The Unclear Hedging Boundary: Should Banks Trade Their Own Money?
Attention is focused on the fiscal showdown that looks like it will accompany a serious disagreement (hopefully) over how to navigate the fiscal deficit and structural debt position. There are numerous unknowns that accompany these kinds of dramatic occurrences, even though we ran through this in largely the same fashion in 2011. Part of the efforts of the US Treasury Department will be focused on stretching the federal government's statutory borrowing authority as far as possible, including issuing as little new debt as they can get away with.
The unintended consequences of such measures, again as we saw in 2011, go right to bank liquidity. My colleague Doug Terry alerted me to the 4-week bill auction results from September 17, the day before the FOMC decision. The "high" rate was 0.00%. According to the Treasury Department itself, that matched the record low yield of December 9, 2008. While some observers pointed to this "market" action as proof of resilience in the face of the looming shutdown, it was nothing more than preparation for this expected period of diminished Treasury borrowing, and thus less collateral.
In the days since that auction, bill rates have dropped below zero here and there as demand for T-bill collateral faces this restraining supply (already short throughout much of the year). As part of the auction process, primary dealers are the major source of bid "money" in most auctions. In that September 17 bill auction, primary dealers submitted $140 billion in bids (at 0%) and were awarded $22.9 billion in bills, or almost exactly two-thirds of the total offering.
That followed another unusual bill auction from September 9, where Goldman Sachs was "glitched" out of its bids for 3-month bills in total. The Treasury Department made up for the malfunction by over-allotting Goldman 6-month bills, briefly causing a stir in repo and bill markets as Goldman had to adjust its client book.
For the most part, the primary dealers take Treasury offerings into their inventory with the stated intent to sell them over time to clients. These client orders are, in fact, incorporated into the auctions themselves as the Treasury Department surveys primary dealer appetite as a proxy for their clients' overall demand. That information is used to set the auction parameters.
The overall auction/primary dealer system is certainly beneficial as it almost assures there will never be any "busted" auctions, and that the federal government will be able to fund its needs without interruption. However you feel about the level of debt and debt funding, the smooth ability of treasury mechanics is unquestionably a net positive for not just the government's operation, but the financial system as a whole. It is one of the most beneficial services provided by the Wall Street banks, and one that contours most closely to the core idea of intermediation.
Since there is a clearing function involved here, due to the mismatch between the timing of client demand and the treasury auction schedule, the primary dealers use their balance sheet capacity in a warehousing function. When they take down their proportion of the government bond auction, the bonds sit in inventory while the vast and motivated Wall Street sales-force scours the investment landscape (their client "rolodex") to move the inventory out to the world. That opens the primary dealers to some risk.
Even though the primary dealers know a lot ahead of time about which clients have demand and for how much, it is not foolproof. Trade failures, clients that renege on promises, etc., mean bonds in inventory may not move as fast as anticipated. Market conditions also change between auctions, meaning client demand that was there before auction may not be the same (in terms of price and quantity) afterward.
To counteract these risks, the Wall Street banks hedge their inventory, and the regulatory environment has created special accounting rules for "warehousing." By and large, there are not any capital charges for bonds in inventory with the intent on being sold. Again, that is a good idea for ensuring the smooth operations of treasury auctions since the government has a direct interest in primary dealer willingness to buy at auction. It is, most assuredly, a mutually beneficial arrangement.
Carrying and hedging bond inventory, even with minimal risks, does lead to losses, and gains. If a primary dealer buys a small amount of an interest rate hedge for a certain period that expires unused, the "loss" on the contract is run through the income statement under "principal transactions." On the other hand, if bond prices move favorably while bonds are in inventory, that move will be recognized as a gain on the same line. In other words, if between the time the primary dealers buy treasuries at auction and the time they sell to clients interest rates in general fall, the banks will have "earned" a premium on their inventory since the effective price they sell to clients has moved further above carrying costs (banks always charge a small spread between what they pay at auction and what clients pay to them; that is the stated incentive for engaging in this market service).
As a result of the rather thin spreads that are available in the inventory of auctioned securities, hedging costs need to be managed expeditiously. Attention on risk and hedging, then, is vital to keep from blowing the spread and making the whole service uneconomical. The smallish spreads themselves are an indication of competition, which is certainly desirable. But there is a window of opportunity here that complicates everything.
Banks, by and large, hedge not on an individual security basis but as a portfolio in toto. Risk may be divided by "desk", but the overall risk management is done by looking at portfolio risks combined. Individual hedging would be far too cumbersome and inefficient on a transaction basis to make it practical, particularly since bank inventory is not limited to just treasury securities. Wall Street banks also underwrite corporate bond offerings, municipals and so many other securities. Again, this is an extremely beneficial arrangement where Wall Street "greases" the wheels of real economy finance. This is the core mission of the banks themselves, or what it is supposed to be.
So taken from a portfolio perspective, a bank's inventory is really a complex risk matrix where there are securities from all sorts of places and classes. Risk management, then, has evolved with the complexity of inventories into a black box system - the banks have internal (and very secret) calculation engines and models that tell them how much to hedge and exactly where. The idea, again, is to be the most economical to keep costs down so that overall spread returns are positive.
For example, if a hypothetical bank's inventory included a mix of credit assets totaling about $10 billion, the bank would not hedge the entire carrying cost of that inventory. Instead, through its black box models, it will calculate expected "risks", and determine just how much hedging will cover the most likely risks. This is called dynamic hedging, or delta hedging. The models incorporate many factors, but the most important are correlation, delta, gamma and vega. If the combined calculations of these risk estimates determine that the anticipated potential change in the underlying prices is only $500 million, then the bank will only hedge for the $500 million (at most).
However, when does a hedge no longer become a hedge? Where does the line exist between hedging and speculation?
Hypothetically, there is nothing stopping a bank from applying, in the situation described above, $1 billion in hedges instead of $500 million. Again, the idea is to be cost effective, but what if that additional "hedge" position instead expresses an intent to profit off movements in the hedges themselves? If the bank "guessed" correctly, the "losses" in the $10 billion portfolio are not only recouped by the hedging positions, the bank would actually profit potentially to a greater degree.
There is an almost infinite range of possibilities in the ways that this relationship could be exploited. Not only can banks express a risky position (speculation) through their hedges, they can do so through the inventory itself, or in some combination of intertwined trading positions that, at the outset on the surface, look exactly like the vanilla hedge/inventory relationship. The bottom line here is that while securities are in inventory banks have an opportunity to make money for themselves given the environment.
That is not, itself, necessarily a bad outcome or process. But the potential for distortion in the intermediation process is not neutral. We already know this because this was the heart of the collapse in 2008.
In the three months ended March 30, 2007 (the date used in the regulatory filings), Merrill Lynch reported total net revenue of $9.8 billion. Net profit was $2.2 billion. Of that revenue, $2.7 billion came from "principal transactions." That meant 28% of revenues came from trading activities as I described them above, often called "proprietary trading."
But even that understates the importance here. If we think of Merrill Lynch's inventory service as its core function, then absent all the principal transactions the bank actually lost money; it was entirely unprofitable without appealing to the monetary opportunities in principal transactions. That was likely a powerful incentive to stretch the meaning and function of hedging positions and warehousing activities.
It wasn't just Merrill Lynch, either. All of Wall Street's results view in much the same way. In the nine months ended August 31, 2007, Bear Stearns posted net revenue of $6.3 billion. That was down from $6.8 billion in the comparable period of 2006. The decline in revenue was all due to the first rumbles of the looming crisis - principal transactions revenue dropped from $3.7 billion to $2.9 billion, meaning Bear was deriving more than half of its revenue from proprietary positions. In terms of profitability, however, there would have been multi-billion dollar losses absent principal trading.
Even the dramatic increase in the importance of principal transactions does not fully capture the element of corruption contained in the changing incentives of Wall Street banks. No doubt most people, even those with low awareness, remember the huge banking losses that were taken in the period between late 2007 and early 2009. Almost all of those losses were due to principal transactions.
Merrill Lynch, by the time it was "forced" into Bank of America, had seen its prop trading swing from such a massive moneymaker to near insolvency. For the nine months ended September 26, 2008, revenue in principal transactions was minus $13.1 billion. At Citigroup, where the Smith Barney subsidiary was a much smaller part of the overall bank, losses on principal transactions totaled minus $22.2 billion. Citigroup could better absorb such losses because the integrated, multi-faceted conglomeration had $53 billion in total net interest revenue and a much larger liability structure to cushion the blow. Nonetheless, Citigroup was in dire straits by that November, nearly failing because its principal transaction losses were staggering.
How does a bank lose so many billions just warehousing inventory, providing clearing services to the US Treasury and other asset classes? It was entirely due to the incentive structure of the business itself, and the window of opportunity provided by regulations that assumed this was a relatively benign business. To put it most harshly, the inventory servicing was just a front, a window dressing for the banks to use as a means to trade their own accounts without scrutiny.
That is where the subprime mess itself fit most directly. The banks, through warehousing, were only too willing to lend to lower credit obligors because it offered not just a market opportunity for clients to short subprime, but an opportunity to do so themselves. I have recounted many times the Morgan Stanley story of a huge loss taken while purportedly hedging a large subprime structured product. While warehousing the piece, they were "hedging" their exposure with the open intent on making big money off the hedge. It ended up blowing up on them as the convexity of the end pieces did not conform to modeled assumptions.
In the years since those scary months, proprietary trading and principal transactions have finally garnered scrutiny, largely through the worthless Dodd Frank bill even though it includes the Volcker Rule. However, how can you even begin to separate what is a legitimate hedge or principal transaction from what is pure, risky speculation? The markets absolutely need inventory and warehousing, but there will always be the temptation to use that inventory to trade the banks' own accounts (London Whale).
This may take increased importance now that interest rates and credit markets have experienced the first major selloff in decades. The first estimates of bank trading results that have been "leaked" in recent days and weeks are not optimistic. And, perhaps more importantly, this extends to derivatives trading positions where Wall Street banks act as both dealers and speculators. If I am correct about the swaps markets, as large trading positions were clearly favoring taking fixed in the days after QE3 was announced in September 2012, there might be an "unanticipated" string of trading losses tied to these derivatives.
I have no doubt that banks will respond as they always have, proclaiming the innocence of true hedging positions. And there will be no means to counter those claims unless, as in 2008, they are so obviously incongruous to the actual level of their services. We won't know how much of the potential losses are derived from providing hedging services to real economy participants (companies seeking interest rate hedges to guard against an increase in rates on their floating rate debt), and how much was just pure speculation taken in anticipation of QE-forever. I have no doubt there was some of both going on.
It is a legitimate question to ask whether banks should be able to trade their own money. Clearly, that is something they desire, spending an inordinate amount of resources trying to maintain that ability (and long before proprietary trading became a mainstream issue in the panic). From my own analysis, I cannot find anything beneficial to the real economy system from banks doing so. The gains in trading revenues ultimately distorted and corrupted the function of these firms to the point the vital functions they were supposed to perform were pushed so far down the list of priorities as to be unimportant and ineffective. Intermediaries are supposed to use specialized knowledge to separate potential "investment" opportunities by sustainability and profitability to the real economy. Instead, intermediaries stopped acting as that kind of gatekeeper, using their specialized knowledge to profit for themselves at the expense of the real economy.
Consider that Wall Street warehousing embraced subprime largely because of the principal transactions revenue opportunities it provided. If there were no such opportunities available, these banks might have actually performed that intermediation function because there was no other means of profiting to be gained. In other words, if there was no real and large profit opportunity in warehousing subprime loans (and then speculating off that), the costs of actually hedging the warehoused loans, since they carried a high degree of even recognized risk, would have made them uneconomical in total. If Wall Street stuck to the warehousing function, subprime might never have been made to the degree it had. We will never know for sure, but the trading results throughout the housing bubble/bust period suggest, to me anyway, that absent the immense profit opportunity of the distortion toward principal trading, there would have been little reason to engage massive and dangerous proportions.
There was no single cause to the housing bubble and collapse, but the window of opportunity to distort a basic investment banking function was certainly one of them. And like the corruption of AIG's insurance business, it likely would never have happened, and certainly not to the degree that nearly caused total chaos and financial destruction, absent monetary policy so favorable to it. Without an open-ended funding commitment on the Fed's part through interest rate targeting, not to mention the squeeze of low interest rates, there never would have been enough scale to make it all work. That it did turned out to be a disaster.