How does one value a bank? That's an interesting question. When looking at how much a bank's shares are worth as a multiple of its earnings, analysts usually consider its "underlying" profit. This sounds like what a bank keeps under the bed – which is where many people think their money would be safer, given the way banks have behaved. It is a number that retail shareholders (you and me), as opposed to City institutions, might view with a degree of scepticism.
But here's why it's important: the underlying earnings figure aims to give a picture of a bank's actual operating performance, taking into account only the ongoing costs of doing business and what the bank makes over and above them. It excludes lots of one-off factors, many of which are outside management's control.
Take the value of a bank's own debt. This can vary wildly, for all sorts of reasons largely beyond a bank's control. And yet if there's any change it has an impact on the headline pre-tax profit, despite the fact that it has no impact on the amount of cash a bank brings in and pays out. So it isn't included in "underlying" earnings.
The same goes for changes in currency values. Banks "smooth" these for the purposes of the "underlying number" because, again, they have little control over, say, the dollar's value against the pound.
More controversially, changes in the value of acquisitions are also excluded. If, say, HSBC paid £1bn for a company that underperformed and the accountants said would really be worth only £500m were it sold, the value has to be "written down".
There's no actual cash involved, but £500m gets lopped off the pre-tax profits whenever the bank 'fesses up and takes the hit. Because such writedowns are "one-off", non-recurring events, analysts ignore them for the purposes of their underlying earnings forecasts. Banks do also sell businesses for a profit, and these are not included in the "underlying" profit figure for the same reason that writedowns aren't.
Other one-offs excluded from underlying profits include unusual tax charges or benefits, accounting anomalies and legal judgments. Oh, and fines and compensation payments.
That's where the debate starts to get interesting. Since the end of the financial crisis there has hardly been a reporting period when the big banks haven't had to report some sort of increase in provisions for fines or compensation to mis-sold customers. So, should these really be considered as "one-off" losses?
You could easily argue that they have become a regular and ongoing cost; that banks view fines and compensation as a necessary part of doing business, launching new products and selling them aggressively while paying little heed to their impact on the customer, in the knowledge that there could be a price down the line if things go pop.
The trouble is, it is very hard to put an estimate, or even a guestimate, on the cost of the scandals this behaviour generates. And of course we don't yet know where the next scandal is coming from, only that there will probably be one.
As a result, any forecast of future "underlying" earnings really ought to come with an asterisk.
A case in point were yesterday's HSBC results. The headline pre-tax profits halved to $3.5bn as a result of those movements in the value of own debt and various one-offs. But "underlying earnings", at a shade over $5bn for the third quarter, were more than double the $2.2bn in the same three months in 2011 (but still short of the $5.6bn the City had hoped for).
Not included in that figure was a $1.1bn increase in money set aside to cover scandals. Some $353m was for compensating those mis-sold payment protection insurance, but the real nasty was $800m more to cover a forthcoming US fine for money laundering.
The total HSBC has set aside is now $1.5bn but the bank admits it could get worse. And then there's the Libor scandal, which may hit later on.
Now you see why stripping this sort of thing out of "underlying" earnings might not be so sensible. These provisions are not "one-offs".
HSBC is no longer empire-building, but is trying to streamline a bloated operational structure with too few controls.
With lending principally funded by deposits, enough capital for even the most conservative of regulators, and a huge Asian business, HSBC still deserves to trade at a premium to other UK listed banks. And it does, at about 12 times this year's forecast underlying earnings. Barclays, by contrast, trades on just 8 times forecast earnings.
I'd like to say hold HSBC. The dividend yield, at 4.5 per cent, is solid and its management is finally doing the right things. But those skeletons in the closet are still rattling, The shares are going to endure a bumpy ride until there is some more certainty on those "one-offs". For the moment, I'd steer clear.
- More about: