On 1/22/19, I responded to a Bloomberg post slamming quants. Now, it’s time to address another attack, this one by Forbes Senior Contributor William Baldwin who, on 3/2/19 blasted AAII (American Association of Individual Investors) stock screening strategies suggesting that use of these as a way to build wealth is “delusional” and that you should put the bulk of your money into index funds and use screening only to “entertain yourself” but that with the latter, you should “not expect to beat the market.” That’s wrong.
© Cam Stock Photo, gwolters
Let’s Start With Some Disclosure
I post to Forbes as a “Contributor,” an outside contributor, and am one of many who does likewise. I’m not and never have been employee of Forbes. As such, I speak for myself (so long as I don’t write anything stupid, indecent etc.), and not for Forbes. Mr. Baldwin is a “Senior” Contributor who, according to his bio, has “been a journalist for 44 years, and was editor of Forbes magazine from 1999 to 2010.”
I am not employed or paid by AAII although I do have a “lifetime membership” and I do appear occasionally as a guest speaker at AAII local Chapter functions (I’ll be in Pittsburgh on 5/21/19); travel expenses are reimbursed but that’s all — the appearances are gratis (and enjoyable too; AAII members are smart interesting people and I love engaging face-to-face with them). I should also mention that nobody at AAII asked me, participated with me, or even had advance knowledge of my writing this post.
You can find my professional bio here. In addition to what’s there, from mid-2010 through mid-2015 I ran the screening-based Forbes Low Priced Stock Report, on behalf of Portfolio123 and, of course, Forbes in consultation with Matt Schifrin, who is presently Forbes Managing Editor of Investing, Markets and Personal Finance and who, in my opinion, has far more understanding of and respect for stock screening than seems to be the case for William Baldwin.
I consider myself a quant, a factor-oriented investor, but not in a conventional Fama-French way. As with the latter and the many inspired by their work, I sort a group of stocks and select the top “ranked” names, but my approach to factor selection is more pragmatic and fundamental than empirical and academic. Also unlike traditional factor investing, I believe screening, pre-qualification of stocks to be ranked rather than ranking against a broad “universe” is important, analogous to the way sales calls limited to those on a purchased list of pre-qualified prospects is superior to cold calling from a telephone directory.
The Point of Stock Screening
I’ll refer to the first paragraph of the Introduction to my book Screening the Market (Wiley 2000) — after pausing to give myself the obligatory pat on the back:
“Screening is absolutely positively the best way to find stock investment ideas. We’ll explore why this is so in Chapter 4, but for now, suffice it to say no other approach can match screening when it comes to calling stocks to your attention based on at least some objective showing of merit and without regard to what periodicals you read, what finance broadcasts you see or hear, what analysts you encounter, what tips your friends provide, or where you live, work or shop. And having the right stocks come to your attention is a major determinant of investment performance. No matter how great you are at deciphering financial statements, understanding business dynamics, or reading stock price charts, you have no real chance to achieve investment success if you always wind up applying your skills to stocks that aren’t really worthy of being looked at in the first place. “
Does that sound like a bad thing? Is that, as Mr. Baldwin suggests in the headline of his post, a harbinger of “bad news?”
Sure someone might hurt themselves doing it badly. Screening is hardly alone in that. Fire, dynamite, automobiles etc. were great inventions offering great contributions to humanity if used properly, or destruction if misused. So, too, it stock screening, capable of great benefits if used strategically or financial damage if treated like a high-tech toy.
Why Not Index?
OK. So screening is the best way to pick stocks. But that leaves open the question of why bother trying to pick stocks at all, when it’s so easy to index nowadays by purchasing and holding the S&P 500 SPDR ETF (SPY), or some other index-based ETF.
Well for one thing, picking SPY, is not at all a passive choice. It’s an active decision to favor U.S. large-cap companies and to do so with a momentum bias. You’d have to justify picking SPY in lieu of a small- or mid-cap index-based ETF. You’d have to justify a decision to stay domestic instead of going global (speaking for myself, I do have a home country bias, but and I acknowledge it’s a deliberate and active choice). You’d have to justify choosing a market-cap weighted index which includes a built-in inherent momentum bias, as opposed to a true smart-beta ETF (since that phrase has been and still is so often mis-used, if you want to learn about it, stick to Rob Arnott and Research Affiliates) or even an equally-weighted ETF that holds S&P 500 constituents such as the Guggenheim S&P 500 Equal weight ETF (RSP).
And then, too, how can the SPY be truly passive when it has no exposure to fixed income, commodities, real estate or who knows what other asset classes (crypto I suppose if you think that’s legit). The only genuinely passive investment you could make, if you can implement it, would be to track a value weighted portfolio containing all global asset classes with each according to its representative portion of the whole, or as a former boss of mine (Gabriel Burstein, are you out there?) put it, “God’s investment portfolio.”
OK. I know I’m being something of a you-know-what, so I’ll just pretend, for the rest of this post, that you really can put all or most of your money in SPY and say you’re not making active (albeit possibly unwitting) choices.
That raises the need for me to address Mr. Baldwin’s claim that you shouldn’t screen (except for entertainment) because you can’t consistently succeed in doing well against the S&P 500.
Not everybody can use screening to beat the S&P 500. Not everybody can write well enough to get published. Not everybody can act, sing or dance well enough to become make a career in the performing arts or even get in front of a room full of people without making a fool of one’s self. Not everybody can cook well enough be a chef, or even outperform take-out junk-food. Not every physician is good enough to avoid losing malpractice cases. Not every realtor can sell enough to make a living. I think you get the point.
Now, cue the violins: Does the inability of everybody, or even a majority of people to be good at something mean that nobody should try to do better? Are you listening that Tom Brady? Close the laptop and stop studying video because most quarterbacks can’t hope to do what you can. Seriously? (Actually, Tom, I’m a NY Jets fan so feel free to do me a solid and shut the #&%$ laptop!) How are you consuming this content (and Mr. Baldwin’s article)? Would you be able to do what you’re doing if countless people just threw up their hands and refused to try to be better because the average person is only average? OK, that’s enough; I’m sure you get the idea and am even starting to annoy myself. (For those who would like to see this topic addressed in a more literary manner, consider “Harrison Bergeron,” a fascinating dystopian tale contained in Kurt Vonnegut’s 1950 Welcome to the Monkey House short-story collection (Warning: Although the story pre-dated the popularity of buying the market, reading it might make Indexers ashamed to continue advocating for it or doing it.)
Let’s all admit that in many (probably most) walks of life, it’s possible for some (or many) to be better and that it’s a good thing when people strive to do just that. The only fly in the ointment would be if we’re dealing with an area in which it’s genuinely not possible for anyone to do better, in which case, trying could cause more problems than it’s worth. That was the case, for example, with human flight for a gazillion years — until someone figured out that wing surfaces had to be curved. Mr. Baldwin states, at lest by implication, that this is still the case with stock screening. Let’s test his conclusion.
The AAII Screens
I don’t know exactly where Mr. Baldwin got his AAII screen performance numbers since he did not provide a link. I get different numbers. I assume though Mr. Baldwin is not making things up. More likely, I’m guessing, is that he’s using an older data set (he refers to some sort of comprehensive AAII review and if it it’s from one of their magazines, there would be a lag time for publication). I’ll be using numbers I got from the AAII web site on 3/2/19; anyone with an AAII password can see it (or newer data if you click in the future) here.
I went through the whole presentation and show, in Table 1, how many out of 59 screens outperformed the S&P 500 each year from 2004 through 2018, year to date 2019, and in the aggregate for the entire period.
Table 1: Summary of Performance of 59 AAII Screens
Mr. Baldwin says only 30 of the screens beat the market. I’m guessing (as I must because he did not specify) that he was working off of data that was complied late in 2017-18, before the rallies that occurred in early 2019.
In the quant community, 2018 will long remembered as an epoch firestorm. There’s no way around it. I stunk it up in 2018. So, too, did pretty much everybody who invests based on data. My above-cited 1/22/19 post will provide a flavor of what was going on.
Was Mr. Baldwin doing wrong for using the dataset he used? Am I doing wrong using the dataset I’m using? Answer: None of the above.
The challenge, when reviewing investment performance (as opposed to a movie) is that there’s never a clear-cut start date (we can always debate choices) and there is NEVER a clear-cut end date — because there is always tomorrow, next week, next month, next year. So it’s never proper to run one set of numbers and say one has the answer. One must instead run many sets of numbers and using the constellation of all the answers, try one’s best to discern the differences between abnormal and normal occurrences. (Imagine if we stopped tracking after 12/31/2008!) Based on the more comprehensive approach I propose, I’d say AAII and its screens are doing quite well for its members, as are folks like Charles Rotblut and Wayne Thorpe, who work hard to explain and educate.
As to Muhlenkamp in particular, a screen to which that Baldwin give especially harsh attention, I’m perplexed by his thesis. According to the numbers I’m seeing, the AAII Muhlenkamp screen is OK, not the best but definitely not nearly the worst. What Baldwin does, however, is try connect the AAII Muhlenkamp-inspired screen to the performance of the Muhlenkamp Fund (MUHLX) which looks as bad as Baldwin says it is.
But it’s wrong, absolutely, positively, completely and indisputably wrong to tar AAII for the misfortunes of the fund. Personally, I don’t know about the Muhlenkamp approach to stock-picking and found pretty much nothing on that firm’s web site. AAII’s web site, on the other hand, is magnificently informative. It describes at length, and in highly readable language, how it built its screen (as it does for all of its screens) and there is one especially relevant tidbit: “While much of Muhlenkamp’s approach is qualitative, we have some quantitative data to construct a basic screen to seek companies with high return on equity ratios that are trading at reasonable prices” (emphasis supplied).
Bingo! Qualitative. That’s not screening. Qualitative is subjective and may or may not be disciplined. Screening is objective, quantitative and disciplined. AAII did not replicate Muhlenkamp. It used principles associated with him to design and build a their own screen (as they do for all “gurus,” many of whom achieved maximum fame before screening was invented).
So Baldwin is 100% wrong, The AAII Muhlenkamp screen cannot be used to bash AAII. To the contrary, it’s something about which AAII should be proud. AAII’s objective screen outperformed the subjective judgment of the guru whose ideas were being applied!
Can a Screen Outperform It’s Own Creator?
Yes it can and often does.
The comparison between the human Muhlenkamp and the AAII Muhlenkamp screen is not an aberration. Much to the chagrin of my ego, I’ve experienced it many times on occasions where I subjectively tried to pick the best ideas from my screens or cross out what I thought would be duds. Usually, I’d have been better off just leaving the screen alone and going with those.
Case Study: The (Former) Forbes Low-Priced Stock Report (LPSR)
The LPSR was born in 2010 out of a shared belief held by myself and Forbes’ Matt Schifrin that stock screening could be a successful way for investors to navigate the often treacherous wild-west-like world of stocks priced at single digit levels, an area regarded by many investors as a guilty pleasure but one filled with too many companies that, to put it politely, were not exactly stars of either the business or financial community.
I built a Portfolio123 screen to identify potential candidates and used a Portfolio123 ranking system to identify a Model Portfolio that could consist of the 15 top-ranked stocks from among those that passed the screen. Wishing to offer more variety (and concerned that 15 names be represent too much concentration for such a volatile area of the marketl) I also presented a list that added the next 25 ranked stocks — I referred to the total (the 40 highest ranked stocks) as the Master List. From among each month’s Master List, I selected some for text write-ups. Needless to say, I tracked results of everything.
Table 2 shows the performance results of the screen, taking into account gradual escalations in the maximum allowable share price (which started at 3 and by the end, wound up at 10) and a late 2014 change in the publishing cycle (and screen refresh cycle) from the 15th of each month to the end of each month. The benchmarks used are SPY and the iShares Russell 2000 ETF (IWM). (Note: Believe it or not, the two consecutive 10.02 figures appearing in the SPY columns are accurate; I I double checked. They do, of course, cover different intervals so performance was not really identical.)
Although performance was successful and I was happy with the money I made by owning the entire Master List each month (I invested via Folio Investing, where transaction costs posed no obstacle), I decided to end the newsletter in mid-2015 as it became apparent that selling the screen (the screen results) was too darn difficult. Screening requires thought, but media works best with quips and clicks.
But I never did vaporize the screen, so just for the heck of it, I checked Portfolio123’s back tester to see how it did from 8/1/15 (the first month of non-publication) through the present. (Although I use the phrase back-test, this is actually out-of-sample since the screen was built in early 2010).
Table 4 shows comparisons of the performance of the stocks written up to these benchmarks over the time period when each recommendation was open and averaged them out over 612 write-ups during the life of the newsletter.
As I said above, I find it hard to compete in terms of my subjective judgment versus the screen. But at least the results here were still positive.
That said, in terms of subscriber dollars and cents, the main value was in the screens (and the willingness of subscribers to set up an account with a firm like Folio, where they could own the whole list without being burdened by commissions). But subscribers most valued the writeups, which produced the least and were a huge drain on my time. That was a shame. I really wish this could have kept going. Sigh . . .
But please, please do not try to tell me the S&P 500 cannot consistently be beaten.
The Pre-Built Screens I Built For Portfolio123
As is the case on other platforms (including, obviousy, AAII), Portfolio123 makes pre-built screens available to users, many of which can be used as is and others of which are for educational purposes (examples that can be saved by users in their own accounts and edited by them as they wish).
One such group of pre-built offerings is known as Reuters Select. This is a group of screens I originally created in 1999 for the old CD-based screener Market Guide for Windows. I called it Market Guide Select. When Market Guide was acquired by Multex, the screens migrated to that company’s version of the screener and the collection was called Multex Select, and when Multex was acquired by Reuters, it was renamed Reuters Select. I brought the screens with me when I moved to Portfolio123 (by which point, I gave up renaming them).
As screening techniques go, they are rather old and creaky. Before Portfolio123, I had no ability to create and build ranking systems (hence the screens were more useful as idea generators than as algorithmic portfolios since they often listed many passing stocks), less sophisticated custom formula capabilities, as well as other deficiencies the most draining of which was no testing capabilities (before I got to Portfolio123, I had to do it manually using old CDs I saved).
Figure 1 shows the performance of these screens over the last 10 years. What’s important, though, in todays world characterized as it often is by claims that quants rely too much on extreme backtesting (data mining, curve fitting) and produce models that looked good in the past but fail when used with live money going forward is that these results are very very “out of sample:” The start dates begin about 10 years after the screens were created. In fact, I can’t recall having even looked at them much, if at all, over the past decade.
Not all of these screens beat SPY. But enough did to make it clear that screening is not just for entertainment.
Figure 2, meanwhile, shows an interesting collection of pre-set screens. These were all built for educational purposes, to demonstrate to Portfolio123 users how to approach different styles. For each style there are four variations; Basic, Market, Industry and Comprehensive.
Basic demonstrates what I might call screening for dummies. These are the sort of naive filters that might be used by a non-investor to give a trade-show type demo that does, indeed, treat screening as little more than a toy. The Market screens use factors that compare companies to the investment universe while the Industry screens compare companies to Industry peers. The Comprehensive group are, well, what the label suggests; comprehensive, meaning they make multiple comparisons. And contrary to the old Reuters Select group, these demonstrated invest-ability: The results of all screens were limited to the passing stocks that were ranked in the top 15 as per the Comprehensive: QVG (Quality-Value-Growth) ranking system I built for Portfolio123.
Notice how most of the screens beat SPY over the 10 years after they were created.
Also, compare the Basic versions to the Comprehensive versions. As in most walks of life, grownup approaches tend to outperform for-dummies methodologies.
Finally, like AAII, I created a set of guru-inspired screens (which I combined with guru-inspired ranking systems to limit the number of passing stocks to an investable level). But in an effort to present an original name, I refer to my collection as All-Star Screens. Figure 3 shows how they fared relative to the S&P 500.
So please, please do not try to tell me the S&P 500 cannot consistently be beaten.
Some More Screen-Based Strategies For The New Portfolio123 Invest
Portfolio123 is in the process of introducing a web site that will contain a new Invest section for free use that includes a variety of investable strategies in stocks, fixed income, and mixed assets. If you sign up — for free — you can access the stock strategies, as well as detailed explanations, by starting here.
Table 5 lists the strategies and their respective total returns from their respective launch dates (i.e., the dates as of which each strategy was completed and turned loose, so to speak, for automated updating and monitoring).
NOTES TO TABLE 5:
- Secondary benchmarks are USMV (iShares Edge MSCI Minimum Volatility USA ETF); IWD (iShares Russell 1000 Value ETF); MTUM (iShares Edge MSCI USA Momentum Factor ETF); IWF (iShares Russell 1000 Growth ETF); DVY (iShares Select Dividend ETF); and VIG (Vanguard Dividend Appreciation ETF)
- Yields are 3.34% for Returning Capital to Shareholders (looks at dividends and share repurchases) and 3.89% for Sweet Spot Equity Income versus 3.24% for DVY.
- The Expecting Dividend Growth strategy is more yield oriented (2.89%) than the more-equity-growth like VIG ETF, which yields 1.87%.
- The Genuine Value strategy is courtesy of my Portfolio123 colleague Riccardo Tambara. Considering how horribly Value has performed during much of this strategy’s life, the performance record delivered by Tabara’s strategy is nothing short of amazing. But he won’t boast. We all know about the great equalizer — Tomorrow.
- Built for Stability and Sweet Spot Equity Income are revised versions of two Portfolio123 Designer Models in which I’m invested and which will be presented below.
- It’ll be interesting to watch the progress of Swinging for the Fences. I’m not cut out for pure unadulterated Momentum investing such as what we get from MTUM (the latter having been a major winner in last year’s quant bloodbath). I openly explain, in the strategy’s short description and detailed white paper how and why my version is moderated by considerations of Quality and Value. While it’s off to an inauspicious start, anybody investing in it should be going in with eyes wide open and a full opportunity to decide for themselves whether or not they think the balance I’m trying to achieve will ultimately be doable. Nobody can promise superior performance. But anybody can and should deliver transparency and the ability of a user to make a thoughtful choice, and is something I do here.
Real Money – Mine
Figures 4 through 8 are screen shots taken on 3/2/19 from my account at FolioInvesting.com. These show real-word (as real as it can possibly get) results of screen-based models.
First off is a model I refer to as Cherrypicking the Blue Chips, which has long been available for free on the Portfolio123 Designer Model platform and about which I’ve often written in the past; see, e.g. this 2015 Forbes post. I invested my own money in it as well, and Figure 3 shows my results through 3/2/19.
Figure 4 – Cherrypicking the Blue Chips
Next up is a Low Volatility S&P 500 Portfolio123 Designer Model in which I also invested. My Forbes post describing it can be seen here.
Figure 5 – Low Volatility S&P 500 Stocks
This strategy started well (delivering the decent performance and low volatility I sought) but hasn’t been a star lately. That’s OK. I’m comfortable with the strategy, I know that nothing works all the time, and understand that the last few years were dominated by high-risk high-momentum stocks. I choose to refrain from racing the S&P under conditions like that because I know that sort of story never ends and how badly momentum chasers often wind up after they get too giddy.
This is an important point. It’s never about a footrace between the investor and the S&P 500. It’s about rational choices that balance potential reward and risk. This is an example of where I think it’s time to tell finger-wagging scorekeepers to take a hike.
Now let’s look at an equity income strategy, another Portfolio123 DesignerModel in which I invested my own money. Here’s a Forbes post addressing it.
Figures 5 and 6 compare its live-money performance in my account to the S&P 500 and to the iShares Select Dividend ETF (DVY) respectively.
Figure 6 – My Equity Income Model Compared to SPY
Figure 7 – My Equity Income Model Compared to DVY
By the way, as of this writing, my model yields 3.76%, versus 1.82% and 3.24% for SPY and DVY respectively.
Next is one of my oddball Portfolio123 Designer Models, one which I refer to as Underestimated Blue Chips. It’s a value model but uses no value ratios. Instead, I adapt a concept introduced by Robert Shiller and expanded upon by Stanford’s Dr. Charles Lee, and adapted for invest-ability by me to focus on S&P 500 stocks for which a smaller than usual portion of the stock price is accounted for by “noise.” Don’t ask me to explain here; if you’re interested, check out this prior Forbes post.
Figure 8, which shows how it’s been performing in my Folio Investing account, is particularly instructive.
Figure 8 – Underestimated Blue Chips
Talk about starting off in a really rotten way, and then staying the bad course for quite a while. Naturally, as a continuation of basic due diligence, I re-checked the model after its initial pothole to assess whether I still believed in the concept. I answered “yes,” and then did what many gurus advocate. I put it put of my mind and let it do its thing on autopilot, intervening only to upload refreshed ticker lists to Folio Investing once every three months.
Then, on 1/31/18, I got an email alert to the effect that a Portfolio123 member had commented as follows: “A new all time high! I couldn’t find any other Designer Models that are at a new all time high.” I clicked on the email and, much to my surprise, he attached the comment to this model, this former dud. I checked again on Folio and saw the same thing. I know better than to take a bow for its recent relative strength; I never do that knowing there’s always tomorrow. So I can’t say whether the strategy will retain its winning ways into the future from here on out. I’ll just stick with my original thoughts upon launch and on my quick double check. I continue to like the strategy and believe it will work well over time, even if that time is not the next few months.
Again, scorecards and the like are for the Morningstars and William Baldwins of the world. I’ll stick with sound implementations of sensible financial theory — regardless of what happens to be in today’s box score.
And please, please do not try to tell me the S&P 500 cannot consistently be beaten.
A Final Diversion
I’ve tried my best to offer logic and evidence in opposition to Mr. Baldwin’s hit piece. But frankly, there’s only so much logic and evidence one can discuss before one’s brain starts to hurt. I think I’ve reached that point. So before closing, I’ll share one more strategy-performance screen shot from my Folio Investing account.
Figure 9 – Thematic ETFs
This is a collection of oddball-theme ETFs in which I invested. What, you might well ask was the logic behind my choices? Don’t ask. There is no logic, except that I thought they seemed cool.
The lesson: Even quants can have fun — and still not lose their shirts (at least not so far). And for having read this far and staying awake, I suppose the least I could do is share my fun ETF list. So subject to the disclaimer that the selections are backed by absolutely no discipline, logic or objectivity, here they are:
- Invesco Water Resources ETF (PHO)
- ARK Israel Innovative Technology ETF (IZRL)
- ARK Innovation ETF (ARKK)
- Invesco Global Water ETF (PIO)
- American Customer Satisfaction ETF (ACSI)
- Obesity ETF (SLIM)
- Equbot with Watson AI Total US ETF (AIEQ)
- ProShares Long Online/Short Stores ETF (CLIX)
- Innovator Loup Frontier Tech ETF (LOUP)
- GaveKal Knowledge Leaders Developed World ETF (LDW)
- Long-Term Care ETF (OLD)
- Amplify Advanced Battery Metals and Materials ETF (BATT)
As can be seen, I’ve been beating the S&P 500 for a long time. But I make no claim to having been gifted with unique genius. I’m just a guy, and by no means the only one, doing my thing with tools at my disposal that were unimaginable to past generations of stock market investors.
Am I an aberration, whether or not I want to believe it? I don’t think so. But one who wants to try to prove I am a special genius and that few can match my abilities (and hopefully do so without bulling me for PR services rendered) should be aware of the massive task they’d be undertaking. The reams of studies purporting to document the failure of active management (the sort of evidence on which my PR heroes would be tempted to rely) have nothing to do with screening or with what many quants nowadays do. They are old studies that explored the efficacy of seat-of-the-pants stock selection or supposedly systematic approaches done before personal computing, screening databases, and usable platforms became available.
I’ve gone so far as to argue that the dichotomy between active and passive investing is obsolete because it fails to address objective, systemic screening-based and screen-like strategies. I had proposed that this third approach, which sits between passive indexing and traditional active (subjective) investing be thought of as something like “Moneyball” investing (data driven decision-making in what was traditionally a subjective field like what author Michael Lewis described in his book about baseball). But the active-versus-passive dichotomy continues to be the rhetorical norm. That being the case, critics of non-index approaches now need to be extra careful about what, exactly, it is they bash, dig harder than in the past to try to get things right, and open their minds to new answers that might come from a proper research effort.
Opponents who continue to insist that the S&P 500 cannot consistently be beaten and that my track record suggests special superior intellect rather than professional use of contemporary resources will find that eventually, I’ll tire of debating and just give in — and allow myself to be worshipped. 🙂