I would like to focus this morning on how we at Treasury look at chronic settlement fails such as occurred last year, and to discuss initiatives aimed at mitigating or preventing a repeat performance.

The large and persistent fails in the May 2013 10-year note can be described as a consequence of either too little supply or too much demand.  Looking first at the supply side, some market participants would argue that Treasury did not sell enough of the note to satisfy demand – although at $18 billion the issue was equal to the largest 10-year auction ever.  Others might cite the reportedly large holdings of foreign central banks – although their take-down in the auction did not stand out as exceptional – or the reluctance of some owners of the issue to lend their securities in the collateral market.  On the demand side, commentators have cited the enormous hedging needs of mortgage and other fixed income investors who were caught offsides when expectations for the future direction of interest rates changed in late June. 

But whether you focus on too little supply or too much demand, the striking fact is that the collateral market in the May 10-year note failed to clear.  Borrowing demand drove the specials rate for the note down to zero in early July and excess demand at the zero rate then spilled over into settlement fails - forcing buyers to become involuntary lenders of the issue rather than holders.

In thinking about initiatives to mitigate or prevent chronic fails, I think it is important to focus on the functioning of the collateral market largely for practical reasons: we are unlikely to see basic changes in the forces driving either supply or demand in this sector of the Treasury curve.  Treasury is unlikely to abandon its commitment to providing you with offerings as regularly and predictably as possible because we view “regular and predictable” as a key ingredient to achieving the more basic objective of financing the federal deficit at lowest possible long-run cost; foreign central banks are unlikely to stop buying Treasuries; and the hedging requirements of fixed income investors are unlikely to shrink appreciably in the future.  

The heart of the pricing problem last year was, unquestionably, the low federal funds rate and the consequently low ceiling on the cost of financing a short position.  I would argue that the low ceiling fostered complacency about hedging and short selling.  In particular, market participants didn’t have to face the prospect that they might have to finance short positions at specials rates 4 or 5 percent below the general collateral repo rate.  This may have left them willing to carry larger long positions in the first place and may have increased the attractiveness of hedging (rather than liquidating) their long positions when intermediate term debt yields began to rise in late June.  This would account for the unusually large short base in the 10-year note and the concomitantly large borrowing demand for the note.

After the specials rate for the May 10-year note hit zero in early July and demand to borrow the note spilled over into fails, some market participants took comfort in the idea that Treasury would bail out the shorts by creating additional supply.  Certainly we at Treasury received ample advice to that effect.  But reopening the note, while possibly advantageous to Treasury in the short run – why not sell more of what is in demand? - would only exacerbate the moral hazard that market participants will come to expect us to fashion debt management policy to accommodate the collateral market.  It would introduce a new speculative component into securities prices and lead to unwelcome pressures on debt managers from shorts who want an issue reopened and longs who do not.

The rock-bottom specials rate on the May 10-year note, coupled with speculation that Treasury might reopen the issue, also led to some particularly dysfunctional discretionary behavior in the collateral market.  Market participants could speculate, cheaply or even costlessly, against an increase in the specials rate by entering into a term repo contract to lend the 10-year note even if they did not own the security and were unlikely to fulfill their delivery obligations.  Because the repo rate was at or near zero, a putative collateral lender would owe a borrower little or no interest if the lender failed to deliver the note.  If Treasury reopened the note and the specials rate rose, the lender could reverse in the security – earning a positive return on the cash it lent – and fund the loan by delivering the security on the original, low or zero rate, repo.  The availability of such a “free (or low cost) option” added to the fails problem at a singularly inopportune time.

The current regulatory architecture provides two responses to a situation like last year in the Treasury market.  One response requires buy-ins of fails outstanding for more than 30 calendar days. Last summer, numerous firms asked for relief from the buy-in rule, arguing that strict application would damage customer relations and would not, in any case, contribute to a reduction in aggregate fails.  Indeed, while the buy-in rule seems appropriate for curing an idiosyncratic aged fail by a customer to a dealer when the dealer can actually obtain the security in a conventional transaction with an alternative seller, strict application of the rule may be inappropriate when transactions are generally failing to settle on a widespread basis.  Little is gained if the buy-in itself fails to settle.

The other regulatory response to chronic fails is capital charges.  Here the case for the response is clearer: settlement fails create risk for both buyers and sellers.  The prospect that a counterparty might ultimately default on a purchase or sale obligation suggests prudent reserves should be established when the obligation has been outstanding an inordinately long time.  This is true regardless of whether an aged fail is an idiosyncratic condition or part of a more widespread condition.  Additionally, capital charges resulting from chronic – widespread and persistent – fails soak up dealer capital that might otherwise be used to support profit-making activities; thereby focusing managerial attention on the underlying fails problem and incentivizing managers to remedy the problem.

We at Treasury, in conjunction with the BMA, the SEC, the NASD and the NYSE, have been discussing ways of addressing chronic fails through the regulatory process, including an examination of the Treasury buy-in rule.  However, you should not conclude that just because we recognize that the current regulations may not address the problem in an optimal fashion that we therefore know what regulations would do the job.  We welcome your thoughts and suggestions.

To a degree, I have already hinted at what I think is the appropriate response to last summer’s events.  The problem began with the complacency that preceded the increase in intermediate term rates and continued with the short hedging that followed the rate increase.  This complacency was a consequence of the low ceiling on the cost of financing a short position; a ceiling created by the market convention that a seller who fails to deliver on a timely basis suffers no direct cost beyond the loss of the time value of the proceeds of its sale.  My assertion is that, had the cost of failing been 5 percent per annum instead of 1 percent, so that the specials rate could have been driven to 500 basis points below the general collateral rate, market participants would have been more cautious in acquiring long positions, more cautious using the 10-year note as a hedging vehicle, and more aggressive in covering their shorts before August.  Hedging needs would still have been substantial, but it is likely that some investors would have sought alternative hedging vehicles.  We might still have had a fails problem, but it likely would have been smaller and resolved more quickly. 

Avoiding market failure in the collateral markets when short-term rates are low requires a market structure that facilitates the seamless movement of the specials rate through zero when borrowing demand exceeds supply at a rate of zero.  Aside from the contractual, documentary, and operational aspects of such a structure, this requires something that imposes a cost (over and above loss of the time value of the proceeds of a sale) on those who do not deliver securities on a timely basis.  Frankly, I do not know what an appropriate cost might be.  It could be regulatory in nature – perhaps a penalty fee assessed on a seller who has failed to deliver securities for some time – or it could result from the explicit terms of the contract between the buyer and seller.  In either case, I look to you for guidance on how to proceed.  Clearly, we don’t want a repeat of last summer.  Avoiding a situation like that, I believe, requires action now.