1st February 2012
The US retailer's sales were some $800m below analysts' consensus figures. But even though sales grew more slowly than in the preceding nine months, the $17.4bn achieved was still up 35 per cent on the same time last year.
Cue customer complaints about missed service targets. And cue a discussion of whether Amazon's push for expansion has come at the expense of shareholder reward. It is now renting films through LoveFilm (competing against NetFlix) and selling Kindle readers and the Kindle Fire tablet computer (going against Apple) as well as selling an every increasing range of goods through its "market". Amazon executives may see themselves with different priorities to Amazon investors.
Certainly, Amazon remains a growth company, selling on around 140 times last year's earnings and some 60 times expected earnings per share for 2011. The share price is still discounting way above average growth for years to come. Despite the recent dips, it has made a lot of money for longer term investors. It is a similar phenomenon to the Tesco price crash last month when it announced growth but not as strong as analysts had expected.
But the real story is why the analysts' consensus turnover figures at Amazon were so wide of the mark. Losing $800m from a calculation is not easy. It's not the first occasion – in fact, they seem to be making a habit of over-estimating Amazon's ability to sell stuff. Back in October, just after the end of the third quarter, Reuters reported that sales had fallen well short of expectations.
US analysts were almost uniformly bullish ahead of the new Amazon numbers. Figures from Yahoo show that 21 analysts rated Amazon a "strong buy", two a "buy", and 10 a "hold". How many rated the stock a sell? Zero.
It is time to analyse the analysts.
Analysts often come up with similar numbers with few going higher than the consensus and even fewer aiming low. Academics Narasimham Jegadeesh and Woojin Kim looked at this question in 2007 for the US National Bureau of Economic Research.
They come up with two reasons for herding. One is that individual analysts may all be basing their figures on similar information. In both the UK and the US, analysts often have briefings with companies and receive "guidance" over the figures. When this guidance turns out to overestimate reality, then firms must expect investor wrath.
In many cases, analysts allow companies to preview their material before it is sent to investors. Those who do not follow this code may find that access to the company is restricted or denied in the future.
Investors need to question companies when so many analysts come up with inaccurate figures. Were they misled? What happened between the briefings and the actual results to render the guidance unhelpful?
The second reason for herding covers more situations than investment analysis. They herd because they find comfort from imitating each other. If the forecast is correct, each individual can take praise. But if it is wrong, neither investors nor the media can single out any one analyst. Going out on a limb increases the risks of getting it wrong and being noticed. In some stockbroker firms, the financial incentive structure punishes incorrect analysis more than it rewards getting it right – the analyst defence is "I was in the middle of the consensus".
It is difficult to attribute the balance between similarity of information and the need for comfort. But the consensus can lead to over-optimism among investors who do not recognise the herd mentality. This is dangerous.
An Israeli study on herding and analysts concludes that analysts follow each other to a substantial degree and that the tendency to herd becomes stronger the more analysts follow the company. So the biggest companies will have the most "consensus-driven" analysts.
Working through the analysts' methods
The US Securities and Exchange Commission has a good analysis of analysts . It warns investors to look at relationships between analysts and the firms they are researching and between analysts and the banks that are backing companies they are looking at.
It notes that clients prefer positive buy reports to negative sell advice. Additionally, analysts prefer buy recommendations as they attract clients – telling investors to sell generates little or no income.
Research analysts are not super-predictors with second sight. Most don't even have 20-20 vision. They tend to be overly optimistic, fast to upgrade and slow to downgrade. They may have closer relationships with the firms they analyse than with investors – it's tougher to meet a company person after a downgrade than an investor complaining about the lack of a warning of lower earnings.
Many analysts would rather sell their grannies into slavery than come out with a strong sell recommendation on a company with which they enjoy a relationship.
More from Mindful Money:
To receive our free email newsletter sign up here.