EDITOR'S TAKE: It's not THAT great at Goldman

eFC logo

Don't get too excited. Yes, Goldman reported exceptionally good Q1 earnings in a surprise move late last night, but before heralding a banking revival (or asking whether Goldman Sachs is the new Goldman Sachs), it's worth scattering a little rain on the gathering parade.

Firstly, Goldman pay is looking good, but it's not that good by its own historical standards. After cutting 7% of its people in the past quarter, GS allocated each individual employee an average of $168k for the past three months. This is up 35% on last year, but it's also down 35% on Q107. And as a commenter below points out, Goldman's pay for the last quarter also included severance for the 2,169 people it made redundant.

Secondly, Goldman's success is looking a little one sided. Back in the wonder years of 2006 and 2007, revenues were swelling across all business areas. This time, however, there was swelling in trading and nowhere else. Revenues in fixed income, currencies and commodities were up 34% on last year, and revenues in trading and principal investments were up nearly 40% year on year. In other areas, including investment banking and asset management, revenues were substantially reduced year on year and even quarter on quarter.

As various analysts point out, this implies Goldman is riding on the spreads of a dislocated fixed income market. When markets return to normal, so may its trading revenues.

Thirdly, even though Goldman is going to pay back its TARP money, lots of its senior people have been leaving. In the past month alone, we've counted at least four (Arthur Peponis, Byron Trott, Scott Norby, Ed Watts), suggesting the allure is fading slightly.

And lastly, as Barry Ritholtz points out, Goldman wouldn't have done nearly so well if they hadn't been able to bury all the bad news in the stub month of December.

This is not to deprecate the achievement of exceeding earnings expectations by 100%. But it is to say that Goldman may not be nearly as golden as its earnings suggest.