The Committee convened in closed session at the Hay-Adams Hotel at 10:32 a.m. All Committee members were present. Deputy Assistant Secretary (DAS) for Federal Finance Matthew Rutherford and Office of Debt Management Director Karthik Ramanathan welcomed the Committee, introduced the new Chairman Matthew Zames and the new Vice Chairman Ashok Varadhan, and then gave them the charge.
The first item on the charge asked the Committee what adjustments to debt issuance Treasury should make in consideration of its financing needs and uncertainty regarding the fiscal outlook.
Given the cumulative deficit over the next three fiscal years of nearly $3.5 trillion according to OMB, Director Ramanathan stated that Treasury will need to remain extremely agile through its debt management approach and actions to confront challenges related to the fiscal and economic outlook.
Director Ramanathan said that market participants should expect between $1.5 trillion and $2 trillion in nominal and inflation linked issuance again this year; at the same time, bill issuance may marginally decline while shorter dated coupons stabilize at current levels. Treasury debt managers will continue to remain aggressive in managing financing needs while minimizing potential market implications.
As an example, Director Ramanathan outlined Treasury's successful strategy in addressing the $1.4 trillion deficit in fiscal year 2009. Noting the $1.9 trillion in nominal coupon issuance this past fiscal year, Director Ramanathan stated that Treasury was able to raise $1.25 trillion in new cash in tenor beyond two years, nearly six times the amount raised in fiscal year 2008, while at the same time meeting unexpected borrowing needs through bills.
Director Ramanathan pointed out that outlays in FY 2009 were nearly $550 billion higher (an 18% increase) versus fiscal year 2008 while receipts fell by over $400 billion (or 17%), versus the prior year – just short of a $950 billion financing swing in just one year.
Outlays related to fiscal stimulus and financial stability measures were the drivers of expenditures, including TARP related spending of over $300 billion as well as additional spending related to unemployment benefits which are up approximately 150% year over year. Outlays increased across the board, including Medicaid (25% higher), Medicare (10% higher), Social Security benefits (9% higher), and Defense spending (7% higher). Net interest on public debt was lower though by 10%.
At the same time, all receipt categories fell significantly including withheld taxes (7% lower year over year), non-withheld receipts (28% lower), and corporate taxes (55% lower).
Ramanathan noted that as of today, the gross cost basis of Treasury's marketable financing was less than 1.0 % given the large amount of bills issued. Even with nearly $8 trillion in gross issuance across the portfolio, bills averaged 0.16%, notes averaged 1.9% while bonds averaged 4.2% in fiscal year 2009.
Director Ramanathan then turned to OMB's midsession review, which included a table illustrating the effect of budget proposals on projected deficits. The updated 2010 projections' baseline estimates placed the deficit to GDP just above 4% for 2010-2014. Ramanathan noted that any significant variation in debt to GDP or decline in GDP growth in the coming years could increase Treasury's funding cost.
Despite these significant headwinds as well as issuance for Federal Reserve liquidity initiatives, the multi-tiered approach implemented by Treasury to meet these large financing needs ultimately served to stabilize Treasury's average maturity and actually shifted its direction higher. The reintroduction of the 3-year note and 7-year note as well as aggressively moving to two reopening in the 10-year note and 30-year note took place in an extremely compressed period of time, but led to minimal market disruption.
Moreover, if non-marketable debt such as state and local government issuance reverses course or the economy strongly recovers, these increases in coupon sizes could potentially end sooner. Director Ramanathan then explicitly stated that all of these estimates are subject to change given the uncertainty in policy and fiscal expectations, and any shifts would be gradual.
DAS Rutherford continued the presentation to the Committee, outlining cumulative net financing flows since FY 2007 and noting the transition from bill issuance to coupon issuance in greater detail. The large financing needs and lumpy nature of cash flows has led to large cash balances which remain elevated and volatile. DAS Rutherford noted that Supplementary Financing Program (SFP) bills would gradually decline to $15 billion as previously announced from $200 billion to in anticipation of the constraints surrounding the debt ceiling legislation.
DAS Rutherford reviewed the composition of the portfolio, noting that bills as a portion of the debt outstanding fell to about 27%, while bills excluding the SFP program fell to close to historical averages of 23%. DAS Rutherford noted that Treasury will work in close consultation with the Federal Reserve in considering the future path of the program.
Furthermore nominal coupons have risen to 65% from 57% of the portfolio, with particular reliance on the 5-year coupon, while TIPS have fallen to about 8% of the portfolio. Given recent dealer estimates of $1.4 trillion for the fiscal year 2010 deficit and marketable borrowing needs estimates between $1.2 trillion and $1.75 trillion, DAS Rutherford expected the current trends in issuance to continue but cautioned that the outlook remained uncertain as demonstrated by the $550 billion range in marketable borrowing expectations.
DAS Rutherford noted that deficit projections remain unacceptably high and that he expects the FY2011 budget to outline ways to address this. He did inform the Committee, however, that given the number of TIPS coming due early next year and Treasury's sizable borrowing need, TIPS as a proportion of the overall portfolio may continue to decline.
To better understand the reason for the shifts in the composition and profile of the portfolio, DAS Rutherford reviewed the circumstances of last year that sparked the decline in average maturity, noting that the bulk of bill issuance occurring last fall. DAS Rutherford discussed Treasury's maturity profile, noting that over the next 5 years, 73 days will have maturities greater than $20 billion and 46 days will have maturities greater than $30 billion DAS Rutherford noted that approaches to addressing these sizable maturities will be a topic for TBAC discussion in the future.
Looking at the a number of forecasts for the next three years, DAS Rutherford pointed to continued reliance on nominal coupon issuance as well as additional issuance of inflation indexed securities to meet the large borrowing needs while shorter dated issuance eased. By gradually increasing coupons incrementally over the next three years, DAS Rutherford expected the average maturity of the debt to increase back to the historical average of 60 months by fiscal year-end 2010. Eventually, though it could take five to six years, Treasury's marketable debt portfolio will stabilize at a new level between six to seven years.
DAS Rutherford then discussed a change in policy relating to Treasury bill issuance. After repeated rescheduling of the 4- and 52-week auctions this year, Treasury proposed moving all bill auctions to 11:30 a.m. from 1:00 p.m. to minimize conflicts with coupon auctions. DAS Rutherford presented a chart depicting improved coverage ratios in the 4-week bill auctions that occurred at the 11:30 a.m. close.
DAS Rutherford then addressed Treasury's intention to eliminate the 20-year TIPS and reintroduce the 30-year TIPS. Citing an internal ODM report, DAS Rutherford noted zero-coupon inflation swaps data depicts inflation to be upwardly sloping. Assuming that 10-year forward and 20-year forward inflation expectations are not much different, Treasury would be capturing more inflation risk premium by extending issuance from 20-year to 30-year, making 30-year TIPS more cost effective for debt management.
With the presentation complete, DAS Rutherford asked the Committee for its thoughts on debt issuance and the policy changes being considered.
The Committee turned its attention to what changes to the current auction calendar, if any, were needed in order address Treasury's future borrowing needs. Members agreed that at this time, no additional securities to the nominal calendar were necessary, and that gradual increases in coupon sizes would be sufficient to address the large borrowing needs. Any coupon issuance to increase the average maturity should also take place in a gradual manner, and market participants should not be surprised with slow shift.
The Committee opened with a discussion on TIPS and the idea of eliminating the 20-year TIPS in favor of 30-year TIPS. One member began by discussing the September 2009 GAO Study "Treasury Inflation Protected Securities Should Play and Heightened Role in Addressing Debt Management Challenges" (http://www.gao.gov/new.items/d09932.pdf) . One member stated that the study was generally balanced and that the study highlighted reasons why there were market perceptions that the Treasury was not committed to the program.
Another member stated that the GAO study pointed out that there are two potential valid ways of considering the cost of TIPS – an ex-post analysis and ex-ante analysis. Ex-post analysis, over the last 13 years, had shown that TIPS were an expensive form of financing for the government; by other metrics, including asset swaps and auction tails, Treasury was paying a premium to issue TIPS. Another member stated that on an ex-ante basis, TIPS appeared to be less expensive than on an ex-post basis. Another member stated that it would take years to determine if ex-ante analysis is the correct way of measuring TIPS costs.
A member stated that the measure of TIPS cost should perhaps be broadened so as to consider any positive externalities associated with the government accepting the risk to sell inflation. The member stated that investors may perceive the government's willingness to short inflation as a sign that policy makers are confident that inflation is contained. Also, issuing TIPS may be pro-cyclical and serve as a hedge to the government's balance sheet. Another member pointed out, however, that the government currently issues significant amounts of short-dated Treasury bills which are less expensive to the taxpayer and could be considered to be "pro-cyclical issuance". The member noted that the "pro-cyclical benefits" argument for issuing TIPS also breaks down in stagflation environments.
A majority of TBAC members generally believed that despite legitimate concerns surrounding TIPS costs and liquidity, given its funding requirements Treasury would need to increase the size of the TIPS program. In addition, TIPS could help Treasury in its stated goal of extending the average length.
In terms of TIPS issuance, there was general consensus by committee members to eliminate the 20-year TIPS and replace it with 30-year TIPS issuance. This change might allow Treasury to capture a greater inflation-risk premium and would also create a TIPS issue that could be better compared to a comparable on-the-run nominal issuance point. The additional duration associated with a 30 year TIPS would also be consistent with Treasury's desire to increase average maturity.
The committee generally recommended that the overall issuance of TIPS be increased over the next couple of years. For FY2010, TIPS issuance should be increased from the current run rate of $58 billion per year to an overall issuance amount of between $70 and $80 billion per year across securities. In FY2011, overall TIPS issuance should be further increased to between $100 and $125 billion.
In addition, given that TIPS auctions are liquidity events in the TIPS market, Treasury should consider increasing the frequency of TIPS auctions so that there is either a 5-, 10- or 30-year TIPS auction once a month. One member suggested that Treasury could issue a new 5-year TIPS in April with August and November reopenings; a new 30-year TIPS in February with June and October reopenings and new 10-year TIPS in January and July, with March/May and September/November reopenings, respectively. Some members, however objected to increasing the size of the TIPS program so dramatically, citing concerns about costs.
Members recommended that it was important to continue to extend average maturity but there was no need to change the nominal auction calendar to achieve such an increase. One member stated that there was some confusion in the markets regarding how fast Treasury was intending to increase the average maturity and that Treasury should provide some clarity around that issue.
Director Ramanathan reiterated that market participants should expect between $1.5 trillion and $2 trillion in nominal and inflation linked issuance again this year; at the same time, bill issuance may marginally decline while shorter dated coupons stabilize at current levels.
The Committee then turned to a presentation by one of its members on the likely form of the Federal Reserve's exit strategy and the implications for the Treasury's borrowing program resulting from that strategy.
The presenting member began by noting the importance of the exit strategy for financial markets and fiscal authorities. It was noted that the near-zero interest rates driven by current Federal Reserve policy was pushing many financial entities such as pension funds, insurance companies, and endowments further out on the yield curve into longer-dated, riskier asset classes to earn incremental yield. Treasury securities have benefitted from the resultant increase in demand, but riskier assets have benefitted even more. According to the member, the greater decline in the indices for investment grade and high-yield corporate debt relative to 10-year Treasuries and current coupon mortgages displays this reach for yield. A critical issue will be the impact on the riskier asset classes as market interest rates move away from zero.
The presenting member then looked at the likely sequence of the Federal Reserve's exit strategy. The member acknowledged that the central bank must address the uncertainty and fragility of the economic recovery and the dependence of the housing market on low rates. It was suggested that the most likely sequence would be the draining of excess reserves from the banking system, the cessation of the mortgage-backed securities purchase program, and only then raising the Fed funds target rate.
Several members at this point asked why draining reserves before ending the MBS program made sense. The presenting member noted that the program was already set to expire, and other measures, such as a revival of the Supplementary Financing Program, could be utilized by the Federal Reserve at the same time.
The presenting member then addressed the options for draining reserves from the banking system. The problem of excess reserves could persist through the end of 2011 with up to one trillion in excess reserves remaining after liquidity facilities and on balance sheet securities have rolled off. One approach, raising the Fed funds rate to increase the opportunity costs of banks using their reserves, carries the attendant problems of increasing interest rates too soon in the economic recovery. A second option, taking in term deposits, lacks a clear mechanism for rate setting and bank use. Selling assets may run into difficulties if the public appetite for debt at that time is sated, especially considering the impact on the housing market and the major role the Federal Reserve currently plays in the market.
According to the presenting member, these less than optimal solutions leaves the Federal Reserve the option of reverse repurchase agreements (reverse repos) as the most likely option although the potential of the mechanism for draining reserves is unclear. If it is to undertake these reverse repos, the selection of counterparty is important. Depending on how the program is designed, whether it is made to work with dealers or money market funds or to pursue a TALF model with banks as agents, there will be different impacts on the scope of the program, the ease with which it can be set up, and the term of the contracts. In all cases, the program will compete with other short-term investments and put upward pressure on Treasury bill rates according to the presenting member. Moreover, draining excess reserves may dampen the demand for Treasury securities by banks given that banks are investing in securities – particularly Treasuries - in the absence of loan demand.
Several members noted the graph discussing net fixed income supply in 2009 and 2010, and how issuance will ramp up dramatically in 2010. Federal Reserve purchases have taken an enormous amount of supply out of the market this past year across fixed income markets, but next year, financial markets should expect even greater issuance with no support. Such an outcome could pressure rates.
The presenting member then addressed the Treasury market implications of this likely strategy by the Federal Reserve. Using information from the Flow of Funds data and internal projections, the presenting member suggested that the program of Federal Reserve purchases of securities has artificially reduced the supply of fixed income securities coming into the market. According to the member, by contrasting 5-year/5-year forward TIPS expectations with skew data for the 1-year/10-year high and low strikes, the real concern in the market is higher real interest rates rather than inflation.
The presenting member then offered several recommendations to possible market outcomes given the outlined scenario. According to the member, the Treasury should increase TIPS issuance to diversify and broaden its base in light of future competition for market demand. Further, Treasury should extend the average maturity of the Treasury portfolio. These actions must be balanced with the risk that further disinflation or deflation could impact the concentrated TIPS buyer base when it is most needed. The presentation then concluded.
Members generally agreed with the recommendations offered in the presentation as well the potential outcomes from pursuing specific exit strategies in relation to Treasury demand. One member noted that the end of the mortgage purchase program could impact overall rates by 50 to 100 basis points depending on the economic outlook and housing situation. Another member noted that Treasury rates could simply remain stable as other rates fell as the exit strategy took place or if less of an inflationary scenario took place.
The Committee then turned its attention to the third item on the Charge regarding characteristics of Treasury's debt portfolio. Specifically, given recent trends in the economy and the government's fiscal position, the charge asked members to discuss Treasury's plan to lengthen the average maturity of the portfolio in the medium to long term. In addition, Treasury sought the Committee's opinion on the optimal range for average maturity given structural financing needs in the medium and long term. Another Committee member gave the presentation.
The presenting member began by stating the four conclusions of the analysis. Namely:
1. Inflation, higher interest rate and roll over risk should be the primary concerns in Treasury's debt management strategies.
2. In most scenarios, it is prudent to lengthen maturities significantly in a gradual manner from the current average maturity of 50 months. The base case recommends an extension to 74 months, while more pessimistic scenarios suggest an extension to 96 months.
3. The objective of lowest borrowing cost could lead to higher yields that conflict with monetary policy objectives.
4. Clever debt management strategy could potentially reduce debt service cost meaningfully, but still can't completely substitute for prudent fiscal policy.
The presenting member then went on to provide background material before discussing a model developed to help in determining optimal average maturity.
The Committee member noted that in doing a review of G7 countries' debt management strategies, several facts were notable. One point was that the United States had the lowest average maturity. Another point was that the U.S. had the largest percentage of foreign ownership of its debt. The member also noted that while the UK economy appeared to be in similar straits as the United States, it had the highest average maturity of all the G7 countries. Finally, the member noted the general lack of use of an asset-liability management framework for debt management.
The Committee member went on to review charts depicting various characteristics of Treasury's debt portfolio. The member presented charts showing that the current average maturity is lower than it has been in nearly 25 years, that the federal debt to GDP ratio was only higher than it is presently during World War II, and that this ratio is poised to increase significantly according to current Administration and CBO forecasts.
The member stated that much of the increase in the government's expenditures was structural in nature rather than temporary and presented charts indicating that mandatory spending was growing five times faster than discretionary spending. The rapid growth in entitlement spending as a percent of GDP beginning between 2010 and 2020 also remained a major challenge. The member then presented a chart showing an alternative possible budget outlook that indicated that the debt to GDP ratio could grow to as high as 98 percent by 2019.
The member continued by noting that the budget deficit has benefited from low rates but cautioned that this could change in the future since approximately 40 percent of the debt will need to be refinanced in less than one year.
Director Ramanathan noted that a significant portion of the shortening of the average debt was related to over $1 trillion in Treasury bill issuance as a result of Federal Reserve liquidity initiatives, fiscal stimulus, and financial stability measures. Director Ramanathan noted that the transition from bill financing to coupon financing was in process, but would not take place in an abrupt manner.
The member ended the background discussion with a chart that warned that historically, large fiscal expansions that were coupled with debt monetization could lead to inflation.
The presenting member went on to discuss a model developed to help in determining the optimal average maturity of debt issuance and began by issuing a disclaimer that the model was a stylized model meant to aid in how to think about the problem rather than to determine the actual optimal average maturity. The member also indicated that the model was developed over a very short period of time and therefore some simplifications were necessary, including using only 3-month and 10-year securities to finance the debt and excluding factors that would likely affect the output of the model.
The model projects funding needs across 15 economic and credit scenarios over the next ten years and attempts to find the optimal average maturity of debt issuance given different risk scenarios over the next three years in order to minimize the total debt cost of debt service over the ten-year period. The model employs assumptions for the determinants of the 3-month and 10-year rates. The presentation focused on four scenarios a base case, a low growth, a low inflation case similar to Japan, a moderate growth, high inflation case and, finally, a high credit loss case.
The results from the model indicate that in the low growth, low inflation scenario average maturity of issuance should be as low as possible, while in the high inflation scenario, average maturity should be extended to lock in low rates.
The member said that the model indicated that the macroeconomic environment had the most impact on average maturity. The model indicates that real growth of 2% combined with inflation of 2% results in an optimal average maturity of issuance of 55 months, while 2% real growth combined with 5% inflation increased the optimal average maturity to 116 months. In contrast, given a 0% real growth and 0% inflation environment, credit losses of $575 billion lead to a debt maturity of 26 months and losses of $1.4 trillion only increase the optimal debt maturity of issuance to 34 months. Interestingly, optimal average maturity of debt issuance is not significantly reduced by an increase in tax receipts of 30%.
The model results are highly dependent on the impact of duration supply on yields and, all else equal, if issuing more long-dated debt has a larger impact on rates, the optimal average maturity of issuance will be shorter.
The member noted that the lowest cost strategy for optimizing average maturity may lead to yields that conflict with monetary policy goals, and that constraining near-term yields would lead to shorter average maturity.
The member noted that the model indicated that clever debt management could reduce costs by a surprisingly high 13% of government revenues in a high credit loss, high inflation scenario by issuing long-dated debt from 2009 through 2011. However, the member also noted that even with optimal debt maturity, debt service cost would be unbearable and that prudent fiscal policy was needed to bring debt service costs under control.
The presentation ended with the member noting that there were many areas that could me more fully explored. The model did not fully consider entitlements and state and local government as potential contingent liabilities. And, therefore, the risk to the model is to the upside. The model can be enhanced on duration supply going forward. The member noted that current literature is focused on historical regression.
The Committee members all agreed that the Treasury should extend the average maturity of the debt an indicated that the presentation just solidified that view. Members however noted that this process would not take place in a very short period, but may take time to occur over several years.
_________________________________
Karthik Ramanathan
Director, Office of Debt Management
United States Department of the Treasury
