One would hope that the Volcker rule, which will have far-reaching effects on the financial services industry on these shores, is founded on a sound understanding of what went wrong during the crisis and how to prevent a recurrence. It targets proprietary trading, or trading for a bank’s own account as opposed to market making, as the culprit. But what if the massive bank losses that heralded the recession were not generated by proprietary trading, but by activities that will remain permissible? Namely, what if the trading losses were the result of underwriting activities, which must remain permissible for the health of the financial system? Then the very disruptive legislation will prove to have been sadly misguided.

Proprietary trading is notoriously difficult to define or track. Proprietary positions, meaning trades where the risk is for the bank’s own account, can be of three types, but only the first two will be restricted:
• Clean proprietary positions: put on by traders with no client contact
• Gray area: trades that are the result of a transaction made to accommodate a client but are maintained on balance sheet for a longer period either because no market exists or to speculate on price appreciation.
• Underwriting positions: positions in securities that raised financing for clients. The securities are intended for distribution to investors, which can take some time.
Underwriting remains a permissible activity under the Volcker rule. While underwriting positions are proprietary in the sense that they do not result from very short-term holdings characteristic of market making inventory positions, they are not speculative in nature, but rather necessary to the process of intermediating credit. However, when markets become dislocated, the distribution process stalls, and banks get caught with “hung” underwriting positions that may be illiquid for an extended period of time. Historically, “hung” positions have caused large losses at institutions. Other permissible trading activity is market making, which involves taking on short-term inventory positions to facilitate client transactions, or hedging client positions.

Contrary to popular belief, the greater part of the losses in 2007-8, particularly at the three banks most severely hit (Citigroup, the then independent Merrill Lynch, and UBS), were not in fact the result of proprietary trading. They originated instead in the underwriting activities of these banks. Therefore, they would have been permissible had the Volcker rule been in place.

Before the crisis, Citigroup, Merrill and UBS were the leaders in underwriting mortgage CDO’s. Underwriting structured transactions involves accumulating a “warehouse” of loans or securities sufficient to fill out a pool of assets of a size that can be distributed in the marketplace. Then it involves converting this pool into a series of securities, each bearing a different portion of the risks of this pool. Both the “warehouse” and the securities, as underwriting assets, are held in the trading portfolio, and changes in market value of these positions are reported as trading income or loss. These gains/ losses are indistinguishable to readers of financial statements from real proprietary trading losses.

Given that the trading losses were concentrated in these three underwriters, it stands to reason that many of their losses were in fact underwriting losses. In the US, of the $151 billion of trading losses on so-called “legacy assets” (meaning primarily mortgage-related assets and leveraged loans) reported in 2007-8 by the six large banks, $113 billion, or 75% were at Citigroup and Merrill Lynch/Bank America. Indeed, their annual reports clearly describe the ways in which their securitization structuring and loan syndication businesses generated those “legacy assets”. Only Morgan Stanley indicates that many of its $10 billion of losses on mortgage securities and leveraged loans during those years were on proprietary trading positions.

Even the GAO report (http://www.gao.gov/new.items/d11529.pdf) on proprietary trading found only $10.5 billion of net losses in 2007-8 from the type of trading positions that can be cleanly identified as proprietary. It mentions that one institution had two quarterly losses totaling $10.5 billion during that period, so likely that institution was Morgan Stanley.

The onerous part of the Volcker rule has to do with the gray area of proprietary activities that are a part of market making activities but are difficult to separate out or even define. How large could losses from such activities have been? If identifiable proprietary losses in 2008-9 were $10.5 billion and underwriting-related losses were $113 billion, that leaves only about $39.5 billion of the remaining “legacy asset” losses as potential candidates for proprietary losses of the more difficult to identify type. Split six ways over two years, that is not the magnitude of loss that would sink any one of those large institutions. It is also good to remember that total trading losses over the two years were only $63 billion for the six firms, meaning that the rest of the trading operations generated $98.5 billion of trading income including any gray area losses from legacy assets.

If the problems of proprietary trading have been misidentified, then the chances of the solutions proposed by the Volcker rule being à propos are slim. The rule will impose huge expenses in rewriting computer tracking systems and staffing compliance organizations. It may also have an impact on the liquidity of certain markets that will disadvantage investors and render the financial system more, rather than less unstable. Given that the US is alone on this, and that no other country is following suit, an important part of the US economy, the financial services industry, is placed at a disadvantage. Policymaking may be better off thinking about how to improve underwriting risk management, both for trading and for bread and butter lending, than addressing a problem that didn’t exist.