The $700 Billion U.S. Funding Hole; Desperately Seeking A Very Indiscriminate Treasury Buyer
A month ago we observed that in 2010, the supply/demand picture for US fixed income would be very problematic, as there was no immediate apparent substitute to fill the void resulting from the departure of the constant bid provided by the Federal Reserve's Quantitative Easing in both the UST and the MBS markets. The conclusion was that there would need to be a dramatic increase in demand for debt securities across the board, with an emphasis of Treasuries and MBS.
Today, we focus on the most critical segment of debt issuance for 2010 - those ever critical US Treasuries, without whose weekly uptake by various investors, the multitrillion budget deficit will become unfundable. Using estimates from Morgan Stanley for 2010 Treasury supply and demand, the conclusion is that there will be a demand shortfall of at least half a trillion, and realistically $700 billion, to satisfy the roughly $1.7 trillion in net ($2.4 trillion gross) coupon issuance in the upcoming year.
The implication is that back end prices will decline sharply due to an ever increasing supply overhang, even as nearly $800 billion in Bills are paid down, thereby further accentuating the steepness of the bond curve. And with ever more emphasis put on the coupon supply, the marginal yield on long-dated Treasuries will likely find it needs to be increasingly more attractive to find bidders, which in turn will jar mortgage rates out of hibernation. We are now certain that Q.E. will continue: the weakness in the mortgage backed-market is already becoming a topic of contention, and when it becomes apparent that there is an additional $700 billion demand void in Treasuries, then it is merely a matter of time before Ben (or his successor) realizes the dollar destruction comeback tour has to resume asap. Those cynically inclined may wonder why Bernanke's reconfirmation should take place prior to any potential Q.E. 2 announcement. Perhaps this country's Senators would further evaluate their support of the Chairman once they experience the popular anger which will accompany the next leg down in the US currency the minute Mr. Bernanke announces that the Fed will need to continue being the market in treasuries and mortgage backed securities, further eroding the collateral behind the greenback.
First, based on Morgan Stanley's expectations, and further corroborated by yesterday's disclosure that the next increase in the debt ceiling by $1.9 trillion net, to $14.3 trillion, would last the country only through early 2011, we present the estimated supply of gross coupon issuance in the upcoming fiscal year (keep in mind one quarter of issuance has already been absorbed and the run-rate validates the projections).
After issuing a $1.9 trillion gross amount of coupons in F2009, in 2010 this amount is expected to increase by 30% to $2.4 trillion, with an emphasis on long-dated maturities: per the chart above, the average age of new gross coupon issuance (excluding Bill impact) will increase from 5.9 years to 6.4 years in 2010.
In 2010, net issuance will be substantially lower than gross according to MS, due to an increase in maturities, and "only" $1.7 trillion in net new coupon bonds, $425 billion more than 2009, are expected to be issued by the US Treasury: this number may well be an underestimation as the Senate, which likely has far more granular issuance projections, is calling for $1.9 trillion in net issues (in addition to the $300 billion temporary increase which passed late last year) which would fund the US budget for about a year. One offsetting feature of net issuance in 2010 will be a surge in paydowns in Bills, which are expected to be a net negative contributor to issuance to the tune of $775 billion (of which $275 billion has already taken place in Q1 of fiscal 2010, primarily as a function of the $195 billion in SFP bills rolling off).
So far so good- the supply picture is clear, and in reality the final amount will probably end up being substantially higher than $1.7 trillion net, as the runaway deficit-creating machine in D.C. will stop at nothing to prove that any one failed auction will destroy this country.
Where things get tricky is on the demand side.
As we pointed out previously, the number one defining feature of 2009 was the Fed's blatant support of the bond and MBS markets. Bernanke monetized $300 billion in Treasuries, and indirectly will have purchased another $1.4 trillion in bond/MBS hybrids (we say indirectly, because Fed MBS purchases effectively allowed MBS holders to switch their holdings to Treasuries at preferential terms, better known as the "reallocation trade" in essence achieving the same effect as if the Fed has purchased these - see Bill Gross). With the Fed out of the demand picture (at least temporarily), the questionmarks emerge.
Combining the supply and demand for Treasuries yields the following chart. Fact: in 2010, a best case of demand projections, indicates there will be a $400 billion shortfall for total Treasury supply... and a worst case of a stunning $700 billion funding shortfall. This is "just" a little worse than Greece, yet the latter's CDS trades trades nearly ten times wider than the U.S. Logical? You decide.
The key variable in this exercise is quantized and overall demand, which is why a detailed analysis of each end segment must be performed to understand the demand mechanics.
On December 31, 2009, a majority of U.S. debt (marketable Bill, Coupons, TIPS) was held by foreigners, making America a net foreign creditor nation. Compare this with Japan, where 93% of sovereign bonds are held by domestic accounts. Yet over the past several years, the US has become increasingly reliant on foreign generosity: foreign demand has grown from $143 billion in 2007 to $794 billion in 2009. And even as foreigners have purchased an increasingly greater amount in absolute terms, the relative composition has in fact declined in the past year: foreign demand dropped from 76% in 2008 to 46% in 2009.
An even more granular analysis of foreign purchases, indicates that as foreigners rushed into the safety of Bills, demand for coupons actually declined. Also notable is that foreign demand for coupons has never moved too far, and has stayed in the range of $192-$370 billion each year.
The biggest problem this data indicates is that foreign demand will not go willingly with the Treasury's demand to extend the average Treasury maturity from 4 to 7 years: foreigners purchased 145% of the Fiscal 2008 net issuance of $255 and a meager 26% of the Fiscal 2009 of $1,271 billion.
And herein lies the rub, as MS points out, the foreign bid is usually a direct function of the amount of global trade and the associated trade gap experienced by the U.S. Historically, the excess trade gap was not an issue, as China, Japan and net exporter partners had to recycle their otherwise useless dollars back in the U.S., and they did so by purchasing U.S. bonds, thereby allowing U.S. consumers to borrow ever cheaper and to purchase yet more Chinese and Japanese trinkets, rinse, repeat.
As Zero Hedge pointed out some time ago, the deputy governor of the PBoC, Zhu Min, said the most logical, yet scariest, thing for the US Treasury.
"The United States cannot force foreign governments to increase their
holdings of Treasuries," Zhu said, according to an audio recording of
his remarks. "Double the holdings? It is definitely impossible."
US current account deficit is falling as residents' savings increase,
so its trade turnover is falling, which means the US is supplying fewer
dollars to the rest of the world," he added. "The world does not have
so much money to buy more US Treasuries."
Zero Hedge has previously demonstrated the problem associated with China's trade surplus, which while still positive, saw a significant drop from the prior year. And compounding this is the concern that while China is still accumulating FX reserves, it may now be diversifying its US-denominated holdings. Yet setting diversification concerns aside, the bigger picture indicates that China UST purchases usually are a function of FX reserves: should the US continue on the recent protectionist path, this will implicitly make Chinese demand for Treasuries even scarcer.
Based purely on global trade surplus/deficits, it is likely that the foreign bid would purchase $300-$400 in coupon Treasuries in 2010. However, in evaluating foreign demand in 2010 one has to consider the Bill/MBS reallocation trade. A big question mark for 2010 will be whether foreigners will reinvest Bill holdings purchased at an above average rate in 2008 and 2009 (see Foreign Bill Vs Coupon Purchases). The demand for Bills occurred due to reallocation away from Agencies/MBS and corporates, which can be seen from the below chart. Here it becomes visible why the Fed's MBS program was the practical equivalent of a Treasury QE extension. The Fed was acquiring foreigners' MBS and Agencies at prices that would allow them to buy Bills (and sometimes Coupons) in kind.
With non-Fed demand for MBS still non-existent (and, in fact, everyone selling into the Fed's bid), and a reduced issuance of Bills in 2010, it remains to be seen what assets foreigners will reallocate to. This "reallocation" trade will likely add another $200 billion to the $300-400 billion estimated above, thus bringing total demand for Coupons to $500-600 billion, offset by a Bill outflow of $300-400 billion.
Recently the "Household" sector as defined in the Federal Reserve's Flow of Funds, attained some notoriety after, as Zero Hedge disclosed first, Eric Sprott brought up allegations of covert monetization and general ponziness by the Fed via the "Household" sector. We will stay away from semantics, and present what is known: at the end of 2009, "households" held 12% of Treasury debt, or $800 billion: less than a quarter of Foreign holdings of $3.6 trillion. The inappropriately-named household sector consists of individual households, nonprofits, hedge funds, private equty, private foundations, labor unions and others, and Treasury holdings allocated to it, are calculated as a differential between total USTs outstanding and known amounts held by other investors. Basically, it serves as a plug to "everything else."
Regardless of semantics, a critical point must be added to the Sprott analysis, and also to Goldman's optimistic outlook on bonds, which is predicated on increased household purchases. As a reminder, Goldman speculates:
Increased saving by households and businesses creates a
potential demand for Treasury securities as well as less competition
for lenders' funds; flow of funds data and bank balance sheet reports
confirm that the domestic private sector is increasing its allocation
to Treasury securities.
Is Goldman overly optimistic on their expectation that U.S. households will finally do what their Japanese equivalents have been doing for decades? The answer is yes. But before we get into this, we need to point out that the recent surge in "Household" buying has not been effected in one bit by actual households and individual investors. Why is this? After all, the household savings rate has increased from 0.8% in April 2008 to 4.8% in December 2009. Yet as a reminder, the two key components of Household Treasury holdings include Savings Bonds, which are what households actually buy when they wish to purchase government debt, and Other Treasuries, which are marketable Treasuries, and which average households have no access to. It is a notable observation, that while the savings rate has indeed increased, holdings of savings bonds have not only stayed flat, but have declined over the past year: this is perfectly explainable by the combination of an increasing savings mentality coupled with a desire to deleverage: i.e., Rosenberg's new frugal normal. Goldman, which has bet the house on household Treasury purchasing to keep rates low, will be disappointed.
The chart below demonstrates the historical holdings progression between Savings Bonds and Other Treasuries. As can be seen, actual households have not been active purchasers at all in the recent bond buying spree.
Yet while it will take much more to convince Goldman in its faulty assumptions, what is without doubt, is that the same "reallocation" trade that has taken place in Foreign purchasing, has been paralleled in the Household sector. As the chart below shows, while Treasury holdings have surged over the past year, this has been purely a function of a collapse in Agency/MBS holdings. In fact, in the past year, MBS holdings in the Household category have fallen by a stunning $772 billion, from $840 billion a year ago to just $68 billion most recently. This has been accompanied by a less than half increase in Treasuries in the last 12 months: from $493 billion to $860 billion, a $367 billion increase, and less than half the decline in MBS.
Just like the reallocation trade has been critical to spur demand in foreign purchasers for USTs as they have rotated out MBS with the Fed lifting any and all foreign offers, so has the Fed been busy domestically. Comparing the action over the past 3 years, from the peak of the housing bubble (2006-2009 period), indicates that the reallocation trade accounts for nearly a dollar-for-dollar move out of MBS, which declined by $352 billion from $420 billion to $68 billion, into Treasuries, which in turn increased by $344 billion, from $516 billion to $860 billion.
With just $68 billion left in Households' MBS holdings, the reallocation is over, which means that the household sector will no longer be a major purchaser of Treasuries, and all of this on the backdrop of actual consumers, whose Saving Bonds holdings have dropped from $197 billion to $192 billion over the past two years. On the other hand, should "Households" end up purchasing substantially more than expected, then the Sprott thesis will have to be seriously revisited.
A major wildcard for 2010 Treasury demand will come from commercial banks, whose $1+ trillion in excess reserves, courtesy of flawed monetary policy, may be used if not to spur consumer lending, then at least to acquire treasuries. As was shown previously, banks held only $200 billion in Treasuries at the end of 2009, making them the second to last holder, yet the massive dry powder on their books, as well as possible political prerogatives, will likely make this sector a major purchaser of Treasuries.
Empirically, banks add to their Treasury holdings at the end of recessions, when banks have capital to allocate, yet consumer and small-business loan opportunities remain weak. This can be seen on the chart below:
This is also evident when one considers that change in bank UST holdings, compared against the steepness of the yield curve: it makes all the sense in the world that banks would increase Treasury holdings in a steep yield curve environment.
Yet even if banks unleash the full power of their excess reserve holdings it will likely not do much for back end supply. The reason is that banks traditionally purchase USTs in the 2-4 year sector, as they get most of their duration via their mortgage holdings, and with rising rates, existing duration has grown. As banks receive much better returns by lending direct, moves along the curve are i) rare and ii) merely placeholder measures until the economy improves, which explains their unwillingness to stray far on the back end of the curve.
The reason why this may be problematic is that there is an incremental $350 billion in new gross issuance in the 5Y - 30Y part of the curve alone, which is precisely the part that is least attractive to the banking sector.
Another major concern to banks is the prevalent uncertainty about possible future inflation: the Fed's liquidity spigot is as worrying to banks as it is to all but the staunchest deflationists. Today, inflation uncertainty is near decade highs. Furthermore, even as 10 Year yields remain near all-time lows, 10 Y inflation expectations are rising fast.
In order to determine the pace of Treasury purchases, a comparison with prior recessions (including those of the 1970s, 1980s, and 1990s) indicates that following major recessions, banks increase their UST holdings by 1.6% to 2.7% of total assets (with an average increase of 2.2%), and this increase takes two years on average.
Of the $16.9 trillion in total banking assets as of June 2009, USTs accounted for $141 billion or 0.8%. Growing this number in value to 2.2% of projected total bank assets of $18.6 trillion in June 2011 in the low case, and 2.7% in the bear case, implies that between $421 and $513 billion in Treasuries would have to be purchased over the next two years. As $82 billion was purchased in H2 2009, this implies banks need to ramp up purchases to $225 billion per year in the low case, or about $4 billion per week. This represents about one-third the pace with which the Fed was monetizing/buying back bonds in 2009. The high case corresponds to an annual pace of $287 billion per year, or $6 billion per week: about half of the Fed's rate of purchases. Both cases, as noted above, would focus on Treasuries in the front-end of the curve.
As a result, it is expected that banks will purchase between $190 and $240 billion in Treasuries in 2010, which number also includes the $24 billion already purchased by domestic banks in Q1.
Broker dealers, unlike the other mentioned purchasers, do not have an outright preference for Treasuries as a yielding instrument, but merely as a hedge for spread-product books (including corporates, CDS, MBS and agencies). As banks deleveraged in 2008 and 2009, they covered massive amounts of UST shorts as they sold off the underlying hedged securities. Indeed, in Fiscal 2009, B/Ds purchased a record $119 billion of Treasuries, following $86 billion in 2008. Not surprisingly, these coupon purchases occurred in the front end of the curve (1Y-5Y), again indicating B/D's aversion toward dated paper.
As of June 2009, the B/D deleveraging process appears to have ended, and in fact has reversed as leveraging has once again commenced: B/Ds sold $21 billion of Treasuries in Q4 2009. Therefore, Broker Dealers are expected to sell $25-50 billion in coupons in 2010.
Insurance funds are essentially banks-lite: they prefer to purchase treasuries in a steep yield curve environment. In 2009, insurance/pension funds were the fifth largest buyer of Treasuries ($56 billion from insurance firms and $37 billion from pension funds). With expectations of a steep yield curve (for now) likely staying in the 270-280 bps range, Insurance funds are expected to purchase about $100-$150 billion.
An upper ceiling to purchases is likely to come from the discount rate on defined benefit pension plans (around 6.5%), implying the yield on purchased Treasuries has to be at least 5.25%. Currently the highest yielding P-STRIPS in the 30 year sector offer just 5%. Insurance companies could very well become a purchasing force... however at materially lower levels.
In 2009 the fixed income mutual fund/ETF space saw unprecedented activity: doubling the $104 billion in 2008 inflows (2009 closed at $204 billion). Yet the vast majority of this amount went to chase higher-yielding, riskier assets: only $33 billion (16%) was allocated for UST purchases.
As allocation to these buyers seeks to outperform benchmarks, the allocation to USTs has traditionally stayed limited, and as a result 2010 demand from mutual funds/ETFs is expected to stay in line at around $50-75 billion. Furthermore, as has been repeatedly pointed out, equity inflows have been negative in 2009. If there is a reallocation trade whereby investors seek even riskier assets, 2009 could see a rotation out of broad fixed income into equities and even riskier assets (CDOs are already stirring).
Money Market Mutual Funds
As money-markets only invest in ultra-short dated Treasury products, this demand category would not have an impact on the back-end. Furthermore, money-markets will probably continue to unwind the $332 billion in front-end paper purchased in Fiscal 2008: already last year $34 billion in Bills and short-end coupons was sold. If ZIRP persists, and if the Volcker doctrine manages to make money markets sufficiently unattractive, this category will at best have a neutral impact on USTs and more realistically will continue to be a net seller. As such, in 2010 this segment is expected to sell $100-200 billion in Bills and front-end paper.
Municipalities are expected to provide a token amount of demand, to the tune of $25-50 billion: this source of demand is low in a rising rate environment. In 2009 only $3 billion in demand came from money market funds.
Up to this point, we have demonstrated that under realistic assumptions, the traditional buyers of Treasuries will be insufficient to plug the demand hole. As the Fed will not sell any of the roughly $770 billion in Treasuries on its balance sheet with a ZIRP policy still in place, the only question is whether Ben Bernanke will step in and roll out QE 2. Of course, the implications to the stock and currency markets will be drastic should the Fed relapse to its old financial heroin-dispersing ways.
While near end supply will likely not be as difficult to satisfy, the back-end will face increasing yield pressure in order to stimulate demand. This means that long yields will begin a slow trickle higher to attract the missing demand that currently is unaccounted for. Should this happen, and should the likes of Morgan Stanley be correct in expecting even further steepening, the implications on mortgages will likely be severe. Which is why we are confident that the Fed, which is all too aware that the economic situation is far worse than what is presented in the mainstream media, will expand quantitative easing not only to more MBS purchases (mostly to facilitate yet more reallocation trades), but to direct Treasury purchases once again. In doing so, the Fed will surely short-circuit the market beyond all repair.
A practical idea on how to approach this binary outcome, would be the implementation of the kind of barbell trade that has made John Paulson a billionaire: should the Fed announce QE 2, the dollar will plunge, and gold will surge. Due to negative convexity between these two asset classes, we anticipate a non-linear acceleration in the price of gold compared to the DXY. Alternatively, should the Fed stay pat and do nothing to prevent the verticalization in the yield curve, the other side of the barbell would be to reward those who would benefit the most from the resultant even greater curve steepness, expressing this with long financial exposure (the more levered, the better). Another levered way to play the increasing curve steepness would be putting on the Julian Robertson-proposed Constant Maturity Swap trade (discussed previously in depth here).
Lastly, should the Fed attempt to stimulate an endogenous flight to safety and boost demand for Coupons artificially, we believe, as we have said before, that the FRBNY will certainly implement a stock market crash. The alternatives, an interest rate hike and QE. We believe that while the probability of QE 2 is increasing with every day, the likelihood of a rate raise is negligible, leaving the market crash theory as the wildcard. We will not handicap this outcome and instead let every reader decide for themselves. Nonetheless, as this week demonstrated all too well, once the market gains downward momentum, even the much expected daily offer-lifters may be mysteriously elusive. Hedge appropriately.