Wednesday, June 18, 2008

Fair game?

While summer doesn't officially start until Saturday, in many ways it feels as if summer markets are already here. Vol sellers have emerged in equities, fixed income, commodities, and currencies, while conviction and trading interest is reported to fairly light across different products. Closer to home, both traffic and commenting in this space have dropped off over the last few days, though that may say more about a dearth of insight from your author than any underlying market condition.

In any event, yesterday saw a pretty limp reaction to what Macro Man thought would be a touchstone event, namely the release of Goldman Sachs' earnings (more on which below.) Action has been similarly tepid today in Europe, though Asian markets seem to have de-coupled, if only temporarily, overnight. Macro Man retains a short in various European indices, and has seen little to dissuade him that this is the right position to have.

Indeed, perhaps the biggest threat to equity shorts at the moment is that the regulators change the rules of the game. In the UK, for example, the FSA has taken Deutsche Bank to task for proposing a perfectly legitimate trading strategy. Indeed, the UK regulator seems to be more interested in squeezing/screwing shorts than they do about addressing malfeasance on the part of companies, particularly finance firms, which are the very reason that shorts are established to begin with.

Not that misguided regulatory zeal is confined to the Square Mile- far from it. Although the pre-ho Spitzer revolution has cracked down on the worst of the excesses in Wall Street research, the fact remains that analysts are, in aggregate, habitually overoptimistic on the market. The chart below compares the 12 month forward I/B/E/S consensus earnings estimates for the SPX with the actual result; as you can see, over the past 20 years the consensus has overestimated earnings by more than 7%. To put that into context, the actual annual earnings growth over that period has averaged 8.5%.

In fairness, the recent degree of over-estimation has been much less than previous cyclical peaks. But these index-level (over) estimates are only half the story. Because bizarrely, as far as Macro Man can make out, over history a majority of company-level earnings surprises tend to be positive. That is, comfortably more companies beat expectations than fail to beat. (Macro Man has tried and failed to access hard historical data on this phenomenon going back longer than a few quarters; if anyone has the data, by all means pass it along!)

So we're left with a situation where most analysts overestimate earnings on a 12 month forward basis but underestimate earnings on an ongoing quarterly basis. Hmmmm.....something smells kind of rotten there. If this earnings estimate game were truly a fair one, it would be reasonable to expect companies to beat earnings 50% of the time. yet clearly that's not the case.

Consider our friends at Goldman, who currently enjoy universal acclaim as "the smartest guys in the room." That may well be the case, but unless the analysts covering GS are at the other end of the scale, wouldn't we expect them to mark their estimates accordingly? Evidently not. Bloomberg stores quarterly estimates for GS since the beginning of 2006. Since that period, GS have beaten those estimates ten out of ten times. In a "fair game", the cumulative probability of that occurring is less than 0.10%. Indeed, the best model for predicting GS earnings appears to taking the consensus estimate and adding a buck per share.
Judge for yourself whether this is the result of analyst incompetence, or whether GS is gaming the system. It's not like every other IB hits the ball out of the park every quarter. Lehman and Merrill have famously had recent difficulties, and even Morgan Stanley (reporting today) has disappointed expectations three times since the beginning of 2006.
Unfortunately, regulatory zeal apparently doesn't extend to determining just how Goldies (and to be fair, they're hardly alone in this game) manages to beat expectations quarter after quarter. Naked Capitalism highlights potential dodgy goings-on at Lehman with respect to how they managed de-lever in such a crappy market.

That equity markets are gamed isn't exactly new news, and Macro Man doesn't want to come across as a doe-eyed naif who's just discovered that Santa Claus doesn't exist. But just as equities were de-rated in the wake of the late 90's/early Noughties corporate scandals, Macro Man can't help but wonder if a similar de-rating won't occur now, sourced in the twin concerns of inflation and the fact that stocks aren't a fair game.


Apprentice said...

Hi, Macro-Man! has some more history for earnings and estimates,
enter GS in "quick search" field.

Acording to data, GS beat estimates 26 out of 28 times from Sep 2001.

spagetti said...

one place where Goldman may have been right about 'decoupling' is its own fortune.

but, one swallow doesnt make a summer

Anonymous said...

MM, Did you ever get round to buying any short-sterling? The strip of call condors playing for on hold bought today look a pretty sensible 4-1 wager.

Macro Man said...

I did some call flies last week. Fairly low risk and high reward, but will need rate cuts to be priced at some point this year to really make a lot of dough. Profit: investment ratio was ~ 7:1. Still, in short sterling, I think it's the "low risk" bit that's the most important part of the equation.

Anonymous said...

Goldman Suks is the dirtiest whore house on the street and with the coming change in administrations some of their luck may soon run out.

Anonymous said...

great points about the earnings "surpises"

Andy said...

Hi Macro Man, great blog.

The charitable explanation for the continuing pervasive influence of analyst earnings estimates is that everyone knows that these pesky analysts are constantly playing three card monte with their estimates, but we're stuck in a local equilibrium where everyone expects everyone else to slavishly follow said analysts, and hence it becomes self-fulfilling.
The uncharitable explanation is that equity managers are idiots.
Take your pick. (obviously, the two explanations are not mutually exclusive).

Personally I find it more than a little bizarre to be living in a world where ratio of positive to negative surprises or whatnot is bandied about by seemingly sentient creatures as a buy signal, and with a straight face too!

Macro Man said...

Apprentice, spagetti...I did a back of the envelope calculation on the chance of Goldies getting 26 of 28 beats in a "fair game"....suffice to say it is a very small number, less than 0.1% chance. Judge for yourself whether this game is, in fact, fair.

Even MS today managed to "beat" the number...yet every single particular looked bad, with every analyst I've seen saying that this is really a "miss." It seems there's been some jiggery-pokery to massage the fig towards the street estimate; refreshingly, thanks to fedex, the market's looked through it...

Anonymous said...

The reason of course is obvious: the estimates are publicly known and available to the company's accountants.

They make small adjustments every Q to beat by a little, until this stops working, and they take a ginormous """one time""" miss, just like a few years ago.

So, total cumulative earnings can be overestimated and yet most quarters beat estimates by a little bit.

Anonymous said...


Two ways the discrepancy between the fact that more companies beat expectaions individually while the aggregate expectations fall short of reality can be explained.

One, AFAIK (and I don't know enoguh). When they say Goldman beat expectations 26 out of 28 times, are we talking about expectations 12 month ahead or beginning of quarter or month ahead of earnings release? Earnings expectations undergo substantive revisions over the course of a quarter let alone a year.

The second point is that when companies beat expectations, they usually do so marginally whereas when the fall short they do so by a large margin (of course this is related to optimizing the earnings management game). So, even if a majority of companies beat expectations, the few that do not more than offset that.

Macro Man said... in other words, it's not a fair game. Someone remind me again why we want to pick the bottom in financials?
Oh, that's right, because Warren's gonna buy Wachovia. Just like he wanted to buy a slug of Bear and Lehman...

Andy said...

Anlysts overpredict year-ahead earnings in order to sell a story and generate nice low forward P/Es, which they use to impress the impressionable by comparing said forward P/Es to historic backward-looking P/Es. They then spend the next few quarters massaging the expected earnings number lower so that the stock can impress the impressionable by beating expectations, while projecting year-ahead earnings to grow by an unfeasible number in order to...

It's a shell game; think the earnings growth is under this shell? Nope; it's been quietly palmed and is under the other one. Wanna play again?

This dynamic is pervasive, and is independent of the 'eat like a bird, shit like an elephant' dynamic correctly highlighted by other posters.

cordura21 said...
This comment has been removed by the author.
cordura21 said...

hey MM, I like your site cause your view is well...macro. And me being from Argentina, it's cool. Anyways, here is a link to an EPS Beat Rate analysis made by the Bespoke guys from 1998.