How reliable are the City's scribblers and should private investors pay heed to their recommendations?
It's a question worth pondering if you happen to have shares in Carillion, the support services group and general jack of all trades whose most recent claim to fame is building the £255m Olympics media centre. Its shares tumbled at the end of last week in response to some decidedly negative analyst comment.
The life of an analyst is not an easy one. They work long hours for ruthless investment banks which demand they justify their existence.
In the past this wasn't too hard, because fund managers paid for their research through various "soft" commissions, but these have been outlawed and life has got tougher.
Regulators have also cracked down on bankers pressuring analysts to write favourable reports on clients. Conflicts of interest have to be clearly disclosed on any published research, and there are supposed to be Chinese walls within banks to prevent this sort of thing – although they are made of paper.
Companies too aren't generally very happy if analysts write critical things about them. Again, the watchdogs have tightened up on various past malpractice. Stock market quoted businesses aren't supposed to play favourites and what is available to one is supposed to be made available to all. That's the theory. Behaviour in the real world doesn't always live up it.
When it comes to bigger companies, you can often find more original (and piquant) views from the smaller broking houses which don't work for the giants (and don't have much prospect of doing so in the future).
In general it always pays to be cautious, using analysts' views as guidance, but doing your own research at the same time.
That said, when three analysts at once downgrade a company at the same time (as happened with Carillion) you have to at least pay attention. And the downgrades came from across the piste, from the enormous UBS (which went from neutral to sell) to the more medium-sized RBC Capital Markets (from outperform to sector performer) to the boutique Liberum (from hold to sell).
Carillion's statement last week said trading was in line with expectations. It also tried to talk up its order book despite admitting that things are "challenging".
"The value of the group's order book and probable orders is expected to remain strong at around £18.0bn and the group's pipeline of contract opportunities is expected to have increased to some £35bn," it said.
But Liberum noted that margins have fallen sharply in the Middle East, where the start dates for various projects have been put back. Carillion noted that revenues will be below this time last year and said the latter was part of the reason.
Liberum, however, also noticed that "margins seem unsustainably high" in Britain.
Support services groups have been talking up the opportunities for outsourcing work available to them at a time of austerity. But if the Government and its civil servants get tough over costs, the extra opportunities might not yield as much as in the past.
UBS also called construction margins "unsustainable" and worried about growth in support services.
The shares slipped after the trading statement, although it wasn't a profit warning. But they slipped much more on the analysts' views.
Sentiment has changed quite a bit from last year when we were buyers at 323p. Well, you can't win 'em all.
Carillion trades on just 5.7 times 2012 forecast earnings, while offering a prospective yield of 6.3 per cent, which is comfortably covered by current earnings.
The group's order book still looks OK, but is below the £19.4bn when we last looked at the stock. Borrowing (at £125m) is manageable.
I'm not going to ignore the analysts' concerns. Even on such a rock-bottom valuation with a very tasty yield, I'd be wary about buying any more shares before seeing some more financial data from the company. That said, I'd be reluctant to sell immediately. The yield is certainly worth having. But it would pay to keep a close eye on Carillion's shares and act quickly if the outlook continues to decline. Hold.