. I developed a custom screen recently that focuses on narrowing a wide universe of qualitative information (the Medallion Funds of the world can’t game that with algos) and the hits are significantly better. Was trying to see if there is some way to cross reference the screen with a quant screen but I’m a bit skeptical.
Generally, if it can be screened for, it tends to get incorporated into prices automatically on the same day. So with LBY, for example, the stock just went parabolic after they had a good quarter based on price increases, cost cuts and a refinancing that lowered interest expense. You could have known all of that was coming in prior quarters if you had been watching their stock or read last year’s annual report letter, and somehow it still managed to surprise the “all knowing market.”
Anyway, LBY came up on the qual screen, and this would have been such an obvious long if I had seen it in time. Bought some at $8 in 2010 and sold around 15 after the big run up, would have bought back in after the refi in May if I had been watching it. If you looked on a quant screen, by traditional metrics it would have looked like a fairly priced boring company with debt – nothing to see here, move along, move along.
I find a lot of the free screeners to have horrific interfaces for executing any kind of nuanced screens, and I’ve been a bit lazy to code up one myself.
I’ve often wondered how valuable screens can be, because anything that can be screened for is likely to be arbitraged away rapidly by people bigger than me, and it is likely to get worse and worse. What passes a screen is increasingly likey to be something that hasn’t been picked up for a good reason. When fewer people knew how to code (which wasn’t so long ago in stock-market terms), screening was probably a significant advantage becuase not everyone was doing it.
I’m not saying that screening can’t be useful, but certainly simple screens aren’t going to work as well as they used to. So I wonder what other people’s experiences are.
Then again, I mostly work with indexes rather than stock picking, so you guys probably have better insight, and I’m curious to hear what it is.
While we wait for an enlightened post on screens, I’m curious as a tangent to know how you time your entry to an index. Presumbaly, this is part of an overall allocation strategy using indexing for equity exposure. How do you know when the right time is to purchase an index?
I’ve always wondered that because some people say you shouldn’t buy individual stocks because the market is efficient, so just buy the index. But then the key question is, when? You can easily lose money on the index as well with poor timing.
In terms of screening, I assume anything that can be quantified by computers has been and is cut to the nth degree by various firms looking for statistically cheap stocks. My impression is that, at best, you end up finding stuff on the low end of its trading range that may move to the higher end of its trading range. However, that’s not really “value investing” (or if it is, it is only in the loosest sense). Something isn’t a great buy because it usually trades below 13 and 15x earnings and it happens to be at the low end of the range, for example. Hopefully people are doing much better analysis than that.
All of the best stock picks I have seen require people to get in there and dig around on a company specific basis. If anyone here has SZ or VIC access, check out that ACTV report. That was solid work.
Well, remember that with tactical asset allocation there are two types of decisions you can make: one is a decision about whether to be in a market, out of a market, or short a market; the other is a decision about how much to hold if you are in. I’ve seen strategies that are just about going in and out, and strategies that are always exposed but have a high and low range.
Effectively we make decisions about whether to be in or out or short using some fairly simple techincal indicators that are mostly momentum. I can’t go in to the specifics here, but there’s nothing particularly arcane that we do. In the macro portfolios, there is also some fundamental analysis, but the thing about fundamentals is that even though you know something is overvalued, it can still go higher or stay high for a long time, so the technicals are actually pretty useful for gauging when people are starting to pull out. Not foolproof of course, but useful nonetheless.
Then there’s the decision on how much exposure to have. Volatility has a lot to do with that particular decision, so even if you still want to be exposed, if your asset is getting more volatile, you may reduce the exposure without getting out. It also can be done by budgeting your risk. Our process doesn’t do this, but I personally want to look into volatility targeting for my own stuff.
Finally, it helps to have more than one asset class or at least relatively uncorrelated things. The big challenge with things that resemble market timing is that you have relatively few investment decisions you can make at any one time. There are maybe 10 or 20 different things you can put into a TAA portfolio. Perhaps 50 at the most (but that would be pushing it, because lots of things would have to be either correlated or illiquid to get to 50). And you don’t want to be racking up too many transaction costs jumping in and out all of the time. In contrast, there are 1000s and 1000s of stocks to analyze, and even more individual bonds, so in some ways the size of the universe can help you find something that someone else isn’t watching. On the other hand, stocks tend to be fairly correlated with each other in the long term, which can make you wonder whether there is really all that much payoff in doing all that research if you’re just going to hold them for a long time. In major indexes, it’s a lot trickier to find something that isn’t being watched. But having more than just one or two helps a lot. If the indexes are relatively uncorrelated, you can get bettter risk reduction through diversification with just 4 or 5 asset classes than you could achieve with 20 stocks.
I might agree with this in large caps. In small caps, the market doesn’t doesn’t look out long-term due to institutional constraints (quarterly mandate, minimum market cap requirements, $5.00 minimum price rule, etc.). If you can build a book of longs and shorts in companies that will do well / poorly over time, you can both create significant alpha (assuming reasonable stock selection), and hedge out a lot of exposure to the market.
It’s hard to see why you’re going to have a lot of pay off researcing WMT, KO, etc.
I’m curious to know also if you can comment – do you use fundamentals to determine when to exit the index or just technicals? Some people seem to use historic P/E analysis, estimating S&P earnings, etc. I haven’t ever really gotten into that, but if there’s an argument that trying to forecast company specific results, then it would seem significantly harder to forecast index results, but maybe that’s not correct somehow.
The macro portfolio is more discretionary, so there is no hard and fast rule for fundamentals and exits. Also, I don’t directly manage that - I am often asked for opinions on trading ideas, either commenting on the PM’s thoughts, some other article that’s been published, or submitting some of my own.
But here’s what I can say about fundamentals vs technicals:
You might think that a change in fundamentals would be a good reason to enter or exit a position. But the reality is that fundamentals seldom change all that quickly, so they don’t really make good entry/exit signals. On those occasions when fundamentals do change rapidly, the technicals tend to parallel these changes anyway, so it’s just simpler to follow them in both cases. In the options part of the macro portfolio, fundamentals can help with determining reasonable strike prices, such as puts sold at prices where we think it’d be good to enter based on fundamentals, or determining the width of strangles.
To me, fundamentals are actually more about figuring out where the risks are. If fundamentals deteriorate, it generally makes sense to narrow your stops, but usually you still wait for a technical signal to exit, because you never know how high it can go. If something is clearly unsustainably high by a wide margin and you have a decent profit, then it generally sense to reduce your position or close it, but the decision is usually to reduce the position rather than to exit it entirely.
Thanks, bchad. In more liquid equities, I’m generally okay with decent fundamentals at a low valuation as long as I think I am trend right over time. You do not have to be the first through the door on a large cap stock provided it’s got a long tail to whatever is driving the stock higher (which should be knowable within a range based on fundamental analysis, particularly if you have a multi-year time horizon). The chart will tend to move first, and even if you do not catch the bottom, you can still make money by getting on the choo-choo train. I assume it is the same for other types of securities / markets.
In small caps that are less liquid, the approach is a bit different, but I agree with the general premise that once something is in the numbers (fundamentals), it’s “too late” and the stock has already moved. If it’s a first inning move of a long game, that’s fine, but the fundamentals are usually the lagging indicator. Sometimes the market does get surprised and you have gap ups and downs though.
What’s most interesting to me is watching stocks behave when they change from one shareholder constituency to another – i.e., a “value stock” becomes a growth stock or dramatically ramps as momentum investors get involved. This is the heart of the question – finding those sort of stocks via screens before they have major price breaks to the upside. I’m pretty convinced that quant screens can’t do that consistently, but I would love to hear someone prove me wrong.
I’m very much a beginner at this, and I’d imagine it takes many years to get experienced, but I tend to look for small-to-mid cap non-financial stocks that pay at least a nominal regular dividend, have low price-to-earnings, not too much debt, no recent radical price changes, below-average “hype,” publicity, and appearance in the news, and a solid history of steady earnings. If insiders are selling large quantities of stock, that’s a red flag to me. I like to check the credit rating on the company’s debt, if that is publicly rated, as well.
Those innumerable little bank stocks, they need different criteria for screening. Publicly traded REITs, yet other criteria.
But I try not to be too strict with any one criterion, because otherwise I wouldn’t have any stocks left.
I’m not active in the equities market lately like I was back in college. However, when I used the Bloomberg terminal to use screens for equities, generally it was to find things that looked interesting. I don’t think you’ll ever get a screen where you find stock picks every time you run it (Although, there is the ‘Buffet Screen’ that people say Buffet may use). I think screens are more for trying to find some stock that for some reason is trading differently than it should. These don’t happen all the time, but constantly running screens could find the stock whenever it becomes mispriced based on inconsistent metrics. Then of course, you have to go figure out why its priced that way and if its actually a market ineffeciency or not.
rawraw, good points. I’ve had success with the Bloomberg “MOST” function looking for stocks that are most down on a particular day – there are certain patterns that are interesting (like multiple gap downs in the stock of a good company after something has gone wrong and shareholders have thrown in the towel). You have to be very targeted with that, only focusing on high quality names, but it’s possible to find stuff that is irrationally sold off.
I have not really seen this discussed elsewhere, but from observation, the best time to get involved in a stock is when it has no natural shareholder constituency but there is a reason other investors will want to own it soon.
FBC was one I bought recently on that premise. It is a large Midwest bank that was valued at
So you have:
Stock that cannot be purchased due to institutional constraints – it looks like a penny stock even though it’s not by market cap (reason it is not owned)
Stock is added to the index and has a reverse split (ability to own)
Improving results (desire to own)
It was and is a cheap stock, so given all of the above, it was very likely that the stock would be rerated higher. It wouldn’t have screened well because until recently they were reporting losses.
My screening tends to vary a lot, depending on what I am looking for. If I want preferred’s/Exch traded debt I usually use http://www.quantumonline.com/, where I will either look at the new issues, or a specific screen (variable rate, industry specific, etc).
Otherwise, there are a lot of free screeners and even some pay screeners I have used to sort out based on different criteria. some may be growth based, GARP, value based, etc. But I would agree that typically it is hard for me to find value (since I would have to find something the market hasnt). Where I have found investments I liked it is usually in the micro/small/medium cap space, with either some specific events (e.g. spin-off) or trends to attract me (years ago I bought RAVN because I liked the area it was in).
The last type of screen I may do will be based on something I found in another stock/situation. For example, ORI, which I cant recall what the other company that I was looking at was that led me to it, but basically I started screening for high dividend, money losing, insurance stocks, trying to find something where I felt the legal entity structure might protect the ultimate dividend and book value.
In any case, screens for me are just the start in terms of time, which I never seem to have anyway…
I do a lot of the usual : low bv, p/e, etc. I do like low p/s in certain situations as well. I have had the best success actually with using screeners on regions as weird as that is. It seems to be useful for small caps though, where they may be operating almost exlusively in a particular region and if you know that region well it can be helpful. Utlimately screens are not that helpful though. I generally end up making a lot of adjustments anyways, so i’m more interested in business models at first.
I think a lot of the discussion to date here has focused around characteristics of specific companies, which is fine. That’s how many people set up their screens, and I do have screens that are designed for similar purposes. The problem is that screens aren’t all that useful in any given instant – they’re more useful for companies that you’ve been watching for a while, which means that you have to check screens regularly to figure out whether there’s an attractive entry point at a given time. If you’re only looking to dig into a company after it’s triggered something on your valuation screen, you’re probably too late to the party…you won’t have enough time to ramp up on the name to generate alpha unless you’re just that quick.
One thing I personally screen for (more now than ever) is coefficient of variation in estimates. It’s quite simple – for stocks that have a relatively large dispersion among estimates (be it EBITDA, EPS, sales or whatever), some analysts must be right while others must be wrong. It’s hard to generate alpha when everyone’s clustered around the mean, but much easier to find mispricings when there’s a meaningful spread. If I have conviction that I can understand the key drivers of value better than consensus, I dig in, take a position, and hope I’m on the right side of the trade.
Thanks, numi, that’s legit. I prefer companies with no coverage if possible, but wide variance would be a good screen too (controversy is good if you can do the work and get an edge one way or the other).
I agree also that once it’s in the numbers it’s too late. I’ve been reading about this in recent weeks and it seems that the best screens would be some sort of pre-catalyst screen.
Hey bromion, as you know already, my view on small-cap stocks that aren’t widely covered tends to be more about getting to intrinsic value and finding an information edge. That’s not to say these things aren’t important for mid- and large-cap names, but one reason I *don’t* often play in the large-cap space is because it’s highly unlikely for me to get an information edge there, and because analyst estimates tend to be pretty tightly clustered. That’s less true for the mid-cap space and I try to exploit the dispersion in estimates by doing better research and understanding the downside/upside relative to consensus.
One tip that I got from meeting with a manger of a prominent hedge fund yesterday is that for large-cap names, it’s important to understand downside/upside vs. consensus in all cases, but the “whisper” number matters a lot more for large-cap names. I don’t personally spend a lot of time talking to sell-side analysts, but there are some shops that do and the whisper is often different from consensus. You also get an edge in the large-cap space if you understand the whole value chain because that gives you a sense of the magnitude/timing of catalysts. You can basically get in and out of names if you can time the cycles correctly and use your industry knowledge to know when valuation is too low or too high relative to where you are in the cycle.
Let me know if any of this stuff isn’t clear…happy to elaborate more off-line about it. Thanks for your detailed response on my e-mail as well. We’re all square and I’m starting Monday. Happy Thanksgiving!
It is difficult to get an information edge in small cap names as well. Even in the 250M-1B sector there are a lot of people active in the markets and info gets priced in pretty efficiently, and I suspect these investors are more knowledgeable than the ones who are in larger caps.