Is Goldman Materially Misrepresenting Its Prop Trading Exposure?

Recently Goldman Sachs has been attempting to downplay the impact of prop trading on its operations, with various executives, among them both Lloyd Blankfein and David Viniar, claiming that proprietary trading accounts for a mere 10% of total revenue. This is likely a major misrepresentation and a substantial underestimation of the true impact of prop trading to the firm if an earlier analysis by third party credit analysis firm CreditSights is correct. According to CS analysts, Goldman's true prop exposure is at least 30% and probably inbetween 30% and 40%. This would imply that the proposed ban will have a truly material impact on Goldman, much more so than Goldman's executives claim.

The problem with prop trading, as has been discussed previously, is its narrow defition: it is impossible to isolate prop as a distinct profit center as it has extensive client-facing and risk-mitigating aspects embedded within it. For an OTC trading firm, such as Goldman, which holds large amounts of inventory in stock as it manages it flow demand, prop trading is as embedded in Goldman's business model as dispersing Goldman alumni to various regional Federal Reserve banks.

Following up on the earlier complexities associated with prop trading definitions from Meredith Whitney and Bernstein, here is the suggestion from Credit Sights as to why Obama's suggestion may be somewhat toothless:

Even in the event that the restriction on proprietary trading by banks is passed, we believe the practical separation of “proprietary” trading from “client activity” could prove extremely difficult to delineate. For instance, we sense that Goldman Sachs would have deemed its massive housing industry short position to be simply a “risk management” tool directly related to offsetting its other balance sheet exposures, which were facilitated by clients. So at the end of the day, the actual limits imposed by this proposal, even if it is enacted, may be difficult to execute on clearly.

Yet no matter the practical implications, should Obama be sincere in his desire to "mitigate risk", we are certain that Volcker has a few cards up his sleeve. So going back to the original question of Goldman's actual prop exposure, here is once again Credit Sights' perspective. In a nutshell, the firm uses BBB, and lower-rated counterparty exposure as a direct (and previously corroborated) estimation of in-house hedge fund and prop trading.

Since the mid-to-late 2000s, we have attempted to approximate the level of proprietary trading activity at major banks and brokers. Before the credit bubble burst in early 2007, some big banks/brokers were noting that 20% to 40% of their capital markets revenues were related to in-house hedge fund or proprietary trading activities. Some, such as Goldman Sachs, noted at that time that it was difficult to separate the customer business from the proprietary trading as it had become more integrated. This may now mean, as Goldman Sachs suggested by CEO Blankfein's testimony last week, that the trading businesses have a risk management component related to the relationship between customer trading and demand for yield structures and the proprietary trading business.


Our approximation methodology looked at the percent of derivative trading activity to counterparties rated BBB or lower. Back in the mid-to-late 2000s, this seemed to track the proportion of proprietary trading, in the range of 20% to 40%, some of the big banks and brokers corroborated. From 3Q09 10Qs, we estimate that Goldman Sachs, Morgan Stanley, and JPMorgan were in the range of 32% to 38%, or in the range of the mid-to-late 2000s. Interestingly, Citigroup and Bank of America were at higher levels of about 46% and 69%, respectively. We would note that these are very rough numbers, and they should be understood in the context of their overall diversified banking operations. Also, we believe that Bank of America's acquisition of Merrill Lynch may have elevated its prop trading content as the former broker operated at higher levels than legacy BofA. We would also note that Citi's figures (~46%) for 3Q09 includes its corporate lending commitments, while year-end 2008 is a pure counterparty figure (32%) which it reports only once a year. BofA's figures are related to its CDS counterparty exposures.

One can see why there is a "bit" of a discrepancy here: Goldman reports 10%, while realistic derivative exposure to less than prime counterparties (one wonder if AIG was included in this list) indicate that the number may be up to 4 times higher. Which leave the "risk management" straw man.

As readers will recall, this is an exchange between Goldman Sachs and Zero Hedge from a month ago, in which we specifically asked the question of how (if at all) Goldman's prop positions allegedly mitigate risk exposure:

14)     “Do you have a risk policy?”

Yes.  We think of risk management as being one of our core competencies and it remains integral to our success as a firm. 

Our management team is active in risk management discussions across the firm and open discussion on the subject are encouraged.  By the way, we think fair value accounting is a critically important aspect of risk management.  Another important tenet of our approach to risk management is the independence of control functions from the business units

We also use a variety of approaches to monitor risk.  In addition to VaR, we use multiple stressed-based methodologies, including jump-to-default analyses, to quantify tail risks. 

16) “How do you monitor trading/hedging limits?”

Virtually all of our equity and fixed Income businesses receive VaR based risk-limits, aged inventory limits and balance sheet limits. The limits are reviewed by senior management and Risk Committee on a regular basis.

Since Goldman is adamant that the vast majority of the gross prop exposure is merely to mitigate risk, can the firm provide an extended analysis of how 30% of revenue is generated simply as a function of "risk management?" Furthermore, the implication that 10% is a mere 25% of total exposure, does this mean that there is a walled off section of trading, and if so, why foes Goldman not disclose that the firm controls the risk of customer trades (very profitably at that) by running a flat or balance book of customer trades? Put specifically, our question is whether Goldman can effectively state, just like UGI Energy's Robb Gibbs, and Guy Butler, that it runs a "flat book" with no speculative interest?

In fact, as regulators consider the vast implications of Obama's overture, that one aspect on Goldman's business model they should focus on is how to allow the firm to merely take flow-risk mitigation positions -i.e., positions that allow it to hedge any and all daily inventory it may hold as part of its flow trading operations, and not allow it to piggyback on/ahead of/or behind existing orders, whereby the risk is not client-facing but merely investment capital risk.