Mitch Zacks on Upward Earnings-Estimate Revisions

apply a profitability screen to it. It's awfully difficult to kind of make conclusions on fundamentals without having such a screen, given that about a third of stocks in the Russell two thousand or in small caps in general aren't profitable at the moment, So we take that profitability screen and that whittles down the universe a little bit further. I think as of September we were at about eleven hundred

constituents left over from the two thousand. When we do that, and then we apply our momentum scores on a sector neutralized basis. All these factors are on a sector neutralized basis, but momentum the average of trailing six and twelve month

total returns excluding the last two weeks. We apply our low volatility factor, our value factor, which four small caps, we just look at the trailing twelfth month sales to price ratio, and our profitability factor, which is the average of trailing twelve month ROE and ROIC, and then the top descal of those percentile ranks once we average those out, is the final b MVP small cap portfolio. Now a little bit of stats before I go into the back test result. That results in a little bit higher market

cap weighting than the B two thousand. For example, so the Bloomberg MVP, the media market caps about two point eight billion. In that universe with the largest stock about eight and a half billion, small stock about two hundred and eighty seven million, So the market cap range of the B two thousand as a whole as much wider the same thing with the Russell two thousand. The media market caps about a billion. But onto results, the cumulat

return for this thing is pretty solid. So back to nineteen ninety nine through September twenty twenty four, you're looking at percent significantly lower volatility than the Russell two.

wholly owned money management subsidiary. At the time I joined, we had roughly fifty million dollars in assets under management, and over the past twenty some odd years, you know, twenty odd plus years actually at this point in time, we've been able to grow the firm from around fifty million dollars in asset to a little bit over nineteen billion dollars in assets under management and model delivery and advisement.

So you know, that's kind of my background. I've been working with quantitative equity portfolio strategies my entire career, and I've designed most of the asset management strategies that we implement at Zach's. Zach's Investment Management's a wholly owned subsidiary

of Zach's Investment Research. Zach's Investment Research is the second largest provider of independent investment research in the country, and our contribution to finance was that we created the concept of the quarterly consensus earnings estimate before Za's came around.

What people were doing was pretty much looking at how earnings evolved on a year over year basis, looking at how that earnings growth compared to sort of a trend line analysis based on historical earnings growth, and saying, well, earnings have been growing at you know, ten percent, in at twelve percent year over year earnings growth, it was

a good quarter. What we started doing is we created this concept of the quarterly consensus earnings estimate, where we took to the cell side earnings estimates by all the analysts following these companies, and we created a consensus and then when a company would report earnings, it would be compared to what Wall Street expectations were. If the earnings were greater than Wall Street expectations, it would be deemed a earning surprise, and if they were less than Wall

Street expectations, it would be a negative earning surprise. And we started to look at this data we created and started to say, well, how can we use this data

to effectively manage money? And what we found to be the most important use of the data is not to look at the year over year earnings growth and instead to focus on how the earnings estimates are changing over time and what we found to be the case is that there's some degree, it's a reasonably strong degree given finance, you know, time series metrics of earning of serial correlation with earnings estimate revisions, and that became the basis of

some of our proprietary models that we then started the asset management firm with.

So the first thing I started to do is to try and say, well, what if we create a portfolio where we simply own all the companies that are receiving upward earning customer revisions, we hold them for a certain period of time, and then we rebalance the portfolio. And what happens is if you kind of apply that sort of very very structured approach, you tend to have a

very high degree of turnover in the portfolio. Roughly sixty seventy percent of the companies that are receiving upward earning cestemer revisions at the beginning of the month are not receiving upward earning customary visions over the end of the month, So there tends to be a high degree of turnover.

You also tend to get a very very high degree of sector concentration, so you not only see a high degree of turnover, you see a level of sector concentration that increases sort of the volatility and tracking error of the portfolio relative to the benchmark. So what we started to do is saying, well, how can we control the turnover and how can we control the sector exposure to generate a return that's more in keeping with the benchmark.

And that's what I've been doing for many years, and in the ZAC small cap core process as a mutual fund CIX, which is a very small mutual fund which actually kind of helps us in terms of being able to move in and out of these positions. We're trying to control the turnover so that the turnover is around one hundred percent per year, and we're trying to control the tracking error as much as we can through optimization to try and keep the risk in line with our benchmark of the Russell two thousand.

revision so that they're positive. So we're looking for sort of four factors, agreement to the extent to which multiple analysts are revising their earning estaments in the same direction, magnitude, the size of those estimate revisions upside, where the most accurate or recent earnings estaments are coming in relative to the consensus, and the earning surprise, the frequency and magnitude of earning surprises that have occurred in the process that

we're implementing. We're using the Russell two thousand universe as the initial universe of selection, and we're adding to at any company that's in our current portfolio that may have been removed from the index at some point in time. So the union of those two sets is really our basis.

What I will say is that companies that are getting negative earnings are a little bit less likely to get positive earnings estimate revisions because they're sometimes in some sort of growth phase where the management is spending money for some sort of potential you know, jackpot in the future,

so to speak. So they might be a biotech company where the management has decided, hey, we're going to continue to spend money to engage in the drug development and our hope is that it gets approved at the end

of the entire process. So generally, those companies that are kind of in the development phase that are based on massive capital expenditures to sort of capture a market will likely not get upward earning cestmer visions, And in some of those instances, it's possible they get downward esimate revisions, but they're getting closer to their sort of goals in terms of growth that the stock price will respond positively.

So they are two separate ideas. One idea is, hey, can you look at companies that have positive earnings to do companies with positive earnings or positive expected earnings outperformed companies with negative earnings. And that has historically been the case, but that's been widely disseminated amongst managers. So I'm not positive that's going to be completely the case over time.

I think it might be a little bit better instead of sort of throwing out you know, I guess you know, roughly thirty percent of the Russell two thousand to instead say, okay, are these companies? These companies are expected to lose money, but they expected to lose less money this quarter? Are they expected to lose less money over the last month

because analysts are revising their earnings estimates upward. And what I've actually anecdotally found to be the case is companies that move from a nonprofitable sort of scenario where they're losing money to companies that are going to generate positive earnings tend to maybe not all the time, but have a potential for very very strong returns, and you don't want to necessarily remove those from the portfolio.

the future. But you have to remember that we again, the pure estimate revision universe has a high degree of turnover, and what we're trying to do is use earning sest revisions as an alpha source and then control the turnover and control the sector exposure. So the overall risk of the portfolio is in line with the Russell two thousand benchmark.

So to answer the question, what we're doing is we're using earnings estimate revisions along with another model in a multi factor model, but we're using that as an input to optimization that then can identify companies. So it's a subset of that group that might be potential by candidates, and we might even go outside of that group if the optimization indicates that adding that position could conceivably reduce tracking error.

that is outside of that sector. So, for instance, right now, there are companies that are construction companies that are helping construct sort of cloud computing centers, helping with construct infrastructure on semiconductor equipment, manufacturing plants, and things of that sort, and these companies will be correctly classified as construction companies.

They build things, they issue debt, they use the debt to buy to buy land to build things, and they sell the they sell their what their access or they sell their effective use to a company that needs to use that type of building or that type of source. But in terms of their contribution to risk, they behave much more like semiconductor companies than construction companies. So I tend not to try and think from a top down perspective. I tend to want to think from a bottom up perspective.

And what we're looking for when we do do some qualitative oversight in some of our other strategies is we're looking for lack of crowding in the market. So we don't want to look for a individual company where there's a tremendous amount of information or focus from either the media or from investors on what the outcome of or

how that company is evolving over time. But in terms of like if you had to pin me down in terms of sectors, I think industrials are looking relatively interesting at this point in time.

Some small cap funds are even holding large caps. So I'm just asking you know, for you is it Do you think this is due to small companies growing into midcaps or just they're being less available small cap companies.

cap companies. So we're agnostic in terms of capitalization as long as it is within the Russell two thousand universe or within our portfolio. But we are looking for companies are receiving upward earning cestemer revisions. We are looking for companies that have a reasonable score on our quality metric, and we are looking for companies that, when added to the portfolio, will conceivably provide us with the alpha score

while reducing the tracking error relative to the benchmark. So you know, part of this is that you know, midcaps have been performing a little bit better than small caps. As you go down capitalization ranges, the number of companies that are not quote high quality or good companies tends to increase. So as you go down to sub five hundred million dollar market cap, as you go down to sub two hundred million dollar market cap, the number of

companies that really shouldn't be public tends to increase. The number of companies that are kind of basically completely ignored by the market for usually a good reason also tend to increase. So if you start using quantitative factors that statistically have just gener returns, profitability, momentum, earnings, estimate revisions quality, that will always try and sort of push you away

from smaller cap companies. You also have the case that with estimate revision based strategies there tends to be a correlation between higher market cap and the amount of analyst coverage. So the larger the market cap, not all the time, it really depends on the sector and how much in vogue that company is, but generally the larger the market cap, the more analyst coverage you have. So again I don't

think it's a function of a qualitative decision. It's more a function of the result of the quantitative models we're using to try and generate returns. And my guess is in the process that was described at the top of the hour, there would also be a somewhat of a MidCap bias if you're using low volatility and price momentum and valuation metric to take a look at things based.

we're looking at. My experience has been that as you move lower in capitalization, the returns from sort of the smaller cap universe tend to accrue a little bit more towards technology and more towards some biotech companies because they're able to effectively compound over time, go from becoming a small cap company to a mid cap and maybe even potentially a small large cap company, and that compounded return compensates you for the losses that you have with all

these other companies. So that if you look at valuation amongst very very small cap companies, you can run into a somewhat of evaluation trap where you find companies that are attractively valued and there might be an initial push upward as the price book ratio you know, reverts to historical means amongst historical levels amongst sort of bank stocks.

But these companies don't have the ability to compound on an annualized basis at a high enough rate to compensate you for the smaller cap names that might not perform

well or might be value traps. So what I found to be the case is that valuation metrics seem to make more sense and the larger cap that you moved towards, and as you move towards smaller cap names, you know, you have the traditional Fama French research which shows this anomaloust return amongst small cap value names, and I'm not quite sure if that has been picked over by the market with the amount of capital that's that's looking for

these opportunities. So if you if you think about value stocks, and you're saying, well, do small cap value stocks necessarily always out reform small cap growth stocks? You're looking at this regression analysis that's saying, well, yes, small cap value

gives you a bias, But it's possible. We've seen sort of a such a flood of capital into the small cap space that that thirty five dollars small cap company, or that twenty dollars small cap company is now worth twenty two dollars, or is now worth thirty six or thirty seven dollars, and that would eliminate the annualized return from owning those small cap value companies, And anecdotally, that's

what I happening. So I think that as the market becomes more efficient, valuation as a metric sometimes loses some

of its historical outperformance. And I'm much more comfortable owning companies that are receiving upward earning vest revisions and controlling the risk through tracking you a relative to the benchmark, looking at the PE multiple, looking at cash will multiples, just eyeing them to make sure they're in line with the benchmark in aggregate, than saying let's try and select those companies in the Russell two thousand that have the lowest p multiple relative to the industry group that they're

trading in.

very well over the long term. Do you think that's kind of more of a function of factor crowding given your commentary on profitability and everybody kind of be in there, or do you think it's more, you know, small caps are inefficient or something else.

in any bank financing. They have an idea, they're going to create an online service, they're going to create a new generative AI, whatever it is, and they need a massive amount of capital to grow. Historically, in the seventies, in the eighties, in the nineties, even in the early two thousands, these companies would go public at reasonably small capitalizations.

They would use the capital and grow. Now what is happening is that the companies that are capable of growing with capital expenditure and growing very largely with capital expenditure, are staying private. And the companies that are non profitable are not growth companies that have gone public. They're remnant public companies that really shouldn't be public and are losing money for some purpose, for some reason, because they're under

competitive pressure. So what I think is happening is there's adverse selection within the smaller cap names, within the Russell two thousand that the companies that are going public are going public because they can't remain private and they can't attract financing to grow. So what's happening is that as those companies that show up as nonprofitable in the smaller cap name, they're more likely to be less likely to be able to compound on a very very high basis.

So it's you have this sort of two tiered market system. You have all these remnant companies that are public. They shouldn't be public, they are not generating any money, they have a low return on equity, they should not continue to trade. But there's there and there's no methodology to remove it. Meanwhile, the good private companies are staying private

and coming public at much much higher capitalization levels. So a lot of the research showing that hey, small caps dramatically outperform over time may be skewed until we kind of receive these companies that are going public and can grow in the small cap space. But in terms of you know, what happens is it's it's very interesting in the mark is that as an idea becomes readily disseminated and acted upon it, the market changes and it ceases

to reflect that idea. The biggest example I can come up with this sort of the accrural phenomenon, where there was this indication that there's an anomaly and maybe the anomaly was due to data mining, but everyone was talking about accurals. There was all this research done on purls and then if you go back and you look, you know, five ten years after the crural phenomenon was publicized, companies

that with attractive accurals are no longer outperforming. And I am I would be a little bit wary of going into the small cap space. And because this phenomenon where unprofitable companies have dramatically underperformed for such a period of time, and it's just accepted wisdom that if the company is not profitable, it has to underperform the profitable company that the market may have adjusted valuations in such a way that that no longer holds to be the case going forward.

And that's why I think it's important to focus on the earnings estimate revisions as opposed to necessarily the profitability or the valuations.

the technological change. And you have this, you know what the US economy is tremendously good at is this creative destruction. But this does may not hold with software for some reason.

So if you think about a company like JP Morgan, and you think about your favorite small cap bank, and you think about which company is going to be better able to increase productivity as a result of you know, if generative AI becomes as strong as you know, sort of optimist are anticipating which company is going to be better able to increase productivity. What should be happening is the it should help the smaller cap company a little

bit more. They can do more with fewer people. It's easier for them to adopt their procedures, it's easier them to change. But what we're seeing in the immediate future is it's helping the larger cap companies because they have more capital to expand, to create these structures necessary to integrate the technological change. And I think the future of small cap stocks is dependent upon whether the technological changes continue to be monopolized or come benefit the larger companies

more than the smaller cap companies. And that has to do with questions of monopolies from these sort of platform type large tech companies and the question of well, why aren't the productivity gains, why are they flowing from the largest companies to the smallest instead of from the smallest companies to the largest, which is how the economy always functioned prior to basically two thousand and I think that's

going to be the key question. The key question is whether a small cap company and you don't need many of them, you need one or two or three of them, can become the next Microsoft, can become the next AOL, can become the next Google, or whether that growth is going to be segmented to private companies and that any technological change is going to be gobbled up at the top of the food chain, preventing the small cap companies from growing. And I think that's the most important fundamental

factor driving whether small caps can effectively grow. I mean, all these companies, all these elements, they obey sort of power laws where you have small numbers of companies generating excess returns and compensating you for the majority of companies that do not perform well. And so the question of whether can small caps outperform, I don't think is a valuation question, and you might get a year or two years where valuations revert to normal levels in terms of

relative valuations. The key question is a growth question, and can within the small cap universe a company arise that can grow or are the sort of monopolistic platforms and the integration of new technology so great amongst the large cap names that it prevents these smaller companies from growing. And I think that is more the fundamental question of whether small caps can regain their historical annualized return levels.

That being said, small caps are very attractive from a valuation standpoint, just from a common sense standpoint, the trees will not grow to the sky, and it's reasonable to have some degree of diversification around a large cap cap weighted index.

don't have enough data points. But what what did seem to be statistically sort of significant is the more divided the government, the better the stock market seems to do.

So if you're taking a step back and you can remove yourself from all the sort of raw rahing on one side or the other, if the government can remain divided so that the structure the regulatory is nothing changes, that is probably the best outcome for an equity investor, So a divided government is less likely to produce changes that are going to cause you know, massive changes and flows of capital are going to provide winners to some

companies and losers to other companies, and effectively, you know, what you have priced in in the equity market is then reflecting what the future earnings growth are is currently. So I think as an investor, one thing to hope for is for a divided government and for the government to make the least amount of changes based on what's occurred in the past.

weighted benchmark to outperform a larger cap weighted benchmark. As sort of the earnings growth tends to sort of broaden and move away from these extraordinarily large, extremely profitable, extremely high growth, and reasonably valued companies that make up the you know, the largest companies in the cap weighted index.

And you know, bringing it back to sort of the small cap space, you know, I when I look at you, I qualitatively look at a couple of companies and I find some very small cap company that looks interesting, and it looks like it has some earnings growth, and maybe it doesn't even have any analysts following it, and I'll start to read the ten k's and ten ques, and what will come up is that in the process of doing this, there'll be two activist investors that will come

on board and start sort of looking for changes to occur.

So the level of sort of pouring over the public equity markets has increased, and as a result, I think the efficiency of the market has increased, and sort of the mean reversion due to valuation metrics has decreased, which I think actively helps estimate revision processes, because what you're doing with estimate revision processes when you're implementing them quantitatively is you're looking for not necessarily misvaluation on any one equity.

You're looking in aggregate that your valuation is basically in line with your benchmark. And the more efficient the market is, the more likely that's going to occur, and the less likely there's going to be mean reversion in terms of

sort of valuation metrics. So over time, what I've seen is an increase in efficiency, an increase of interest amongst smaller cap companies, and a less more likely that a company's stock price, no matter that it's capitalization, is going to reflect current earnings outlooks, So as a result, it's more likely conceivably to respond to upward estimate revisions that materialize that have not yet been published.