The Fed giveth and the Treasury taketh away.
Last September, the Federal Reserve began implementing a program to sell $400 billion of its holdings of Treasury securities of 3-years duration and shorter and use the proceeds to invest in 6- to 30-year securities instead. This June the Fed announced that it would do the same with an additional $267 billion in securities. Some analysts expect the Fed to announce additional related measures after the next FOMC meeting.
The theory behind such measures is discussed in this statement from the Federal Reserve:
By reducing the supply of longer-term Treasury securities in the market, this action should put downward pressure on longer-term interest rates, including rates on financial assets that investors consider to be close substitutes for longer-term Treasury securities. The reduction in longer-term interest rates, in turn, will contribute to a broad easing in financial market conditions that will provide additional support for the economic recovery.
One could thus view the Maturity Extension Program as a deliberate effort by the Fed to undo the consequences of previous Treasury decisions for the composition of debt held by the public. A number of academic studies have looked at whether changes in the maturity composition of publicly held Treasury debt really do affect bond yields. Gagnon, et. al. (2010), Greenwood and Vayanos (2010), Swanson (2011), and
Hamilton and Wu (2012) all found that historical changes in the maturity composition are correlated with small changes in long-term yields. Studies of the particular effects of the Fed’s recent operations include D’Amico and King (2010), Hancock and Passmore (2011), Krishnamurthy and Vissing-Jorgensen (2011) and Wright (2011), all of whom found effects consistent with the theory.
I was interested to learn this week that the U.S. Treasury in turn may be deliberately trying to undo the consequences of the Fed’s Maturity Extension Program. David Beckworth calls attention to the Office of Debt Management Fiscal Year 2012 Q3 report. In trying to understand some of these charts, I was led to the minutes of the latest meeting of the Treasury Borrowing Advisory Committee, where I was surprised to read this:
The presenting member noted that the original MEP lowered Treasury’s borrowing capacity through 2016 by $400 billion and that the extension of MEP lowered it by an additional $267 billion.
I would love to find more explanation of what this statement means (and welcome instruction from any readers knowledgeable of the institutional details). Here is my best effort at understanding it. Suppose that the Treasury owes $400 B in 10-year debt which has been issued evenly over the last 10 years. Then each year it will need to auction $40 billion in new 10-year securities. Alternatively, if the Treasury owes $400 B in 1-year debt, each year it will have to auction $400 B in new 1-year securities. So its funding needs in terms of cash flow operations are bigger if the maturity structure of its outstanding debt is shorter.
But the puzzling thing about this statement to me is that, regardless of whether the public or the Fed owns those $400 B in 1-year securities, the Treasury is still going to have to auction $400 B each year to the public, because the Fed is required to buy Treasury securities on the open market rather than from the Treasury itself. So, prior to the Fed’s MEP, the Treasury had to auction $400 B in new bills each year (which the Fed subsequently bought), and after the Fed’s MEP, the Treasury still has to auction $400 B in new bills each year (which now must remain in the hands of the public).
The statement from the TBAC minutes suggests to me that in practice, these two outcomes are viewed as very different by the Treasury. In the first case, $400 B is auctioned to the “public”, but the dealers bidding on the securities know that the Fed is shortly going to take them off their hands, and this determines their willingness to bid. But in the second case, the demand from the “public” may not be there if the Fed is no longer buying up these T-bills.
OK, but that’s how these operations are supposed to work: force the public to hold a higher portion of bills and lower portion of bonds in hopes of twisting the yield differential. But the TBAC minutes continue:
Next, the presenting member reviewed various scenarios by which Treasury could finance this [MEP-induced] gap, including financing with bills, coupon securities or floating rate notes. According to the presenter, depending on how Treasury finances the effect of MEP, the average maturity at the end of FY2016 could range from 67.5 month to 70.2 months.
Committee members then debated the best way to fund the short-term effect of MEP. Some members noted that increasing coupon issuance to fund the MEP shortfall would be consistent with the Committee’s long standing recommendation to extend the weighted average maturity.
Beckworth was interested in this picture from the Office of Debt Management 2012 Q3 report. As described in the report, this is simply a projection of what the current Treasury issuance percentages (which include significantly more long-term securities and less short-term securities than normal) would imply if continued into the future.
 |
The TBAC minutes provide this added information:
[Office of Debt Management Director Colin] Kim reviewed several debt metrics. As of June 30, 2012, the average maturity of the portfolio was approximately 64.2 months. In the chart presentation showing the projections for Treasury’s weighted average maturity, Kim adjusted future note and bond issuance on a pro-rata basis to match projected financing needs. The simulation showed that the average maturity continues to extend well above the 3 decade average of 58.1 months. By 2016 it could reach the upper end of the historical range.
He emphasized that the average maturity projections and the associated underlying assumptions for future issuance were hypothetical and not meant to convey future debt management policy or an average maturity target. He also reiterated that Treasury will remain flexible in the conduct of debt management policy.
The report also provides this summary of the composition of outstanding Treasury debt and what current procedures would imply for the future composition.
 |
All of which leads me to conclude that, if new policy initiatives are announced at the next FOMC meeting, they will not take the form of more twisting and maturity extension, but instead would consist of large outright purchases.
Professor Hamilton, that the first MEP lowered Treasury’s borrowing capacity through 2016 by $400 billion – does this have anything to do with the fact that when the Fed holds the security the return would be routed back to the Treasury, but when the public holds it, it is not so?
PB: No, remember we’re just talking about a swap of securities. The Fed will not be returning the interest on $400 B T-bills to the Treasury, but instead will be returning the interest on $400 B T-bonds. Since the latter pay a higher yield, there is actually more revenue returned to the Treasury under the swap.
Professor I have a question. It’s probably very basic, but I can’t figure it out
Right now rates are very low and falling all along the yield curve. The story one hears is that things are so bad that people would rather put their money in treasuries than pretty much anywhere else no matter how little (even negative) they yield
Now, on the other hand, a stimulative policy on the part of the Fed would mean that they would inject money into the economy by buying bonds (that’s what an OMO is, right?).
The question is, doesn’t that stimulative policy on the part of the Fed just exacerbate the problem of a very high demand for bonds?
and another, somewhat related question: in an OMO the Fed buys bonds in exchange for money. The question is, under what terms would banks want to engage in that transaction, that is be willing to sell their bonds?
if you could help me out with that, I’d really appreciate it
Professor Hamilton – Your confusion is evident in this inaccurate statement:
“But the puzzling thing about this statement to me is that, regardless of whether the public or the Fed owns those $400 B in 1-year securities, the Treasury is still going to have to auction $400 B each year to the public, because the Fed is required to buy Treasury securities on the open market rather than from the Treasury itself.”
In fact, the Fed “rolls over” maturing Treasury securities at auction (this is described elsewhere, but also in the April statement toward the bottom: http://www.federalreserve.gov/newsevents/press/monetary/20120425a.htm ). As you noted, this practice is not possible for outright purchases, but is the standard option for securities rolling off of the portfolio.
This is not done competitively, but comes in the form of an add-on to Treasury auctions. Note that the SOMA line is now zero (go back a few months and you will see a number in the SOMA line that represents some portion of the roll-over of maturing securities): http://www.treasurydirect.gov/instit/annceresult/press/preanre/2012/R_20120808_1.pdf
Thus, Treasury’s financing needs are increased now that the Fed is keeping its (par) Treasury portfolio at a constant size by engaging in a maturity swap rather than simply passively rolling over at auction. The size of the program (400bn and 267bn – being either sold or redeemed) represents the amount of add-on the Treasury will no longer enjoy from the Fed relative to no MEP.
Of course, assuming the purchased securities are held to maturity, the Treasury will now realize larger SOMA add-ons in 6-30 years than they would have without an MEP, which is why they qualify the statement with “through 2016”.
The Krishnamurthy/Vissing-Jorgenson paper (BPEA, 2011) finds that QE lowers rates on Treasuries but not so much on riskier bonds, when the policy is mainly directed at Treasury purchases. So to get any traction, large-scale asset purchases would have to be directed at riskier assets (something Keynes advocated in the GT). I’d be interested in your opinion on this.
“regardless of whether the public or the Fed owns those $400 B in 1-year securities, the Treasury is still going to have to auction $400 B each year to the public, because the Fed is required to buy Treasury securities on the open marketrather than from the Treasury itself.”
I haven’t followed all of your analysis here, but I believe the Fed is allowed to replace the dollar amount of maturing securities it holds by an equal amount of securities purchased directly at auction. In other words, it can buy direct at auction, provided it does not increase its stock of debt holdings. In that sense, its capacity to buy direct from Treasury is short term “grandfathered” to the point of replacing maturities in order to maintain the level of what it already has in its portfolio.
But I’m not sure what difference this makes either. If the term structure of debt held by the public shortens, due to MEP, then the public’s share of the gross volume of Treasury issuance required to roll over that debt at maturity increases, in the short run. But does that matter to anyone? And if the term structure of debt correspondingly held by the Fed lengthens, then its share of gross volume issuance required to rollover the debt decreases in the short run, but does that matter to anyone?
The more important question is the share of current deficit financing that is absorbed by the Fed, in effect. And MEP has no effect on that.
I don’t know if any of this matters to your analysis. Sorry, I don’t have a reference handy on the auction rule noted above.
continuing …
So, a Fed MEP corresponds to a publicly held MSP (maturity shortening).
“The presenter noted that for FY2013 through FY2016 Treasury would need to raise an additional $425 billion”.
So it looks like they’re measuring borrowing requirements according to what they need to issue gross to the public, which is more than before because of the increased bunching of publicly held maturities at the short end.
The “borrowing capacity” language seems awkward. I can only guess it refers to the net deficit financing effect of a given projection for gross Treasury issuance to the public. The shorter the outstanding term structure, the more gross borrowing is required just to stand still, etc.
E.g. see final sentence on first page here:
http://www.newyorkfed.org/research/current_issues/ci13-1.pdf
“The announcement also discloses the quantity of securities that the Federal Reserve has maturing on the issuance date — securities whose proceeds might be reinvested in the new security.”
honestly, i think too many economists are focused on the particulars of how this or that method of stimulus does x, and not enough on what the Fed is communicating, i.e. how tolerant they will be of 2.5% inflation or 6% nominal growth (when/if the housing market rebounds). I tend to think low rates are not really the problem.
What the Fed really needs to do is engineer faster deleveraging: that could be through higher asset prices (which reduce the need for state/local govts to cut staff to plug pension holes), through higher real growth which accelerates debt paydown, or higher inflation (which does the same). The fact that asset prices right now are highly correlated with inflation expectations means that we just need some old fashioned money printing, but we all know the fed has been very reluctant to do that, which gets me back to my original point: its not this or that stimulus policy, its the speed limit that they are uncomfortable with in the first place.
The effect of outright purchases (QE) and MEP on private sector portfolios is the same. Each involves a Fed swap of a s.t. asset (T-bills or Excess Reserves) for a l.t. asset (T-bonds). All else equal, the effect of both QE and Twist is to transfer duration risk from private portfolios to the Fed’s balance sheet. Switching from Twist to QE will not change that.
A lot of all this is way over my head. From my peanut gallery perch, however, I would think the Treasury would be interested in lengthening it’s portfolio duration in order to capitalize on historically low interest rates. The opposite of a bank, in other words, which chronically borrows short and lends long. Instead the Treasury needs to spend short and borrow long.
If the Fed is buying up longer dated bonds, thus reducing the supply of same, would not that actually help Treasury’s desire to extend it’s portfolio duration and thus capitalize on the lower rates? It could be a wash however in terms of absolute amounts.
I don’t have a clue what role public purchases play in this scenario. Anyway, if you think of the Fed as Treasury’s bank, then this makes sense. Treasury lowers the cost of its debt by lengthening it’s duration and locking in low rates while the Fed, just like a bank, is lengthening it’s Treasury (credit) duration and funneling cash out.
“The opposite of a bank, in other words, which chronically borrows short and lends long. Instead the Treasury needs to spend short and borrow long”
This would make economic sense but it would be political suicide. Even though long term rates are “low,” extending the maturity would drive up interest rate costs to the Treasury, worsening the US gubbermints fiscal position as it is, and making deficits larger, and/or tax hikes, and/or spending cuts.
Thanks to mat, JKH, and other knowledgeable readers who emailed me separately in order to clarify that although the Fed is not allowed to purchase new securities with cash at the standard auction, they apparently are allowed to roll over securities that they already hold through the standard auction, e.g., page 7 of the Desk’s annual report or the second paragraph of
the June 12 Operating Policy.
It nevertheless still seems very strange to me that the Treasury would perceive MEP as directly lowering the Treasury’s borrowing capacity, or think that it needed to make any decision other than to auction to the public the identical bills that the Fed would have requested in its standard rollover. The thing that really stumped me was that the notes to the figures I produced above indicated that they were based on “estimated projections of the Maturity Extension Program”. I am still astonished to have learned that the Treasury thinks it has to condition the maturity of securities that it offers based on the extent to which the Fed decides to implement its Maturity Extension Program.
I agree with dwb. Some faster deleveraging is necessary… That used to mean default and bankruptcy… Now it means public assumption of private debts via quantitative easing.
Against the dollar, wholesale food prices have increased 130% in the last 7 years, gasoline 75%, and healthcare 50%. For the lower half of Americans, the rate of inflation on their most important consumables is already unacceptable. Increasing the money supply to produce more inflation will only make their market access more difficult.
Those of you interested in this issue might find this report by my friend Ed Rombach interesting. Though it is more concerned with the creation of deflation it is related to the topic.
Those of you confused about the huge excess bank reserves might find this especially interesting.
“I am still astonished to have learned that the Treasury thinks it has to condition the maturity of securities that it offers based on the extent to which the Fed decides to implement its Maturity Extension Program.”
I think their unusual language must be geared toward a focus on issuance of debt to the public, after setting aside a sort of benchmark assumption about the size of the Fed balance sheet.
In a way this makes sense, because the effect of the Fed’s balance sheet is to transform the term structure of what they buy/switch/hold into the term structure (i.e. interest rate sensitivity) of reserves. After taking into account the net interest margin that is submitted from the Fed to Treasury as part of annual profits, it’s as if the cost to Treasury of the Fed portion of the debt is the overnight rate.
So the interest rate sensitivity of the entire debt is effectively the overnight rate for the Fed balance sheet, plus whatever is publicly floated. If the first is a given, or assumed, then they need to monitor the second in order to manage the interest rate sensitivity of the entire federal debt. The Fed’s market activity changes the publicly held term structure, so they monitor that accordingly and adjust strategy as a result.
I imagine it will become doubly interesting to interpret this sort of thing if they ever get to the stage where the Fed is actually selling bonds into the market as part of an exit strategy.
Although at this stage more intuitive than methodology,when Central Banks will predominantly own a certain critical percentage of the government debts directly or through proxies, that is debts amounts, maturities, prices, it will display the characteristic of a mathematical sequence where one may intuitively see the convergence through a limit but the limit is yet unknown.
Almost everyone believes that the Treasury is actively lengthening the average maturity of its debt. This is not true. Treasury has actually been selling almost all of its debt with very short maturities.
So why has the average maturity been increasing? The average maturity grows “naturally” because the expiration of maturing debt removes downward pressure on the average. This is like driving up the average IQ of a class by removing the kids with low IQs. The average maturity has a built-in tendency to lengthen (for a few years) since the debt portfolio is so front-loaded.
I recently gave a presentation on this matter. The slides and a summary of the talk are available here:
http://welltemperedspreadsheet.wordpress.com/2012/07/24/garp-presentation-on-interest-rate-risk-and-the-treasury/
Win Smith: The average maturity of outstanding Treasury debt increases or decreases depending on the maturity structure of debt newly issued by the Treasury. To say that a continual increase in average maturity of outstanding Treasury debt is “natural” is complete nonsense.
The average maturity grew slightly in May 2012. One would think the reason was that the new securities sold in May must have had a longer average maturity than that of the whole debt portfolio. But the average maturity of the May issuance was less than half that of the whole portfolio. And if you exclude the new issuance from the May 31 portfolio, the average would have increased much faster than it actually did.
So why did the average maturity increase?
There are three main influences on the average maturity. First, most of the securities continued to be outstanding in May, getting one month closer to maturity and tending to reduce the remaining average maturity of the portfolio. Second, the short new issuance also tended to reduce the average. The third and largest impact came from the securities that left the portfolio by maturing in May. These securities had extremely short remaining lives when they were part of the portfolio in April. They put especially strong downward pull on the April average since they were far away from it. By May 31 they were gone and no longer suppressed the average. The exit of the matured securities lengthened the average more than the aging of the portfolio and the new issuance shortened it.
If the Treasury’s outstanding securities were uniformly distributed by maturity, today’s average maturity would be about 15 years. If the Treasury stopped issuing new debt, the average would follow a straight line down to zero over the next thirty years.
But the debt is heavily front-loaded, with a median maturity of about 3 years. Without new issuance, the remaining average maturity would continue to increase for ten years as short debt burns off. Then it would decay to zero over the next twenty years.
It’s like the average age of a retailer’s inventory. It depends on three things: the aging of existing inventory, the inflow of new goods, and the outflow of sold goods. You can’t look at just one of these and hope to understand the inventory numbers.
This is just math.
Win Smith: No one who knew the first thing about math “would think the reason was that the new securities sold in May must have had a longer average maturity than that of the whole debt portfolio.”
You’re miles behind the rest of us on this conversation.
if housing comes back,and interest rates increase due to a stronger economy, and stock market etc…who will be the major determinent in increasing rates? will it be the fed selling it’s holdings ,or the public leaving the bond market? and by what percentage do you guess will one predominate over the other?