How dumb do you have to be…

I (Phil) just read an article about Apple. Here’s the last sentence: “Apple has beaten earnings expectations in every quarter but one since March 2013.”

[Note added a week later: on July 31 Apple reported earnings for the fiscal third quarter.  Earnings per share was $2.34 vs. the ‘consensus estimate’ of $2.18, according to Thomson Reuters.]

 

54 thoughts on “How dumb do you have to be…

  1. I’ve noticed a lot of internet/news discussion/advice on investing seems hopelessly naive in one way or another.

    Eg, someone will do an in-depth analysis of the current finances, market situation, etc and say at the end what the company is worth right now, then conclude whether a stock is over/under-priced based off that… That’s great, and there is probably useful info in there, but the conclusions are wrong since (rational) people buy a stock based on what they think will be going on in the future, not right now.

    So, I wouldn’t really expect to take it seriously. I’m not sure who it is that makes decisions based on this stuff.

    • “Eg, someone will do an in-depth analysis of the current finances, market situation, etc and say at the end what the company is worth right now, then conclude whether a stock is over/under-priced based off that… That’s great, and there is probably useful info in there, but the conclusions are wrong since (rational) people buy a stock based on what they think will be going on in the future, not right now.”

      With regard to earnings specifically, the problem isn’t whether a given analysis works but whether it works – i.e., proves useful – for a given stock/market/industry. You can see that in the Chinese market, for example, EPS acts as a stronger predictor of future stock prices than it does in the US market, where guidance and earnings calls can override EPS surprises. Statistics and backtesting can answer, in most cases, which methods of analyses work.

      (Backtesting is great but is biased because a company’s growth is inversely proportional to its market share, i.e., its time listed on the market.)

      Like any sort of predicting, we are merely looking to be correct as much as possible via the law of large numbers. The ideal model of a stock includes as much information as possible, but the law of diminishing returns restricts most analysts and traders to a handful of metrics. If the metrics are correlated to stock price, it’s a poor idea to dismiss them simply because investors are irrational.

      And that’s still ignoring the fact that 60% of trades are made by algorithms, as well as the fact that each stock is held by the general public at a different proportion (relative to insiders/institutions/funds). Trading is as complex as any science, and just as we cannot dismiss a data point in the science of global warming in spite of our personal beliefs, we cannot dismiss market data no matter how cynical we are about analysts – if our goal is profit, that is.

    • … but the conclusions are wrong since (rational) people buy a stock based on what they think will be going on in the future, not right now.

      The analysis of current info is used to make a projection about future earnings and the projected future earnings are used to value the stock. Almost all stock analysis is based on an estimate of future earnings.

      It may be a little hard to see how the current info is used to project future earnings, but it is there. Sometimes they explicitly forecast a future growth rate and use that to estimate future earnings. Sometimes they multiply current earnings by an “earnings multiple” (such as a P/E ratio) to get an estimate of the current value (P = P/E * E). In this case, a higher forecasted growth rate for earnings means a higher P/E ratio.

  2. Dumb because of being selective with the presented data, i.e., choosing only one unit from many?

    Dumb because of not explaining the point of the presented facts?

    Dumb because the last sentence does not read as a conclusion and/or call for discussion?

    Could I not call your blog post here dumb for the same reasons?

  3. The sentence itself isn’t dumb. nor necessarily the article. The idea that analysts would have expectations that have been too low (allowing Apple to beat them) for all of those quarters leads one to believe that the analysts — or at least their consensus — are dumb.

  4. I don’t see the issue with the sentence. Companies have a lot of freedom in how they report their earnings. Creative accounting can shift earnings from one quarter to the next, so instead of exceeding expectations a lot in one quarter and missing them in the next quarter, they can exceed expectations a little in both. They can also be strategic about the information they release about future earning estimates.

    That is to say: the earnings that are reported are themselves a function of the model predictions. It’d be more surprising if Apple’s stock kept jumping after they exceeded their earnings expectations, but that generally doesn’t happen: markets become less “surprised” as earning targets continue to be exceeded and stock prices can drop when the targets are exceeded only by a small margin.

    • “Creative accounting can shift earnings from one quarter to the next, so instead of exceeding expectations a lot in one quarter and missing them in the next quarter, they can exceed expectations a little in both.”

      Sure, but don’t believe that such creative accounting goes unnoticed. Those with skin in the game aren’t just looking at EPS beats and making decisions based on those beats. They are poring over the updated financials, and artificial beats are not rewarded like true surprises.

      “markets become less ‘surprised’ as earning targets continue to be exceeded and stock prices can drop when the targets are exceeded only by a small margin.”

      Not necessarily. You will lose money by holding general statements like these as your trading philosophy. Many trends become self-fulfilling prophecies, often because of algorithms relying on those trends or because investors see a “sure thing.”

      I know it’s popular to hold a cynical view of both the market and predicting the market, but those of us with skin in the game know that only research and experience can confirm or deny cute phrases such as “Sell in May and go away.” I’m vocal about this subject because I’m also cynical. When you’re cynical about investing, you realize most of the common sense about the market – and money in general – hurt those near the bottom of the SES the most: Those with little money let that money decrease in value due to inflation almost certainly; the middle class makes “smart investments” in general market funds, which are not nearly as safe as they seem.

      Automation will eat the jobs of all people eventually, but the process starts at the bottom of the SES totem pole. Trading will be the only lifeboat for those who cannot make money creatively (e.g., Youtube, blogging, art). Those who comment on Andrew’s blog seem to be quite smart (myself excluded) and their words hold weight, but we are still susceptible to Gell-Mann amnesia.

      We should remember that lest we poke holes in someone’s lifeboat.

    • It’d be more surprising if Apple’s stock kept jumping after they exceeded their earnings expectations, but that generally doesn’t happen: markets become less “surprised” as earning targets continue to be exceeded and stock prices can drop when the targets are exceeded only by a small margin.

      +1

      Words are cheap. Money is the real test.

      Apples earnings guidance is cheap talk. Analysts’ forecast are relatively cheap talk too. People are people, and it may be considered polite for analysts to pretend to accept Apple’s guidance at face value and play along so Apple can consistently exceed expectations.

      But, trading stock involves real money. You have to look to price movements after earnings announcements to see if anyone is systematically fooled.

      The same thing happens all the time elsewhere. Average employees regularly “exceed expectations” on performance reviews. The most ordinary actions by children are commonly praised to the heavens. In Lake Woebegone, all the children are above average. We are talking about cheap talk made by humans here.

  5. “The idea that analysts would have expectations that have been too low (allowing Apple to beat them) for all of those quarters leads one to believe that the analysts — or at least their consensus — are dumb.”

    If you’re not privy to a field, those working in the field might appear to be dumb. For example, in light of the common saying that “history is always written by the winners,” we might scoff at a historian calling himself “objective.” Or we might call dieticians dumb when they say that fat doesn’t necessarily make people fat.

    Analysts for many big stocks work in a symbiotic relationship with the company, keeping estimates low so that the company can consistently meet and/or beat estimates. This is not true for all analysts, nor is it true for all stocks. Each stock requires its own investigation as to whether this relationship is running in the background during earnings seasons, and perhaps the “dumb” information here is then useful in that investigation for Apple.

    I run an earnings trading newsletter, so I’m sensitive to this topic. But even if this blog post were about another topic, we are still looking at a man’s Gell-Mann amnesia in action. I hope in the future, Phil’s posts will be more like Andrew’s in that the topic and/or call-out will be congruent with the author’s expertise.

    • Here’s something I don’t get: Why do we, as a society, or as analysts, focus so much on pegging performance against “earnings expectations” which itself is a somewhat artificial & gamifiable benchmark?

      Why don’t we see more news of the nature that compares Apple to peers, or segment average or DOW or its own past growth etc?

      Why focus so much on pegging oneself to a subjective expectation?

      • I think the idea is _supposed_ to be that the price of the stock and the earnings expectations both include all of the information about the fundamentals of the company, so if the earnings beat expectations the stock price should go up, because the fundamentals are better than people knew.

        Of course, in reality this is laughable. Nobody with a lick of sense thinks that “expectations” represent what anyone actually expects; certainly they don’t represent the expectations of the analysts who report what they “expect.”

        So, in answer to your question: I have no idea. Maybe it’s just for entertainment value, like when sports pundits predict who will win tonight’s game.

      • I don’t think anyone is benchmarking companies ‘real’ value (to society?) this way; such financial news is looking at stock price changes in the near term (seconds to a few months) after an announcement. If the earnings information had come
        out more diffusely, or in the complete absence of estimates, we might end up in the same place but in a more boring way (and less tradable for professionals, and less newsworthy.) Same (perhaps) final value; playing the expectations game hasn’t helped your books.

        Fact: large price moves (well above typical daily variation) often occur after company’s earnings announcements (*)
        Model: these happen when the earnings are a surprise, not anticipated by the market as a whole and thus not previous
        baked into the pre-announcement prices.

        Turns out (ton of literature on this) the model works well, but to even evaluate it you need a proxy for the
        unobservable (and not even real) value: “what the market anticipates”. It’s a purely empirical question as to whether
        “extrapolate Apple’s past earnings” or “look at analyst consensus, but fudge it up a bit since by convention they
        are always a bit low” works better in the context of this model. There’s no reason to feel surprised or annoyed if the
        latter does.

        (*) For a certain class of investor (who are generally short to mid term, professional) price movements after earnings
        are their bread and butter. (And for good reason, but I’ve rambled on enough.) So these surprises are real news! Not to mention, there aren’t other common big stories you can honestly tell about day-to-day stock price moves.)

        • When I was working in finance one of the projects I had assigned to me was to parse news stories out of the ticket feed and classify them by what kind of story they were, then look at price movements for a week before and after the announcement, rescaled by predicted risk from an existing model, the idea was to use the stories to alter the risk predictions. I found that the news stories predicted abnormal price movements pretty well… Between one and three days BEFORE the announcement. That study taught me a lot about the industry. Trading on whispered information is widespread and almost completely eliminates any short term advantage you might gain by doing smart things with public info in short term.

        • Did your newsflow include estimate revisions and recommation changes? Sell-side upgrades/downgrades are definitely one of the things driving prices.

        • I don’t think so, hard to remember since it was around 1999 or so, but I remember it included things like changes to the corporate officers, earnings actuals, lawsuits, accounting revisions, basically information from the company itself.

        • Also remember the benchmark was the existing volatility prediction model, I’m not sure what went into that model in full.

          It’s very possible that the pre announcement volatility was caused by analysts who leaked whispers and altered their buy or sell recommendations based on already knowing what was going to be officially announced, thereby moving the price early in the 3-4 days before an announcement. There was almost no category of announcement that retained unexpected volatility on the day after the announcement. Most categories had a pattern of ramping up starting 3-4 days before, peak volatility day of, and back to baseline day after. That was true even when announcements came after close of trading. So day of was still before announcement.

        • 1999 is like forever ago… sell-side analysis has changed quite a lot since then, starting with Reg FD in 2000 (https://en.wikipedia.org/wiki/Regulation_Fair_Disclosure) and the settlement in 2003 (https://en.wikipedia.org/wiki/Global_Analyst_Research_Settlements). I don’t have first hand experience of those ancient times, I’m not sure if analysts are now more or less useful than they use to be.

          It should be noted as well that you’re talking about a very specific part of the money management industry (I would say more on the trading side than on the investment side). Other parts of the industry do care about sell-side reports. Not so much about current-quarter EPS estimates, mind you, but that’s just an infinitesimal part of what they do. Of course one cannot just take what they say at face value, a proper understanding of sell-side research depends on many factors including the specificities of the company being covered and the company doing the coverage.

          For example, micro-caps only get extremely positive ratings because why would the analyst waste time with them otherwise? So either the analyst is really very bullish on the company or (quite often) his job is to be very bullish on the company because his employer is in the lucrative business of raising capital for that company.

        • No doubt 1999-2000 was “ancient times” ;-) And I also realize that I’m primarily talking about company official announcements rather than analysts reports.

          No doubt analyst reports, when taken with a grain of salt, are useful to someone in some way. I assume the original report Phil was discussing was targeted at wide-dissemination, retail investors (otherwise why would Phil even have read it?). My general feeling is that retail investors are the ones *least* likely to benefit from analyst reports. There are several factors:

          1) Retail investors should never invest in individual stocks, the performance of ETFs will generally exceed what they can realistically expect from trading individual stocks, at lower risk.

          2) Retail investors are the least likely to understand the nuances we’ve been discussing here: how firms manage their analysts, how analysts manage their predictions, how predictions change over time… who benefits from what kind of predictions etc

          3) Large investment management firms and trading firms also generally benefit from either factor/industry investment practices using ETFs or market-making practices on individual stocks, where the goal is to collect a short-term premium by providing liquidity, carrying either a positive or a short inventory in exchange for a little bit of profit off every trade *on average*. Here basically the goal is to estimate short term volatility in price and in demand and set a proper bid-ask spread for the range of sizes of trades between say 1 share and several tens of thousands.

          This leaves who to read analyst reports? Active/hedge fund managers, and high net worth individuals who trade on their own account? Those guys are all very savvy about how to read between the lines of the reports, and are not “dumb” enough as Phil says to just believe what they read.

          Unfortunately, my guess is that there are still many individual retail investors who buy and sell Apple or Cisco or Facebook or GE or whatever, they have a portfolio with 5 or 10 individual stocks, and they read these analyst reports, and they buy and sell in part based on analyst reports and reading the newspaper articles about their favorite speculative company. I don’t think that this is a good way for people with less than about $50M net worth to behave, but I also don’t think people should ride motorcycles or smoke cigarettes, and yet they do.

        • “Trading on whispered information is widespread and almost completely eliminates any short term advantage you might gain by doing smart things with public info in short term.”

          But, at least in some circumstances, public info can be useful for medium or long term advantage. I know someone who tracked announcements in Science Magazine about drug recalls and other negative information about particular drugs, then compared that with price swings in drug company stocks, and managed to use that to determine good times to buy and sell stock in drug companies. She made about a million profit before she decided to quit.

        • That kind of thing is possible, particularly for smaller investors and particularly when it comes to issues like the one you’re talking about. For example automotive recalls, or laptops / cellphones that catch fire or whatever. If you have access to information that indicates something is wrong with a company’s product or production line or supply chain or whatever, then you can make money trading on that information so long as you do it before there’s a very obvious implication announced.

          I’m surprised you can make money trading on information in Science Magazine since that’s a pretty widespread information source, but if you have specific knowledge about say biochemistry and can “read into” early announcements that there might be bigger / deeper issues that come out later, I could totally see how that would help, again in the medium term (months to a year). In the short term (less than a week) any public announcement that has an obvious implication good or bad for a given company is typically priced out of the market somewhere between 3 days ahead and maybe 1 day after announcement. More ambiguous stuff where you can make educated guesses and wait for weeks or months until they pan out are probably still workable as your friend’s example proves.

  6. I take the sentence as evidence that Apple is very good at managing expectations, at controlling information flow so they can exceed earnings expectations regularly. This is a goal of corporate communications to the investment community: better for the stock price, better for the publicity, even better for internal morale, to under-promise and over-deliver than to over-promise and under-deliver.

    • Nah, if it were this simple the analysts would have cottoned to it and they would bump their expectations upwards. No way would they let Lucy pull the football away every quarter for five years.

      “Expectations” are deliberately lowballed, you can read these comments to learn why that is done, or at least to learn why some people think that is done.

      Evidently people didn’t realize that the title of the post is sarcastic. The analysts aren’t dumb, they’re lying about what they expect. Although since everybody knows they’re doing it, it’s not really lying, I suppose. I find it funny, but considering how seriously everyone is taking it, maybe it’s just me.

  7. Here is an article containing that sentence; the only other google matches for the entire sentence are this blog post. Have a look at the graph of those dumb estimates.

    https://www.marketwatch.com/story/apple-earnings-most-boring-quarter-of-the-year-still-holds-some-intrigue-2018-07-27?siteid=rss&rss=1

    “Dumb” is a good word for replacing a quantitative measure with its sign.

    By the way, I read that Microsoft beat earnings estimates in 41 of its first 42 quarters.

    • First, the issue here is clearly what ‘expectations’ mean. To a layman it would might sound something like ‘what the well paid experts, trained in the field and financially well-motivated to get it right, actually expect” under which these examples surely do look like idiocy or dishonesty. Perhaps you could explain to all what ‘expectations’ really means in this context?

      I would like to ask: What’s the meaning of your complaint about ‘replacing a quantitative measure with its sign’. My first uncharitable thought (sorry!) was that you might be trying to draw our attention to how “close’ the estimates vs actual have actually been (see graph in story) rather than the sign of the difference. But as a complaint goes that would (if sincere) be nuts, and I feel bad for having suspected it – but, so then, please what did you mean?

      • You could argue that these estimates are more like Bayesian Decision Theory with a utility function *not* of the consumer of the information but rather of the *analyst themselves* driving the decision. in other words, the analyst asks “given what I know about the possible outcomes, which number should I pick to maximize the benefit to me?” given that there are various kinds of costs associated:

        1) If you’re far wrong you’re considered inaccurate
        2) If you are bullish and low you will be “proven right” by a higher result
        3) If you’re bullish and high you will be “proven overenthusiastic” by a lower result
        4,5) the opposite sense of 2 or of 3 for bearish analysts
        6) The company itself will like you to say something in particular, and would like to give you incentives to do that for them.
        7,8,9)… etc etc various other considerations, such as the value to your parent company who would like to make money off volume as market makers, and the value to you in terms of what whoever your next employer might be would like to see you do.

        Given all those considerations, almost none of which are relevant to a retail investor trying to make a decision, the result of the decision theoretical choice is rarely all that informative from the perspective of the people who are in theory following the analyst.

        Quant finance guys spend almost all of their time optimizing full portfolios to improve return on average and reduce variation, typically by looking at things like industry membership, firm size, liquidity and daily volume, daily or short term volatility… They don’t really use much of the kind of information that analysts provide.

        Furthermore, from the retail perspective, it’s the case that retail investors do well (maybe not optimally, but very close) when they choose broad index funds with low overheads. Such index funds need no analysis. So who exactly benefits from listening to and following advice of analysts?

        • Thankfully there are also investors who look at fundamentals (how the company is doing and what’s the outlook, the industry trends, the competition, the economic and political environments) and not just traders looking at the liquidity and the intraday volatility. It’s hard to see how markets would work otherwise. The (ideal) market efficiency is the outcome, not a given.

        • Of course just looking at liquidity and intra-day volatility and soforth would be meaningless unless the market also expressed opinions on fundamentals. But there are also observed facts about historical fundamentals (ie. actual reported earnings) and while historical results are not a guarantee of future results, neither are analyst opinions about future results any guarantee about future results, particularly in the longer-term such as several years to a decade rather than the few weeks until the next earning report. In fact, the historical results when combined with historical results of companies in similar industries, can in general give a reasonable view on the overall performance of the industry as it relates to technological improvement in the cost of production, and changes in demand for products, and the appropriateness of the quantity of resources invested into that kind of good or service.

          If, as is generally acknowledged as a good idea, you are investing in diversified portfolios across multiple companies in multiple industries, then the primary thing you need to be doing is designing your industry balance, and minimizing the volatility of each industry factor while retaining the overall industry performance. The question isn’t “is Apple going to meet its earnings forecast this week?” but rather “are desktop computers declining in demand? Are mobile phones continuing to grow in demand? Are component supply chains in flux or threatened by natural disasters? Is emphasis changing from producers of hardware to distributors of licensed content or walled-garden app stores? How will policy changes regarding corporate taxation affect high-tech companies compared to say automotive companies?” etc

          None of these things need necessarily be specific to Apple, and in fact predict the effect of those things *on Apple* is extremely noisy and unlikely to be accurate, whereas predicting average effects on say demand for desktop computers, or mobile phones could easily be more accurate and lead to improved portfolio balance without ever needing to try to predict the specifics of individual companies.

          For the most part, I don’t think fund managers spend much time looking at the kind of analyst report we’re talking about here, I think they tend to accumulate a lot of data and run regressions on them to try to make their own proprietary predictions, but this may be due to my specific limited background.

        • Today institutional investors are flocking to ETFs which allows them to build the kind of portfolio I’m talking about by trading essentially baskets of stocks all at once. If you want to de-emphasize say retail chains and increase your emphasis on healthcare providers… you sell some retail industry ETF and buy some healthcare industry ETF, you never touch ToysRUs and Merck directly for example. This dramatically lowers your asset specific risks, and dramatically decreases the liquidity cost of achieving the balance you want given the large volumes you will have to transact as a multi-billion dollar fund rebalancing your portfolio.

  8. Phil, I am truly shocked that no one here seems to get your point… So i’ll spell out what I think it is:

    Who would be so stupid as to bother to even quote what the earnings expectations are? The whole analyst/earnings sham is ridiculous. The purpose of analysts is to say what the company wants them to say so that the company can cook their books and consistently report slightly more than what the analyst says. It’s as simple as that, like a corrupt cop’s purpose is to arrest anyone who gets in the way of the local drug cartel’s business.

    Unfortunately Phil, vast swaths of the world treat this stuff as if it were in fact meaningful, and people continue to earn decent salaries reporting this crap.

    • Yep. In any gold rush (i.e. word is out and the rubes are on their way) you’re better off selling maps, claims, picks and pans than going prospecting yourself.

      Information asymmetry is one thing but manufacturing mountains of data to create the illusion of expert knowledge amongst the gullible is unconscionable. Of course nobody gets in trouble because their SEC disclosures, once parsed, will always (unless they want real trouble) show that e.g. the third quarter was looking dicey so unexpected second quarter receipts weren’t booked until the third quarter because *enter reason here*. Most retail investors think they’re sharks, never realizing they’re the bait.

    • Yeah, Wall Street and banks are populated by servile idiots perpetuating a sham. One stupid line in Marketwatch pretty much proves that. Maybe if we weren’t so smart we’d be rich.

      Why not go further? Maybe the entire idea of predicting an uncertain quantity in terms of other uncertain, but easier to measure, quantities is only for chumps.

  9. Phil: Things often appear “dumb” to the ignorant.

    Analysts do not try (nor are they paid) to accurately predict Apple’s earnings.

    Daniel: You write:

    The purpose of analysts is to say what the company wants them to say so that the company can cook their books and consistently report slightly more than what the analyst says.

    Uhh, there is an extensive academic literature on this topic. If this were the whole “purpose” of analysts, then why do such a large percentage of companies report less than the estimates?

    • It’s more complicated than that of course, the truth is more like ecology, analysts are symbiotic organisms who receive benefit from their hosts (the firms that hire them). The firms in turn receive benefit from the secretions of the analysts that act like bait to attract attention to the companies that the analysts analyze, thereby increasing the revenue created by market-making / liquidity providing / trade commissions business activities that the hiring firms carry out. The hiring firms make money whether the analyzed firm goes up or down, but they make money in proportion primarily to the trade volume, which requires secreting that attractant.

      The companies themselves (the ones being analyzed) of course benefit from the liquidity provided by the market making firms, and benefit also from the availability of low interest rate loans as well that come along with a long string of “meeting expectations” and “steady revenue growth” and soforth which means that the banking firms can make more money when their analysts pump out regular bait secretions that have certain properties that give the companies they lend to good lending properties to enhance the lending businesses.

      The analysts know which side of the bread their butter is on and so they do what gets them and keeps them their cushy job.

      As someone who has worked in the quantitative finance industry, I can say that the quants I knew didn’t pay any more attention to analysts than they do “chartists” who talk about the psychological levels of support for the trend lines of short term price graphs. If an analyst’s output is in fact worth more than toilet paper, they become a *fund manager* and stop giving out free advice.

      • Daniel,

        A lot of your words are true. However, I feel that perhaps the finance industry, the gentle beast, might have chewed you up and spit you out with a grudge. Objectivists and pessimists hold similar viewpoints in that they both must reject much from the mainstream; only the pessimist does it out of principle instead of in search of truth.

        Not all analysts want to become a fund manager. I’ve rejected multiple offers myself. I still offer lots of free trading advice and even waive coaching fees for those seeking if they are young/poor/military/etc. I don’t believe the people I’m teaching are “chumps” but individuals seeking financial freedom and brave enough to bear the risks involved in seeking it.

        Not all analysts give into the corporate system, either. I have called companies out on creative accounting. I have rejected invitations to become closer with powerful companies and their management, even when those requests have come directly from the CEO. A couple years ago, I had my first such experience, when I called attention to IBM’s manipulation of earnings numbers:

        https://seekingalpha.com/article/3967854-sell-ibm

        I was almost immediately contacted by IBM, who attempted to “show me which side the butter was on” and use me as a mouthpiece. I called them out on this too:

        https://seekingalpha.com/article/3997717-ibm-want-know

        Like you, I’ve met many questionable people in this industry. It’s tempting to indulge as they do. Or you might simply reject the industry as a whole in favor of cynicism and move onto a kinder industry. The hardest choice, however, is to foster the cynicism in a way to serve others.

        I’m sure you’ll agree that the market offers many opportunities for consistent income, provided you’re not led astray by those with hidden agendas. I think we are both on the same page in regard to analysts in general, but too often does gravity act on both babies and bathwater.

        As an aside, this discussion of analysts reminds me of a psychology study on politicians. The general gist of it was that when people quantified their feelings for politicians as group, the rank was somewhere around 2 on a scale from 1 to 10. However, when you asked those same people to rank a handful of individual politicians, the average rank was somewhere around 7. I couldn’t find a link for this article, but I remember reading it back in grad school. Perhaps someone could find it?

        • As you point out, I’m not calling out a grudge against every single analyst, I’m talking about the function that they play in the ecological environment of the market today.

          As Phil says below, it’s not the analysts that are dumb, its anyone who would believe an analyst who *every single time for the last N years* has come in exactly x% lower than actual earnings…. It’s like someone who can call out which card you’re about to deal except for every 5th one… you gotta know that they’re getting their info from somewhere and just making up that 5th one because they think it’ll make you less suspicious.

        • Also, Damon, your admission that IBM tried to butter your bread and soforth just goes to show that my basic impression isn’t *cynical* at all, it’s realistic. The existence of people such as yourself who don’t succumb isn’t really so important for the truth of my general impression, in any population of people there is variation and some will be more noble than others, but the bulk or usual thing is to get your bread buttered regularly in my opinion. It’s not even like they’re trying to hide it really, lots of analysts talk up how great their company they follow is on a regular basis in ways that sound more like the PR department, or at least these are the ones that get the attention whenever some news report needs to insert some analysts opinion.

      • “It’s as simple as that, like a corrupt cop’s purpose is to arrest anyone who gets in the way of the local drug cartel’s business. ” — Daniel Lakeland, 8:54 pm

        “It’s more complicated than that of course, the truth is more like ecology, analysts are symbiotic organisms…” — Daniel Lakeland, 12:21 am

  10. Many of you are pointing out the obvious fact that the analysts who report “expected” earnings aren’t dumb, they’re dishonest. They say they “expect” $x, knowing full well that the number is much more likely to be a bit higher than a bit lower.

    You don’t have to be dumb to be one of these analysts, but you have to be dumb to believe them!

    • Instead of “dishonest”, one could call that a conservative estimate. Anyway, earnings forecasts are just a small part of what sell-side analysts provide. Sell-side research is definitely not worthless. Investing in companies with no sell-side coverage is harder. But it’s not *that* valuable either, as we can see now that European regulations require brokers to sell their research for hard money instead of bundling it “for free”.

      Regarding estimates, I found the following interesting:

      “We hypothesize that analysts with a bullish stock recommendation have an interest in not being subsequently contradicted by negative firm-specific news. As a result, these analysts report downward-biased earnings forecasts so that the company is less likely to experience a negative earnings surprise. Analogously, analysts with a bearish recommendation report upward biased earnings forecasts so that the firm is less likely to experience a strong positive earnings surprise. Consistent with this notion, we find that stock recommendations significantly and positively predict subsequent earnings surprises, as well as narrow beats versus narrow misses. This predictability is concentrated in situations where the motivation for such behavior is particularly strong. Stock recommendations also predict earnings-announcement-day returns. A long-short portfolio that exploits this predictability earns abnormal returns of 125 basis points per month.”

      https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1571619

    • I don’t really know what to expect about expectations; not my area. With that said, I just took a look at the marketwatch link provided by Dzhaughn above. I noticed two things:

      1. Actual earnings was always greater than expected in that chart.

      2. The MSE of the expectations was something like 2 percent of the variance of the earnings.

      So to me it seems like these expected earnings were quite accurate but slightly conservative. The fact that Apple always exceeds expectations seems at least evidence that the earnings were extremely stable and predictable.

      • I don’t disagree, it just puts me in mind of Humpty Dumpty:

        “When I use a word,” Humpty Dumpty said, in rather a scornful tone, “it means just what I choose it to mean- neither more nor less.” “The question is,” said Alice, “whether you can make words mean so many different things.”

        It also makes me think of Catch-22, or a Who’s-on-first routine.

        “Yet again, earnings were above expectations.”
        “Well, of course, but everyone expected that to happen. It would be shocking if it didn’t.”
        “You’re saying you expected earnings to exceed expectations?”
        “Obviously. They always do!”

        Oh, as long as this is reminding me of things, it may as well remind me of The Princess Bride: “This word, I do not think it means what you think it means.”

        • Again, not my area, but it just occurred to me. Surely, the company itself has some ability to manipulate earnings for each quarter. Being able to say that one has beat expected earnings in itself surely has some utility so if it is close, doesn’t the company have both the incentive and the ability to push the residual positive if it’s magnitude is small enough?

          In that case, it would seem possible that the analysts were *not* biasing their expectations, but the companies were intentionally pushing earnings to just top over that. Given how incredibly close all the expected earnings and real earnings are, this seems much more likely (to someone as inexperienced in this area as me, anyways). This is much like how we see so many pvalues of 0.049; this isn’t a realization of a precisely powered study (in the case of p-values) or extremely precise but biased estimate (in the case of earnings expectations), but rather a realization of the ability to manipulate the target measure.

          With that in mind, “exceeding expectations” perhaps should read as “exercising the ability to beat expectations”…which, in itself, is still a meaningful statement about a company.

        • This is *definitely* done, but since it’s so incredibly common the analysts know that it’s being done, and *if they were trying to give some kind of accurate estimate* would increase their estimates slightly above what they’re currently giving. In turn, the companies would respond by using up more of their control to bump the earnings a little higher, analysts would respond by bumping their estimates… company responds… analysts respond… this would repeat until almost all of the “creative accounting capacity” was used up, and which point analysts estimates would spread both above and below the actual. To the extent that analysts forced this to occur they would be providing a good function in the market: making more information transparent. since to the extent that companies are creatively booking things it shows that historical real earnings and essentially guaranteed future earnings are un-accounted-for in current reported earnings, because it’s those things that are able to be moved around in the books to give the company some control.

          But, this extent or depth of control is not well represented in the public information either from analyst or company reports. Of course, people trading on their own money work harder to estimate these things, but an “infowar” between the analysts and the company would in fact reveal more information about the company. The fact that it doesn’t happen is in large part down to the fact that analysts make money off being wrong in a way that helps their employer (brokerage firms, banks, etc) and the company they follow. Analysts who got paid based on how accurate or unbiased they are would make different choices.

          I have a friend whose father has been a stockbroker/money manager for decades. He recently got fed up and quit his job with a large retail investment firm and opened his own fee-for-service company with a few partners. He said it was the best thing he’d ever done, because even though he would try to give good advice, the corporate demands and pressures would leave him with limited tools and requirements to push relatively less good investments with high commissions for the company etc. He realized the extent to which there was no way to keep his retail job without bowing to the pressures from above to some extent, but he could only realize the extent of it *after he had left* and was free to recommend anything at all to his clients.

          And that’s the biggest issue in finance, it’s not that individuals are somehow “bad actors” necessarily, they may well want to help their clients as much as they can. But survivorship bias is a real thing. You can’t help people invest unless you have a job as an investment manager, and you can’t keep that job as an investment manager at a big firm unless you are meeting quotas, pushing the products that your boss tells you to, and soforth. I suspect it must be the same for analysts, in the sense that in general your paycheck comes from your boss not from your “clients” (now Carlos Ungil elsewhere says that this has changed in Europe recently, and in fact the prices charged for analyst info is not that high, indicating it doesn’t provide much value).

          You may be able to mitigate *some* of the damage those issues do, but if you mitigate too much of it, you lose your job, so those who are still in the industry are doing at least some of what the boss wants to the detriment of clients. You can still argue that the clients might be doing better than they would be if they just did stuff on their own… and this might help you sleep at night, but the existence of Vanguard and iShares and things suggests otherwise. Anyone could be told “put all your money in a whole-market index and wait for 30 years” but I suppose not everyone will follow this advice ;-)

          It’s been known for decades that if you take a random selection of fund managers and compare their whole career to “throwing darts at the WSJ” (today you’d just make a list of the top 5000 companies and randomly select 100 from them uniformly with a computer RNG) you will do better on average randomly selecting stocks than randomly selecting fund managers. It actually turns out according to a friend of mine still in the industry that a uniformly weighted random selection of companies actually does better than a capitalization weighted selection like the Russel 3000. This actually suggests that capitalizations tend to be over-inflated at least over the last 30 years or so.

          Since fund managers can do this random selection of stocks easily, it suggests that *on average* fund managers harm performance.

          Those who don’t harm performance may be simply getting lucky since there’s a bunch of variation that would just naturally occur if you *did* randomly select stocks.

          All of this is to say that the stock market is a complex feedback system with many many many forces working on each little corner of the market, only some of those forces are real fundamentals, and many of the forces are instead second-order other stuff entirely.

        • Something that might not be obvious for outsiders is that the earnings that are 1% above today’s consensus may also be 10% below the expectations as they were just a few weeks ago.

          In addition to “earnings management” (shifting revenues and costs from one quarter to the next to smooth profits, see the chart at http://statmodeling.stat.columbia.edu/2012/10/08/ethical-standards-in-different-data-communities/ for an extreme example) the can do “expectations management”. Sometimes companies will more or less subtly let analysts know that their expectations are too high (the drastic version is issuing a profit warning). Estimates will come down, and the stock will sufffer, but at the earnings call they may beat the degraded expectations an everyone will be happy.

          In other cases everyone knows that the estimates are too high, and sell-side analysts will tell you so, but they don’t really know how bad the results are actually going to be so they just wait until the earnings release to reset their expectations going forward.

  11. By the way, The New York Times has a spectacularly well written article about Goop that may or may not be relevant to the theme here. When I was first reading newspapers in the 1970s they always had horoscopes. TV had plenty of articles about EST and power pyramids and alien abductions. My Dad had given me a book by Bertrand Russell that made a skeptic of me (“On Liberty” was the first book he’d asked me to read). He was pretty sure that by the time I grew up people would have outgrown superstition and nonsense. Long after growing up (I think/hope) I must report, alas, that he was wrong. But here’s a great take-down and a cautionary tale of how certain attacks only make the quacks stronger: https://www.nytimes.com/2018/07/25/magazine/big-business-gwyneth-paltrow-wellness.html

Leave a Reply to Jackson Cancel reply

Your email address will not be published. Required fields are marked *