Looking behind the Goldman numbers
Goldman Sachs reported quarterly results last week that were, by any measure, stellar. But some of the numbers look a little bit too special.
Reporting on Goldman numbers is something of a parallel universe. For all of the anguish and hand-wringing about business models in investment banking, structural versus cyclical declines, the disappearance of liquidity/leverage/compensation/fun, Goldman keeps on delivering.
Letters from ex-employees, sexism lawsuits, secret tapes of regulatory conversations and cephalopodic comparisons all seem to slide off (at least, in the minds of the bank’s clients) and the maxim that “no one ever got fired for hiring Goldman Sachs” still seems to hold.
Goldman reported advisory revenues up 41% to $961m, with roles on five out of six completed M&A deals over $10bn in the last quarter — three of them as sole adviser — according to Dealogic.
Capital looks good, fixed income trading is humming along, compensation ratios have improved and the strong numbers prompted upgrades for the stock — JP Morgan raised its earning per share target to $18.6 from $16.3.
But the real puzzler is equity trading. From last week's reports, Citi and BAML were virtually flat on the year, JP Morgan was up 22%, but Goldman was up 46% to $2.32bn.
Most of this was derived from “equities client execution”, delivering a 170% surge to $1.12bn. Just pause for a moment and re-read those numbers. 170% to $1.12bn.
Morgan Stanley on Monday managed a 33% increase, pushing total revenues for equity sales and trading to $2.27bn — an improvement of a similar magnitude. But it does not segment the numbers further, so it's hard to tell if it is the same kind of concentrated performance, or spread across the full range of equity markets activities.
The first thing to consider is just how large Goldman's equity markets improvement is. The financial sector is notorious for volatile results, but this historically manifests itself in investment banking proper.
M&A, for obvious reasons, delivers revenues in lumps, and swings between feast and famine. More commoditised products (equities trading being a prime example) tend to give smoother ups and downs, and with less variation between firms.
It’s partly the law of large numbers — equity trades and associated revenues aren’t large enough for any given deal to skew a quarter’s revenues, so in expectation, a quarter’s numbers reflect market wide activities.
In M&A, meanwhile, deals like Covidien-Medtronic (where Goldman earned $58m, according to Dealogic), are large enough to flatter a whole quarter.
So what to make of Goldman’s 170% stunner?
Equities trading, whether in dark pools or through traditional broking, is often low-margin, standardised business. Structured products and prime brokerage get lumped into a broader equity trading number, but this is, quite specifically, “equities client execution”.
Goldman credited the 170% surge to “strong results in both derivatives and cash products across all major regions” and said it “operated in an environment generally characterised by more favourable market-making conditions, generally higher global equity prices and strong client activity levels.”
Which is fine. It is a good environment, equities are on the way up, the first quarter is usually strong, doubtless Goldman has a good client franchise that people want to trade with. But it isn’t really enough to explain a rise of this magnitude.
A 170% increase doesn’t just merit a shrug, and “good job” — one presumes Goldman’s franchise wasn’t that different a year ago, and its salespeople and their client roster must be much the same.
If GlobalCapital turned out a 189 page newspaper each week instead of our usual 70, we’d expect a response from the market more like “who’s been putting methamphetamines in YOUR porridge?”, not the blithe dismal of “a better backdrop”.
So where does that leave Goldman?
Whatever drove the surge, it wasn’t straightforward commission. “Commissions and fees” were down 2% to $808m.
The surge, instead, was probably captured in “market making”, which cuts across fixed income and equities, and rose from $2.64bn to $3.93bn. That sounds like Goldman is taking on more risk as an intermediary, and benefiting from price increases — a course which sounds like a more plausible source for a 170% jump that clients simply booking a lot of trades.
But here the trail goes cold — Goldman's average daily VaR exposed to equity prices actually went down substantially, from $32m to $24m.
Morgan Stanley's impressive equity number lends credence to the "good market backdrop" theory, and it offered a similarly anodyne explanation — "strong performance across products and regions on higher levels of client activity" — but it still leaves the other Wall Street firms oddly lagging.
Goldman or Morgan Stanley or both might have reclassified some activities into the equity numbers, or changed how they book equity revenues, but this would usually be disclosed — and previous years harmonised into the new category.
So Goldman-watchers have a mystery on their hands. As problems go, it’s a good one to have — but what do numbers this good hide?