Categories

Who is being reasonable now?

Last week Horace wrote about the apparent “reasonableness” of analyst Apple estimates. He explained how the consensus for Apple’s growth was always deeply pessimistic because its performance could be argued to be anomalous. It was just too good to be true. We reproduce the chart here:

The estimates look like characteristic “tell-tales” of a company running strong into the wind.

This conservatism in the face of rapid growth sounds “reasonable” but is it always practiced? And what about the ability of this conservative strategy to predict dramatic changes in growth? To test, we started to look at the predictions for RIM. RIM has also enjoyed strong growth over a similar time frame as Apple. How did analysts predict its performance? The following chart was prepared using the same technique as the one for Apple[1].

Surprisingly, the “tell-tales” look very different.[2] For every year we looked at the first year estimates were over-optimistic with a decline expected after one or two years. This pattern is remarkably consistent within the RIM data and remarkably inconsistent when compared to Apple.

It’s as if there was nothing in common with the methods used. This is perhaps appropriate, but the “quality” of predictions are not much better. With RIM they were too optimistic in the short term and too pessimistic in the medium term. With Apple they were far too pessimistic consistently.

More worrisome, the 2010 forecast did not offer any hint of the dramatic decline which followed only a few months later. And after three quarters of really bad news, the consensus is still higher than what will likely happen (ensuring an uninterrupted string “misses”).

The study of these two companies’ forecasts is a study in contrasts. We can say that the forecasting is disturbingly consistently inaccurate but we can also say that there seems to have been a bias of pessimism for Apple and optimism for RIM. We can only guess at the cause, but the consistency of bias points to something institutional.

Notes:

  1. The three year forecast consensus (sourced from Bloomberg) was sampled one month following the end of each fiscal year. This should offer the analyst enough time digest the year passed and deeply contemplate the future.
  2. If we may stretch the analogy, these are indicative of an imminent stall.

 

  • Guest

    I think this is some type of “reversion to the mean” thinking. When things are going extremely well, people conservatively expect the rate of growth to moderate to something more “normal”. When things are going extremely badly, people tend to think the rate of decline will moderate to something more “normal”. 

    This seems like a general unwillingness by wall street to “stick ones neck out” and predict something to be either an extra-normal success or failure. It is much easier to remain employed on wall street by consistently predicting outcomes that will be predictably wrong, than to take a stand (“make a big call”) and risk being totally wrong.

    In short, being wrong when everyone else is wrong, is safer than being the only one who is wrong. Herd mentality…

    • http://twitter.com/pberry Patrick Berry

      But that wouldn’t explain the almost diametric approaches of analysts to RIM and Apple in growth phases.

      • http://www.asymco.com Horace Dediu

        Right. Both RIM and Apple were growing at similar rates.

    • http://www.facebook.com/profile.php?id=741717344 Dick Applebaum

      I think you’ve hit upon the real issue… today’s analysts aren’t analysts any more.  

      They are merely sycophants gathering data points and providing predictions to support the opinions of their bosses or their clients.  One of the biggest providers of positive analysis of RIMM, over the last few years, is a firm that has a vested interest in RIMM’s success.

      It’s a case of “tell ‘em what they wanna’ hear”.

      IMO, what is needed is more of an explorer/detective who is seeks truth and understanding — rather than to justify a predetermined opinion.

      • Kankerot

        Todays analysts? Analysts have always been little more than cheerleaders – why would you want to spend time researching only for you to then provide the conclusion for free? Their estimates, predictions etc are there to help their brokerages or portfolios – we want to short stock X lets put a sell or neutral rating out etc. 

      • http://www.asymco.com Horace Dediu

        Sell side analysts work for sellers of financial services. They are not working for those who manage funds directly. Analysts working for money managers are called buy side analysts and their work is not shared or published. Some sell side analysts actually survey buy side opinions to see whether there is any difference of opinion. I have not seen any study of how the two cohorts look at the same data, but it would be interesting. Already we are seeing distinct differences between “independent” analysts (aka bloggers) and the sell side. Buy siders would be a great control group.

  • http://wmilliken.livejournal.com/ Walter Milliken

    Interesting result — I’m wondering if this is a bias due to business model: RIMM was generally seen to have a “lock in” with corporate customers, while Apple seems to be widely seen as being subject to the whims of consumers (and potential low-price competition, which wasn’t generally a concern with RIMM, as I recall).

    A good test of this possible business-model bias might be to do the same comparison with Microsoft, which has a similar corporate lock-in perception to RIMM’s. I’m not sure what the Apple-equivalent consumer-oriented model company should be, though…

    • Tim F.

      I’m less interested in looking at a business-model analog to rate analysts in general… It’s probable that we can observe poor analyst performance. But what I want to look at is what data are analysts biased towards (and against) that can favor RIM (which looked to have a poor strategy) while discounting Apple (which looked to have a strong strategy) in this specific mobile revolution. That’s what I think Horace is doing.

  • Canucker

    Apple should start worrying when the regular analysts start being optimistic about the company…..  RIM was a disaster in slow motion and several people saw this, but not the analysts. Many of the signs were there and the only consistency in their performance was the sheer crappiness of their excuses for bad results. Perhaps the analysts were unable to think outside of their personal experience? They all have BlackBerries so the company must be doing really well. Let’s hope they don’t all now jump to iPhones.

  • r.d

    Horace, 
    It is interesting visualization but does it tell you anything.How about for MSFT, GOOG, IBM, ORCL.Trademark this chart and create you own index “Horace Index” and sell it to wall street.

  • http://kaizenity.blogspot.com/ FalKirk

    “A good test of this possible business-model bias might be to do the same comparison with Microsoft…”-Walter Milliken

    I was thinking that too. Then I started to think about a company that might mimic Apple’s results – a company that was viewed as a series of one hit wonders. The only example that comes readily to mind is a movie studio like Pixar but I’m sure this august body will be able to provide me/us with many better examples.

    • http://www.asymco.com Horace Dediu

      The reason I am not excited about Microsoft is that the growth has been modest. I wanted to compare two companies which were growing in similar fashion. Remember the premise of the original “being reasonable” argument: that Apple was growing way too fast to be sustainable. So it’s reasonable to assume the slowdown is imminent. In the case of Microsoft nobody would suggest that its growth was unforeseeable. Microsoft might make a good comparison with Google.

      • Anonymous

        What about Amazon?  I’m very curious about the analyst forecasts that lead to recommend buying an established business at 100x earnings.

      • http://wmilliken.livejournal.com/ Walter Milliken

        Microsoft in the early 90s, though? Weren’t the growth rates high back then? Current MS, I agree, is a bad comparison.

      • http://www.asymco.com Horace Dediu

        Yes, in the 90s, the story would have been similar but we can’t get analyst forecasts from back then. That data would require serious financial archaeology.

    • Anonymous

      Jobs modeled Apple after Polaroid and The Beatles, which were also series of one-hit wonders.

  • Pingback: Thursday links: holiday stress | Abnormal Returns

  • Anonymous

    Perhaps this is an extreme case of “Loss Aversion”, i.e, psychologically, a loss has greater impact than an equivalent reward.

    http://en.wikipedia.org/wiki/Loss_aversion

    I read an article recently that I will have to track down and link

  • Anonymous

    What’s fascinating to me is looking at what’s coming up in mid-January. Several posts on AppleInsider by Andy Zaky have been predicting a near doubling of EPS over last year. That could push EPS to $33-34/share (and compress the P/E to new lows).

  • http://twitter.com/Marcos_El_Malo Marcos_El_Malo

    One difference in the companies is that Apple suffered a great crisis in the 90s that nearly bankrupted the company, before it was acquired by NeXT* and turned around by Steve Jobs. The crisis at RIM only now is coming to a head.

    It should also be pointed out that Apple’s demise had long been predicted and watched as it unfolded. RIM’s demise seems to be taking analysts by surprise. Another difference. In RIM’s case, it’s the founders that are driving it off the cliff. In Apple’s case, it was one of the founders that returned to save it at the last moment.

    *Acquired for a price of (negative) $429 million. Just kidding . . . . or am I?

    • El Aura

      There certainly is a difference if a company has had a big crisis already or not (Apple had, RIM had not) and whether the founders have reigned uninterrupted.
      But the chart for Apple starts in 2002, and the plotted predictions start in 2005, five and  nine years after Jobs returned to Apple. It is not as if Jobs and Apple was an unproven figure in 2005.

    • Tim F.


      *Acquired for a price of (negative) $429 million. Just kidding . . . . or am I?”

      A few people have asked where Steve’s billions went, how are they being used. His wealth is in the companies. (Job cashed out of Apple but poured most of it into Pixar and NeXT, which was, in part, put back into Apple (a thousand fold) and barely taken out… I’d wager it’s still mostly AAPL and DIS.) The question is: what will the Jobs family (and Jobs’s will) do with the Apple and Disney/Pixar equity that Jobs amassed?

  • Anonymous

    I would guess that company guidance has a lot to do with the discrepancy.  Apple consistently guides conservatively.  They have the under promise / over deliver school of thought.  RIM somehow seems to be just as surprised as the analysts each quarter when performance falls below expectations.  Most recently, the company guided lower 2 weeks before earnings and STILL missed.  I know this has been a recent trend, but I don’t know the longer history.

    My hunch is that analysts in both cases treated the company guidance as conservative, tacked on a few percentage points of 1 year upside, threw out a wild guess for years 2 and 3, then called it a day.  They assume the companies know more than they publicly release, and that the companies have a lot to lose by missing guidance by too much.  

  • Anonymous

    First, this predictive model simply cannot explain disruptive technologies.  Apple is in a growth spurt simply because they keep coming out with distruptive technologies.  RIM is suffering from the effects of disruptive technologies and can’t get out of it.  All the graphs look like the models being applied are simple S curves, because it’s the only mathematical model analysts have that remotely make sense.  But all this tells you is that S curves can’t predict things when disruptive technologies are still in force.

    Secondly, the bias is simply against wildly optimistic or views in general.  In both cases, a model is applied to show how things “tail off”, either positively or negatively.  They’d rather show a modest increase and see Apple beat that.  Look at last quarter, Wall Street sold off Apple stock when Apple beat it’s own expectations, but many wall street analysts said it’s earnings would be even better.  That’s not good for Wall Street to be predicting high and have those expectations not met.  As for RIM, they have to hope that things will tail off, but no mathematical model is going to show where things bottom out, both because no one investing in RIM wants them to continue dropping so far so fast, and because no one really knows what’s going to happen next.

    • http://twitter.com/WaltFrench Walt French

      I agree with you that extrapolation is a lousy way to understand disruption, so I much more enjoy Horace’s insights to the structure and business models as a way to make short-term guesses. (Maybe it helps that I don’t really care what the short-term EPS outlook is, but wonder more about who the big movers are in technology, and what is their ability to disrupt in the future.)

      But I’ll part ways with “S curves can’t predict things when disruptive technologies are still in force.” S curves are only about the evolution of disruptive technologies and their becoming lodged in the mainstream.

  • Ian Ollmann

    A: Apple is disruptive.
        RIM is sustaining.

  • Anonymous

    RIM wears ties. Apple does not.

  • http://www.noisetech-software.com/Home.html Steven Noyes

    I think this goes back to what I said in Horace’s post.  We know what is around us.  RIM was, and still is, very heavily invested in corporate America.  Analysts are ver much part of that corporate culture and mindset.  They all use RIM.  All of their friends use RIM.  They are the original Crackberry addicts.

    So with Apple products, they don’t use them and they know none of their friends (all financial types) don’t use them.  Apple is doomed.

    With RIM, they use them, many of their friends use them.  RIM is going to be OK (and actually GROW GROW GROW like mad soon).

  • MOD

    RIM is a great choice for comparison. I’ve been reading the WSJ over the past few months, and did not understand the constant anguish over RIM. Another well-known analyst “throws in the towel” was the weekly anguished cry for many weeks. Just how many analysts did RIM have working for it? Because the string of anguish did not seem to end (and still doesn’t).

    What was the big deal? So a new phone comes out and it is better, easier to use, etc. Just invest in the new company.

    It is not so simple, if you are deeply invested in RIM, and your career is about following RIM. The psychology of discarding 10 years worth of work is a big stumbling block.

    And because RIM was so dominant in the business setting, many Wall St. firms were (and probably still are) invested in it.

    This is a great analysis Dirk:
    “the consistency of bias points to something institutional”

    Well, the institutions are the investment banks and brokerages, and they exist to make money, so any bias that might exist would be towards making money.

    You should align the graphs by date (and slope). I think it points to Wall St. favoring the successful incumbents (whose stock they already own), and discouraging newcomers (whose stock they don’t yet own) which would cut into the market share of the incumbents.

    RIM was a successful phone maker long before Apple became one too.

    One could argue that Apple is now large and successful too, but only since 2005. It is nothing compared with the long-term (incumbent) successes of IBM, Microsoft, Oracle, Intel, even Yahoo and Google…

  • Anonymous

    Hi Horace,
    Help me understand your rationale for charting growth patterns arithmetically. I see all dynamic systems and processes as increases or decreases in percentage change. Thus, to me lines tracking dynamic systems can only be accurately displayed on a logarithmic scale.
    Please advise,
    Bert

    • jawbroken

      I don’t know what “accurately displayed” means in this context. The test from my perspective is to ask how a logarithmic scale would help better demonstrate the point in the surrounding text.

      As I see it, the current graphs show an absolute error in EPS prediction whereas a logarithmic scale would demonstrate something like a proportional error in EPS prediction. I’m not sure of the precise reasons to prefer one over the other in this case.

      • Anonymous

        Thanks for your response, makes sense. As I look at the Apple chart of the misses, I want to see the progression of the accuracy and visualize a trend, so the proportional changes are what interest me. On the RIMM chart, this is n/a as there is no sequential trend. 

    • http://www.asymco.com Horace Dediu

      I’ve done both percent change and log scales in my posts. See: http://www.asymco.com/2011/12/16/on-being-reasonable/ (second chart) and http://www.asymco.com/2011/12/15/why-apple-is-cheap/ for rate of change.

      The method you choose depends on many factors. There is no single right way to represent complex data/systems.

      • Anonymous

        Thanks, I’m sure I’m oversimplifying as suggested. I did miss the log chart on the earnings misses. To my eye, that chart is fantastic and illustrates the point beautifully. Observing the rate of divergence, and thus the amount of correction required, tells me the most. Especially in regard to the future price of Apple as understanding normalizes. All the best, Bert

  • Tim F.

    And yet Michael Mace called it nearly perfectly:

    http://mobileopportunity.blogspot.com/2010/10/whats-really-wrong-with-blackberry-and.html

    Is it possible to attempt to plot the data he was looking at for the various mobile platforms to estimate when each may be hitting the “top” of their logistical curves?

    • Tim F.

      RIM has direct subscribers, unlike other competitors in the mobile platform market. Substitute subscribers with users (as actually stated as active individual users of a platform or an estimate of devices/services still in use by other means — your “First to a billion” posts) for iOS, Android, RIMM, Windows Phone, and Amazon. Then calculate net new users, net new users per devices sold, device gross margin (you already provide that), revenue or earnings per device sold (you may already provide that), etc… Stack this data, and it’s something like an AMP Index for platform health.

      This would have to be an aggregate total for leading (or even include “lower end”) Android manufacturers since there is not a direct connect between the platform and the manufacturers, and the numbers are all the murkier for the small minority share of Windows Phone (but soon we may be able to largely use Nokia as an analog? Do you measure userbase by XBL accounts? Zune subscriptions?), but it would be interesting to try to see if we can see Apple and others approaching the peak of the S-curve.

      Apple provides sales, sometimes “activations”, sometimes userbase (with other means to estimate), AppleIDs, etc…; Horace has analyzed most of Mace’s data visualizations, but I don’t think we’ve seen net new iOS users per iOS device sales. Even if Android and other competition were a very rough estimate, the data may be illuminating. 

      I actually have confidence in Apple preserving gross margin, revenue, earnings… but I would feel much more confident with a strong net new user per sales ratio, relative to the competition.

      Horace, I don’t mean to put the work on you. I’ll try to make the time to see what I can estimate myself. Maybe. Happy Holidays, all.

  • gerry.croce

    If you look at both charts, around the 2006 to 2009 period, it seems as though the analyst thought that the mobile market would grow at a certain rate, but they allocated this market to RIM instead of Apple.  They might have been reasonably accurate as to the projected market growth, they just bet on the wrong horse.  Which makes sense since Apple was the dark horse in this race.

  • Anonymous

    What strikes me is that the analysts’ forecasts do a reasonable job approximating RIMM’s trajectory (up until this year), but are nowhere close to AAPL’s. It’s almost like the analysts could understand RIMM’s business but not AAPL’s.

    It’s also striking that there seems to be a mental block of AAPL earning more than $10/share (and, of course, RIMM’s EPS was always below $10)

  • Tim F.

    The analysts can’t see the disruption, but they were able to start to see the peak for RIM between ’08-’10 (even if a year or more out). Conversely, it took until 2011 for the analysts to start seeing that Apple will not have near-term negative growth after predicting flat or lost growth for years.

    They may be horribly wrong, but I’d wager analysts start calling Apple’s peak well before it is in sight. Will they start predicting it in 2012, or will Apple’s results tell them otherwise?

    • Tim F.

      Does this mean it’s safe to predict 2012 earnings north of $45 per share “based on analysts’s prediction”? I just blew my own mind!

    • nikolaihoffn

      “They may be horribly wrong, but I’d wager analysts start calling Apple’s peak well before it is in sight. Will they start predicting it in 2012″ (probably)”..or will Apple’s results tell them otherwise?” (probably not)

  • Th3uglytruth

    The explanation is simple…
    RIMM “spins” the truth; AAPL tells the Ugly Truth.

  • SamLowry

    Horace, Dirk,
    where do you get your THREE year estimates from?
    Thanks!

    • http://www.asymco.com Horace Dediu

      Various financial data publishers survey analyst forecasts. The three year forecasts we used were available via Bloomberg.

  • Chandra2

    Do analysts’ opinions matter in terms of how the stock actually trades? I hear a lot of people quoting analysts’ numbers but they are usually sell side financial services people. They do not generate the day to day volume.  In fact, the charts ofApple and RIMM show that Analysts predictions of future earnings are not a predictor of future stock prices. ( The current P/E compression of many tech companies have not much to do with these sell side analysts’ predictions )

    • MOD

      As these analysts are widely quoted in the financial media, yes it does matter. It is part of the propaganda to either buy or sell stocks.

      Day traders may not care, but they do not have a long term interest in the stocks, thus not likely to have a long term effect. Ie, if you buy and sell in one day and sell-short then buy to cover the following day, that negates the prior day.

      The sell side of financial services sell stocks. If one stock is better rated than another stock, it is likely to be promoted more and therefore bought more.

      I echo one commenter here who asked how can Amazon be justified at 90+ PE by these analysts? Per WSJ, AMZN has 18 Buy, 5 Overweight, 16 Hold and only 1 Sell ratings.

  • mysterio

    The top graph would show just how systemic their forecast error is if it were plotted on a log y-axis.  Do you make this data available on a google spreadsheet?

  • Anonymous

    The article on AAPL and RIMM, and the comments here, point to large questions as to the nature of the analyst community. From the charts, it appears that there are two communities with little or nothing in common. Probably the Canadian stock market analysts are mostly interested in Canadian stocks, and of course RIMM is a big player in that domain. There is in effect a large home team rooting for RIMM, and of course the Canadian government is a Blackberry community (productivity is not an objective). So there’s the loyalty factor that influences or has influenced the nice things that have been said over the years about RIMM.
    As for Apple, as the profits and sales have increased, the P/E ratio has dropped, so it’s becoming a safer and safer investment compared to outfits like Amazon. Apple’s past performance has very little to do with present and future performance, so projections based on the past are not helpful. But that’s what analysts do, and for that matter so do a lot of accountants, who should instead pay attention to things like future plans, such as untapped markets, sales constrained by production; on the one hand unlimited sales prospects, versus limits to growth (natural disasters, parts shortages, production capacity). 
    The lesson for followers of analysts is that independence and objectivity are not their bag.

  • Anonymous

    The article on AAPL and RIMM, and the comments here, point to large questions as to the nature of the analyst community. From the charts, it appears that there are two communities with little or nothing in common. Probably the Canadian stock market analysts are mostly interested in Canadian stocks, and of course RIMM is a big player in that domain. There is in effect a large home team rooting for RIMM, and of course the Canadian government is a Blackberry community (productivity is not an objective). So there’s the loyalty factor that influences or has influenced the nice things that have been said over the years about RIMM.
    As for Apple, as the profits and sales have increased, the P/E ratio has dropped, so it’s becoming a safer and safer investment compared to outfits like Amazon. Apple’s past performance has very little to do with present and future performance, so projections based on the past are not helpful. But that’s what analysts do, and for that matter so do a lot of accountants, who should instead pay attention to things like future plans, such as untapped markets, sales constrained by production; on the one hand unlimited sales prospects, versus limits to growth (natural disasters, parts shortages, production capacity). 
    The lesson for followers of analysts is that independence and objectivity are not their bag.