Why relying on broker forecasts can only end in disappointment
In the final stages of campaigning for the EU referendum, the outcome was clearly going to be close. One group with a particularly keen interest in the result were the pollsters and bookies who played a role in forecasting it. On past form, the pollsters had a lot to prove after failing to predict the 2015 general election. But they ended up getting it wrong again by mistakenly forecasting a slim victory for Remain.
For most of us, the signals from political polling and the betting markets don’t necessarily matter that much. But in the days ahead of the EU vote, the value of sterling leapt and many European markets saw gains. It suggested that traders believed what they were seeing in the forecasts. When that conviction turned out to be misplaced, those markets were hammered.
It’s an unusual case, but the forecasting errors in the referendum echo a wider problem with forecasting generally, especially in the stock market. Not only do analyst forecasts often turn out to be wrong, but there’s a worrying inclination among investors to rely on them anyway.
The problem with forecasts
In Robbie Burns’ book, Trade Like a Shark, he makes the following observation about the usefulness of broker notes: “If brokers really knew what they were talking about do you think they would (a) continue commuting into the office every day, or (b) take their laptop to Barbados and stay there? (P.S. it’s ‘b’).”
That view neatly sums up a lot of the work that has been done into the reliability of broker forecasts. Among the chief criticisms are that analysts are prone to being overconfident, which can lead to over-optimistic forecasts. In addition, the whole structure of the investment research industry has been questioned for compromising the independence of analysts. Arguably, some feel they need to be bullish in their forecasts in order to keep corporate brokerage clients happy, and ultimately keep their jobs.
In terms of analysts’ success rates, some of the most interesting studies have been done by David Dreman. He’s a well known contrarian asset manager whose strategies are based on stock market psychology.
Dreman studied a sample of more than 800,000 individual analysts’ estimates in the US market between 1973 and 1991 and found that the average error rate was 40% annually. Not only was that a terrible record, but those analysts were just as inaccurate right across the the business cycle. More concerning was they they also had a strong optimistic bias in their earnings forecasts.
As Dreman explained, the high level of forecast errors combined with analysts’ optimism result in “a high probability of disaster”.
In The Little Book of Behavioural Investing, James Montier reinforces this point, noting that when an analyst first makes a forecast for a company’s earnings two years prior to the event, they are on average wrong by 94%. And even at 12 months out, they are wrong by around 45%. Montier is even more damning that Dreman in his assessment: “To put it mildly, analysts don’t have a clue about future earnings.”
Why does anyone listen?
Given the known fallibility of analysts’ research, the question is why does anyone listen to them? At least part of the answer to that could be down to human instinct.
Montier says that he doubts that we actually need forecasts to invest. Rather, he suggests that the clamour for forecasts is partly driven by the natural tendency for investors to cling to a number once they’ve been given it, even if it’s subconsciously. In behavioural finance, this is known as anchoring.
Another possible reason why we’re hooked on forecasts is that they let us to stop thinking. Sounds crazy? In his book, The Laws of Wealth, psychologist and behavioural finance expert Daniel Crosby, says that our brains yearn for forecasts in a very specific way.
He points out that of all the metabolic demands made on your body, the brain is the most energy sapping and accounts for as much as 20% of the calories you take in. As an efficient machine, the body it is always looking for ways to conserve energy and “nothing provides as big a return as slowing down the brain”.
Crosby notes MRI studies which show that listening to financial experts actually shuts off some brain functions. He says: “To be as concrete as possible: when you start listening to Jim Cramer, you stop thinking.”
Take back control
The message from a lot of studies into the usefulness of analysts’ forecasts is that they need careful handling. At the very least it’s worth being armed with the knowledge that this kind of research can be highly inaccurate. Even near-term earnings forecasts can be wildly wrong. But just as important is the evidence that broker notes offer mental shortcuts that play on risky behaviour like anchoring. Or worse, they offer the possibility of completely switching off and relying on somebody else’s potentially over-optimistic work - and that could be costly. So caution is needed.
A better way might be to forget about trying to guess what the future might hold. As Crosby states in his book: “The middle ground must scrupulously avoid conjecture about the future, rely on systems rather than biased human judgment and be diversified enough to show appropriate humility.” This quote neatly summarises the philosophy behind the stock ranking and portfolio analysis tools we’ve built at Stockopedia.com. Rather than trying to outguess the market, it’s perhaps more effective to build an investment process on simple, repeatable, time-proven indicators that have been proven consistently to work.
Disclaimer
This is not financial advice. Our content is intended to be used and must be used for information and education purposes only. Please read our disclaimer and terms and conditions to understand our obligations.
26 comments
Further to the conversation above regarding broker level estimates as opposed to consensus estimates... does anyone know a service providing broker level estimates on UK listed companies? I don't mind paying a reasonable fee. Many thanks.
"Another possible reason why we’re hooked on forecasts is that they let us to stop thinking. Sounds crazy? In his book, The Laws of Wealth, psychologist and behavioural finance expert Daniel Crosby, says that our brains yearn for forecasts in a very specific way.
"He points out that of all the metabolic demands made on your body, the brain is the most energy sapping and accounts for as much as 20% of the calories you take in. As an efficient machine, the body it is always looking for ways to conserve energy and “nothing provides as big a return as slowing down the brain”."
Thank you Ben for this thought provoking article.
I do disagree with Daniel Crosby though, for he is not considering the ego mind set brain.
The brain does indeed use far more nutritional intake as well as our daily available energy for operating, than other parts of the human body
But the last thing it intends to do is slow down.
For the vast majority of humanity is mentally driven by the ego mind set.
In my experience when faced with nothing to do or no problem to solve, the ego mind set brain cogitates or ruminates on the past.
The ego mind set controlled brain is also fearful of the future which in some ways, may explain why the market hates uncertainty.
The cogitating or ruminating aspect of the ego mind set brain may also lead to the monkey mindset which has been talked about much hereabouts.
I have inserted a link by Eckhart Tolle which explains far more about the ego if anyone is interested.
http://www.theworkbook.org/egoic.htm
Just a quick comment - in my experience, brokers just lag reality.
So, as the economy enters a downturn, then broker earnings forecasts tend to be way too optimistic.
Subsequently, as the economy craps out, and corporate profits collapse, then broker forecasts lag behind this process, as they remain way too optimistic, and have to keep slashing their forecasts.
At this stage you get a good opportunity. Brokers are now bombed out, and if they still have a job, are terrified of being made to look fools by being too optimistic about earnings. So they're now too gloomy. This is when you can make shed loads of money - brokers will now be wildly PESSIMISTIC.
They forget operational gearing. So they put +10% on turnover, and +10% on profits. Whereas in reality turnover will rise 20%, and profits will be up 50%!
Over the cycle, profits could rise 5-10 fold.
Useful post. I have read lots of comments about the self interest component of Broker valuations. I am also concerned about the actual number and quality of analysts in a brokerage. I get the impression there are minimal analysts and maximum salespeople in brokers, and what is worse infrequent analysis.
However my question is how persuasive are these comments and valuations. Given that most managers of organisations have poor understanding of what the actual company performance is going to be 3, 6 or 12 months ahead. Most corporate planning I have been involved at Plc level has involved a healthy portion of finger in the air plus a large dollop of contingent cash put aside. This is without the unforeseen, poorly recognised, poorly prepared for significant events.
Good post but personally I think if anything there is a negative correlation. In particular small cap brokers as we know are all hired by the firms they do the reports for and are all fed the same guff by the firms they represent. To think they are in anyway unbiased or even useful for that matter is naive at best. They are total cr*p on the whole.
The only use I can see perhaps is if a very bullish one comes out and you are looking to exit a long position it provides an opportunity as mugs buy in.
I know a single instance doesn't prove the rule but look at the most recent broker note on Avanti Comms (AVN). The company is quite clearly bust it needs cash and is drowning in its own debt yet they come out with a buy an and a price target more than double the current price. It's laughable.
Log
This is a good & pertinent article, with a timely warning for investors. However, I feel that there is some confusion in the comments.
Posters seem to be confusing "price targets" and "buy/sell" recommendations with earnings forecasts. I pay little attention to price targets and recommendations, which are generally driven by the broker's marketing objectives. As Ben & Ed say (based on sound academic studies), even earnings forecasts tend not to be particularly reliable and avoiding overreliance and "anchoring" are important.
However, company trading updates frequently refer to "expectations", meaning precisely those earnings forecasts, so knowing what they are is rather important to interpreting trading updates. Moreover, my own style of investing, which has worked for me over the years, is based on judging the value of a business vs the price Mr Market is offering it (i.e. the same funadamental investing principle Graham and Buffett use). The only way it is possible to do that is by forming a judgement on future earnings potential. I strongly agree, however, that one needs to be humble - none of us are Buffett or Munger - and not be overconfident in one's ability to predict the future. In my eyes, diversification & risk control are key weapons in the investor's armoury to guard against this.
Mark
Hi Mark, sorry I think i have muddied the water here mentioning price targets and buy/sell recommendations. I totally disregard these and am only interested in the eps forecasts. But the method that stockopedia uses to calculate the momentum rank is what interests me. Particularly whether a new broker estimate is compared to consensus or the estimate made by the same broker previously. If it is against consensus the the momentum rank is liable to fluctuate more. Hopefully Ed will be able to confirm.
Interesting post!
I think it's probably true that most forecasts are inaccurate. I mean you only have to look at the divergence between different banks' estimates to realise they can't all be right and probably none a clue. An interesting website is Estimize (https://www.estimize.com/). This is an US site, which doesn't cover UK stocks. It uses crowdsourced estimates and according to them are more accurate that the Wall Street firms 70% of the time. It's an interesting concept I think and shame it doesn't offer UK coverage.
Estimates for smaller stock I think have more purpose. They show that the institution has gone out of its way to cover the smallcap stock, rather than simply following the norm. Having said, this doesn't mean to say they are anymore accurate - in fact, in my experience they are wildly inaccurate. However, I feel - from a personal perspective - I put more weight on a smallcap stock that has estimate attached to it than one that doesn't. I think this is driven more by institutional coverage itself, rather than any insight provided by the actual price target though.
Ed
" It would lead to crowding and higher risk."
Have you seen this happen ?
I find that the broker analyses are useful for pointing up company details that I might not have caught from just reading the accounts and brief statements.
It may be some comment like "South American sales are largely in Chile and therefore unaffected by recent changes in the laws of..." , I might not even have properly looked at the South American sales so it can send me back to check.
I particularly like to see the kind of denial of a problem that I never realized the company had since it tends to highlight a sensitive area that quite often is an problem
I'd like to think that brokers have sophisticated computer/ spreadsheet models into which they feed the latest information such as recent results, and based on a set of reasonable assumptions, out pop their estimates.
Of course, these estimates are wrong before the ink is dry, because the underlying assumptions are constantly changing in real life.
I could also be cynical and suggest that broker estimates are nothing more than the numbers fed to them by the company's management. So it isn't really their estimates.
That's a great comment and is too true. The Financial PR industry is mega powerful and encourages companies to play to the expectations of the crowd. PLCs are explicitly or implicitly nudged to report earnings in nice consistent trends - bringing forwards or delaying revenues quarter to quarter etc etc.
Every now and then there's fallout from this kind of behaviour with a 'big bath' profit warning - brokers are caught off guard and then claim they 'couldn't have predicted it'. "It's only when the tide goes out that you find out who has been swimming naked".
I'm being a cynic... but have been around markets too long. I've learnt not to trust experts too much.
If I remember correctly ValueRank is forward looking using broker estimates to calculate the P/E ratio, whilst the QualityRank is backward looking using values gleaned from the accounts.
With respect to earnings estimates the MomentumRank is more interested in the degree to which forward estimates have changed. So an analyst could be way out in terms of the specific estimates but might be correct in identifying a 10% increase in earnings.
Personally I tend to filter on Quality and Momentum factors and as such the precision of the earnings estimate isn't that much of an issue to me.
I look for positive EPS upgrades within 1 & 3 month periods, so an analyst being bullish about earnings.
The exception with respect to the accuracy of the estimates is that one of my additional checks is 'risk adjusted PEG < 1' which of course relies upon the specific forecast earnings.
Phil
It's best if I clarify here.
The ValueRank is completely backwards looking. It uses historic earnings, sales, dividends etc as its components.
The MomentumRank is composed both of Price Momentum and Earnings Momentum.
Earnings Momentum does rely on broker forecasts though not from the way you might think - it looks at the change in brokers forecasts and recommendations not their absolute level.
To clarify - a stock that has been recommended as a sell by brokers, but has seen:
will rank higher for "earnings momentum" than a stock ranked 'buy' with a vast price target.
The change or trend in forecasts is generally more predictive than the absolute level. This is a key, key thing to consider.
This is all explained in more detail in our ebook the Smart Money Playbook.
I hope with the use of Stockopedia that people will put less weight on broker 'buy' recommendations than they might have previously. If you want an illustration of quite how bad / biased broker price targets and forecasts can be have a read of this article - a gratuitous deconstruction of the stockranks.
I'm not saying all broker forecasts are wrong - just that in aggregate they aren't as great as people think !
Hi Ed,
Thanks for the clarification. Could you further clarify whether momentum is calculated on the change relative to the aggregate position of all broker estimates or the change on each individual broker. If it is based on the aggregate then the effect of one particularly bearish/bullish broker on the overall M rank will be marked (and unjustifiably in my opinion).
Probably best to explain with an example. I hold Sainsburys. Lots of brokers produce estimates and in aggregate the estimates are usually pretty accurate. However if you look at the individual brokers making up the average they vary wildly. You have Goldman Sachs who absolutely HATE Sainsburys and always rate it sell with a target price of around 50%-60% of the current trading price, with a terrible EPS forecast. Others who are more bullish - rating it strong buy, good EPS and a target price in excess of the current price. On average I find they tend to cancel each other out. However, if the average concensus on the 1st of the month is Hold and estimate EPS is 22p and then on the 2nd of the month Goldman issue their regular Strong SELL with an EPS of 10p this is going to have a monumental effect on the Momentum rank (in a negative way!). Then on the 10th of the month the bullish brokers issue their estimates saying EPS is going to be 33p - this is going to have a massively positive effect on the momentum rank. By the end of the month we are all back to the aggregate position - but during the course of the month the Momentum rank will have been flying around all over the place. If however you compare Goldman Sachs estimates on the 1st July with their estimates on the 1st June and use this in your calculation of Momentum then that is fantastic. So if Goldman go from Strong SELL with an EPS of 10p on 1st June to Strong SELL with an EPS of 12p on 1st July this is actually a massively positive signal (even if the aggregate is EPS of 22p)!
I hope that I have explained this ok - it is all a bit subtle and I am aware that I haven't spent as long as I should on re-reading and ensuring it makes sense. All the figures and dates are all made up off the top of my head by the way.
Hi Ed, I know that you'll have 1001 other jobs to do other than reply to every single post asking for your input - but I'd be really happy to know how the Momentum rank is calculated re: comment 12 of 20 on this post. If you consider this as one step too far in disclosing the inner workings of the Stockpedia ranking system then I will quietly back off. But if you could offer some further clarification it would be really useful for me in interpreting the movement of the Momentum rank. Many thanks, Dave
If you go into a share on Stockopedia and click on Momentum it does tell you how its calculated: http://help.stockopedia.com/knowledgebase/articles/182916-what-is-the-momentumrank-
Hi Herbie47, Thanks for your help, I've looked through this info and the question I've got is a little more specific. When a new broker estimate is issued and subsequently processed by Stockopedia - how is this translated into the Momentum score (that then feeds into the ranking process)? Is the new estimate compared against the consensus or is it compared to the last estimate provided by the same broker? I imagine that it is the former. If that is indeed the situation then it is likely to make the Momentum score and therefore Momentum rank a little more volatile over the course of a month/quarter.
I think it's all the estimates, so if one changes it will have a small effect of course depending on how many brokers cover that share, many small companies only have 1 or 2 brokers. Also its only part of the Momentum score. I think the share price movement will have more effect however a broker change in estimates could well do that, especially if it's tied in with a brokers buy or sell advice.
Perhaps worth noting that a lot of analysts and brokers have privileged access to the companies that they comment on and so produce commentary that is derivative of the (often rose tinted) views of the companies themselves on their future prospects. Even those that do not have access presumably know of those that do e.g. house brokers, so tend to follow their lead reinforcing the company view.
Relatively few company bosses are brave/foolhardy enough to buck the trend and call down their own prospects - consider what happened to the Next (LON:NXT) share price when the Chairman tried this earlier in the year - so we shouldn't be surprised that the weight of forward looking projections tend to overshoot the eventual outcome.
Gus.
Isn't the reality that some analysts do have access to privileged company information and/or are better at forecasting than others. I believe Thompson Reuters tracks individual analysts and their forecasting accuracy and allows professional investors to use this scoring data in their stock selection criteria. Maybe will come to Stockopedia in time?
I suppose the take home message is that the "average" analyst is hopeless at getting forecasts right and Ben's article confirms this! Somebody once said that predicting the future is difficult.
Sorry but I have clearly missed something. Doesn't Stockopedia use these forecast figures as part of the metrics to generate the stock ranks?
Of course.
The question is: Do you rely on them for investment decisions?
This article suggests: take broker's targets into account for short term forecasts only and even then with circumspection.
It also reminds us to be wary of 'anchoring', think for ourselves and, as you suggest, do not rely on stock ranks.
It is a well written, entertaining piece.
No I'm pretty sure the stockranks do not rely on forecasts (though I think the momentum rank incorporates among other factors whether previous earnings forecasts have in fact been beaten). Stockranks are backwards looking and look to measure factors that are empirically shown to be related to future success (rather than being based on company specific speculation about the future). I think that is partly the point of the article!
Edit - actually the momentum rank also incorporates whether there have been recent earnings forecast upgrades. Value and quality do not use forecasts though. So overall I think you could say the stockranks do not rely on the forecasts themselves.
Hi Ben. Thanks for a very interesting piece,
So what do you suggest that we do? Should we just base our buy and sell decisions based on StockRanks?
I like to follow my own investment process, which does take into account analyst's estimates, but I tweak them according to what I find out. It's a well defined process, which I strive to continuously improve.
Is there any research done into whether consensus estimates are less accurate when a low number of analysts publish earnings forecasts for the company in question. and more accurate when a high number of analysts publish forecasts?
I tend to mostly invest in smaller companies & normally spend about 3 or 4 hours analysing a company before deciding to buy. My process requires forecasts to be made. Would you and others suggest that I should completely disregard analysts forecasts and totally rely on my own forecasts instead?
Warranstar...
I don't think that's what Ben is saying at all. He's implying that rule based or checklist driven processes can counter the kinds of biases that bedevil our minds as human beings. Ultimately the risk assessment area of our minds evolved to run away from predators, not to make the kind of sound statistical assessments of risk that lead to sound investment decisions when under severe emotional pressure in stock markets.
Personally I use the StockRanks as a complete valuation system, but the last thing I'd want is everyone in the world using the same thing. It would lead to crowding and higher risk. One of points of this site is to let people build their own checklists, processes, rankings and screens - we're not all the way there yet but we will be eventually !