Beware the WSJ’s pay statistics

By Felix Salmon
February 3, 2011
habit: today's WSJ article claiming that Wall Street pay has hit a new record high in 2010 is seriously flawed.  

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

This is getting to be a habit: today’s WSJ article claiming that Wall Street pay has hit a new record high in 2010 is seriously flawed. “In 2010, total compensation and benefits at publicly traded Wall Street banks and securities firms hit a record of $135 billion”, the paper says — but you have to really hunt to find the small print. First click on the “Interactive Graphics” tab at the top of the story, then click on “About our methodology” at the bottom of that tab. A window pops up in Flash, with uncopyable text saying this:

The analysis includes important caveats. Figures don’t reflect results of acquired companies before they were purchased. For example, BlackRock Inc. bought Barclays Global Investors in 2009, while Bank of America Corp. acquired Merrill Lynch & Co.

It’s not that the WSJ couldn’t have done a proper analysis. Certainly Bear Stearns and Merrill Lynch and Lehman Brothers were an important part of Wall Street compensation in 2006 and 2007, and they made their compensation numbers public. If you want to run the total amount of money paid to employees by public Wall Street firms over time, you have to include the firms which no longer exist.

The WSJ doesn’t do so, however — and, what’s worse, hides the fact that its analysis is woefully incomplete. The problem, I think, is with data providers, who are really bad at providing any kind of historical data for public companies which used to exist but don’t any more. Companies tend to be identified by their ticker symbol, and so when the ticker symbol disappears, so does the data — especially when the ticker is then taken over by someone completely different.

But it reveals little to say that Bank of America or JP Morgan are paying their employees more than they did in 2007, when the real comparison should be with how much Bank of America and Merrill Lynch combined were paying, or JP Morgan and Bear Stearns. (And WaMu, for that matter.) The WSJ has the resources to find SEC filings for old public companies and add them in to its calculations. Not to do so is just laziness, plus a natural gravitation towards the most sensational possible headline.

(Cross-posted at CJR)

More From Felix Salmon
Post Felix
The Piketty pessimist
The most expensive lottery ticket in the world
The problems of HFT, Joe Stiglitz edition
Private equity math, Nuveen edition
Five explanations for Greece’s bond yield
Comments
2 comments so far

+1 on the PAINFUL disappearance of acquired companies.
The SEC is a god-send in this regard.

Of course the WSJ should have compared like-for-like, and it was laziness not to. But given it would take what 30mins on Edgar to do this, could you not have done the math for us Felix??

Posted by TinyTim1 | Report as abusive

Is this just XLR from Compustat? I could run those numbers, including acquired companies, un 10 minutes.

Posted by guanix | Report as abusive
Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/