December 1, 2005
js-3029

Remarks by Treasury Debt Management Director Jeff Huther before the Bond Market Association New York

I would like to talk a little about our proposed securities lending facility, the issues that make it a complicated proposal and, where possible, clear up mis-conceptions about the impact of such a facility.   I should start by emphasizing that this is very early in the proposal process.  We are still evaluating how a facility would work and what the consequences of a facility would be.  It is possible that we will conclude, in the end, that we cannot proceed with implementation because the risk is too great that the facility would adversely affect Treasury markets.    

In broad terms, we have set for ourselves the goal of obtaining least cost financing for taxpayers.  For almost 30 years, Treasury has tried to meet this goal by, among other things, providing market participants with certainty of supply of Treasury securities in primary market offerings.  It has been our assumption that greater certainty of supply reduces the market risk of bidding for Treasury securities and thereby lowers borrowing costs. 

In general, certainty of supply in the primary market has translated directly into a high degree of certainty of supply across the secondary markets: cash, financing, and futures.  In recent years, however, there have been occasional, isolated periods where the availability of specific issues in the financing markets has been so acutely impaired relative to demand as to call into question the ability of Treasuries to provide the liquidity and risk management functions that also contribute to lower borrowing costs.

Our proposal to create a securities lender-of-last-resort facility stems from our recognition that availability of supply in the financing market is as important as certainty of supply in the primary market for meeting our goal of least cost financing over time.  The proposal is based on the idea that the taxpayer is the ultimate loser if market participants lose confidence that Treasury market transactions will not always be settled in a timely manner due to an acute shortage of specific issues. 

The foregoing idea is, however, insufficient to create securities lending facility – and poor design could damage markets that work extraordinarily well almost all of the time.  Some characteristics of a facility that avoid damage are easy to identify: borrowing Treasuries securities from the Treasury should be less attractive than borrowing from the private sector (as long as the securities are actually available from the private sector) and use of the facility should be at the discretion of the borrower rather than Treasury officials.  In addition the facility should not impose substantially greater costs on users relative to other contemporaneous private market borrowers.  Other characteristics are likely to be determined by practical considerations: who can access the facility, whether the facility should be bonds-for-bonds or bonds-for-cash, and the time of day that the facility can be accessed. 

Whether a securities lending facility will avoid damage, however, depends most importantly on the pricing characteristics of the facility.  There are two factors that complicate the pricing.  First is the nature of the Treasury market – we auction a fixed quantity of securities and at any given time those securities provide the basis for a larger quantity of long positions with the excess offset by corresponding quantity of short positions.  The quantity of short positions that can be supported at any one time is growing but still finite.  Under some conditions, such as those prevailing in the summer of 2003, the demand for short positions by private market participants can outstrip the capacity of the market.  Fixing an upper limit on the quantity of additional securities available from Treasury may only give some market participants an incentive to use the facility strategically.  The implication is that pricing within a securities lending facility cannot be tied to either current market conditions or a fixed quantity of securities. 

The second complication is the nature of the fails penalty within a repo contract.  Treasury markets have grown faster than the underlying supply of Treasuries for years and market participants have generally found ways to increase the velocity of Treasuries to ensure adequate supply for settlement purposes.  The chronic fails of the past few years have been a result of the market's inability to establish a market-clearing yield on special collateral repurchase agreements in low interest rate environments.  In trying to set up a workable pricing mechanism the complication is that, in the absence of a well-functioning market, we do not know where the market-clearing yield would have settled.  Clearly there is a need, assuming Treasury establishes a securities lending facility, for market participants to review the fails penalty in outright sales and both legs in repo contracts with a goal of developing a market-clearing pricing mechanism that works in low interest rate environments.  The implication for us is that, when proposing a pricing mechanism in the absence of good information about what a market-clearing yield would have been, we should think in terms of a structure that can evolve with changes in market conditions or contract design. 

As we work through these issues, we will continue to seek your views on how to establish pricing characteristics that do not reduce the existing certainty of supply of Treasuries.  The discussions that we have had with market participants so far have been very helpful in working through the characteristics of a facility.  

Internally we are working on the characteristics of the facility that pose fewer complications, but still require careful consideration.  We are considering a bond-for-bond transaction structure – like that used in the Federal Reserve's security lending facility - because it allows for better cash management by Treasury.  Initial feedback suggests that market participants do not see a problem with a bond-for-bond structure. 

Our goal of providing additional supply to the market to alleviate severe shortages points toward broad access to the facility.  In the best case scenario, Treasury would like the facility to be available to as many market participants as possible.  However, operational considerations come into play given the complexities of RP transactions.  A logical group of counterparties would be the 22 primary dealers, given their role in the Treasury market as market-makers and liquidity providers.  However, Treasury must be sure that any additional supply it provides reaches the broader market.

Treasury would be acting as a lender of last resort.  This means we want market participants to access all the available private supply and to use the New York Fed's securities lending program before coming to Treasury to borrow additional securities.  We do not seek to be a part of the normal market for lending and borrowing, but to have market participants access the facility only when all other sources of supply have been tapped out.  Because of this, we are considering use of the facility later in the day, after the Fed's securities lending auction at noon.  This would give the market a chance to know what they still needed to borrow.  However, there are operational considerations and concerns with getting transactions completed before the close of Fedwire that will not allow us to move the access too late in the day.  And, depending on the settlement structure, the time of day the facility is available may be of less importance.

Let me conclude by describing what a proposed facility would not do.  We view this facility as truly a lender-of-last-resort facility.  If we cannot make the facility uneconomic under normal market conditions, we will not proceed.  If we would reduce the incentives for financial intermediaries to make markets in Treasuries, discourage the use of Treasuries for hedging, or lead market participants to express their views on interest rates in other markets, we will not proceed.  And, if we do find a pricing mechanism that works, it is easy to imagine a facility that lays dormant for many years at a time under a wide range of interest rate environments – an outcome that we would view favorably.  

-30-