In the current debate about reform of the housing finance market, proponents of a continuing government role frequently argue that without government backing, the so-called "to be announced" (TBA) market could not exist. This is similar to the argument that without government backing there would be no 30 year fixed rate mortgages—and is equally false. It is perfectly possible to have a TBA market that does not involve government backing, and this memorandum will explain how this might be done.
The memo will begin with a description of how the agency TBA market currently works, and proceed to a discussion of how a TBA market in privately issued mortgage-backed securities (PMBS) would function.
First, however, it is important to put the TBA issue in perspective. Just as commodity futures markets enable farmers to hedge the price risk of their commodities, the agency TBA forward market allows mortgage originators to mitigate their interest rate risk. Today, originators of both agency and non-agency mortgages use the agency market to hedge the risk of a change in interest rates between the time that a mortgage rate is "locked in" and the time the mortgage is actually closed and securitized. Reducing that risk has a positive effect on mortgage rates, but it is only one of the elements that go into the full cost of a mortgage. The mortgage market could function effectively without a TBA market, but total mortgage costs—the principal component of which is the interest rate—would be slightly higher. As will be shown, government backing is not a requirement for the TBA market, just as it is not a requirement for a 30 year fixed rate mortgage.
Thus, as outlined in this memorandum, it is perfectly possible for a TBA market to develop for PMBS.
It is sometimes said that the TBA market is responsible for the liquidity of the agency MBS market, but this is not strictly true. The TBA market works efficiently for agency MBS because of the liquidity in the agency MBS market, but it doesn't create that liquidity. Instead, that liquidity is created by a convention—an agreement—among market participants about what they will accept as sufficient information about a particular pool of agency MBS. That convention, embodied in the Uniform Practices for the Clearance and Settlement of Mortgage-Backed Securities –administered by the Securities Industry Financial Markets Association (SIFMA) and otherwise known as the "Good Delivery Guidelines"—establish that a seller and buyer in the TBA market need only agree on six factors to confirm a trade: issuer, maturity, coupon, price, par amount, and settlement date. For example, as described in a recent paper by the Federal reserve Bank of New York, "a TBA contract agreed in July will be settled in August, for a security issued by Freddie Mac with a 30-year maturity, a 6% annual coupon, and a par amount of $200 million at a price of $102 per $100 of par amount, for a total price of $204 million."
The limitation on the information available to the buyer makes the agency MBS, in effect, "fungible" with other agency MBS already outstanding and thus adds significantly to the liquidity in the market. This is a very important point, to which we will return later in this memo when we discuss how a PMBS market would function. As also explained in the Federal Reserve paper: "Paradoxically, the limits on information disclosure inherent in the TBA market actually increase this market's liquidity, by creating fungibility across securities, and reducing information acquisition costs for buyers of [agency] MBS. A similar argument has been used to explain why DeBeers diamond auctions involve selling pools of diamonds in unmarked bags that cannot be inspected by potential buyers."
Accordingly, an important factor that adds to the liquidity in the TBA market is the homogeneity of the pools that Fannie and Freddie create:
The treatment of TBA pools as fungible is sustainable in part because a significant degree of actual homogeneity is present amongst the securities deliverable into any particular TBA contract. The most obvious source of commonality is the GSEs' guarantee of the cash flows of mortgage principal and interest, which essentially eliminates credit risk. However, the GSEs' standardization of underwriting and securitization practices also contributes to homogeneity as well. At the loan level, the standardization of lending criteria for loans eligible for agency MBS constrains the variation among the borrowers and the properties underlying the MBS. At the security level, homogenizing factors include the geographic diversification incorporated into the pooling process, the limited number of issuers, and the simple structure "pass-through" security features.
Thus, apart from the elimination of credit risk, the liquidity associated with the TBA market is created by the homogeneity of the mortgages and such things as geographic diversification of the pool. This will be important when we begin to consider how the TBA market might function when PMBS instead of agency MBS are the principal trading instruments.
How the TBA market functions with agency MBS
As noted above, the sale of MBS on a TBA basis serves one fundamental purpose in the mortgage market; it allows mortgage originators to hedge their interest rate risk. An interest rate hedge is important because there is generally an extended period (say, 30-90 days) between (i) the time that a mortgage originator agrees to a rate with a borrower and (ii) the time the mortgage is actually closed, added to pool of mortgages for securitization and delivered to Fannie or Freddie (the agencies) for securitization. During this period, market interest rates may change, going either up or down. Interest fluctuations can have two different kinds of effects on an originator:
- Changes in interest rates can significantly change the value of a pool of mortgages between the time the rate is "locked in" and the time the mortgages are actually securitized and sold. If interest rates rise during this period, the mortgages become less valuable; unless hedged, the originator will receive less compensation when the mortgages are ultimately securitized and sold. On the other hand, if interest rates fall, the mortgages become more valuable. This would ordinarily be profitable for the originator, except for the fact that borrowers generally have an option not to close. When interest rates fall, many more borrowers than the originator had anticipated when developing the pool elect not to go through with the closing. This is known as "pipeline (or "fallout") risk."
- Fallout risk, as noted above, is affected by interest rate changes. Normally, originators assume a particular fallout rate on a pool of mortgage commitments. If rates rise, the original locked-in rate looks more attractive to borrowers, and increases their incentive to close. Thus, just as a rise in rates reduces the profitability of a pool of loans for the originator it increases the number of borrowers in the initial pool who want to go through with the transaction. The opposite is true when interest rates fall. At that point, the pool is more valuable for the originator, but many more borrowers will elect not to go through with the closing. Originators generally handle fallout risk by hedging fewer locked-in loan applications than are in the pipeline. If rates fall that estimate is likely to be too high; if rates rise it will be too low.
Originators can hedge both these risks to some extent in the TBA market by selling a pool of mortgages forward for a closing and securitization at a later date. The sale, priced based on the market interest rate prevailing at the time, requires that the originator deliver a specified amount of MBS at a closing in the future. This assures that the originator will not lose money if interest rates rise, since the originator will have hedged this risk. On the other hand, the originator must cope with the lower than expected fallout rate, which means it must now find a buyer for the additional lower-value mortgages the originator now has in the pool. The liquidity in the TBA market allows this to be done efficiently because it is possible to sell the excess mortgages through additional agency MBS. The originator will of course receive less for these mortgages than in the forward sale, and that reduces its profit on the transaction as whole, but it at least has a market in which the sale can be made.
If interest rates fall, the results are the opposite for an originator that has hedged in the TBA market. The mortgages in the pool become more valuable, but the originator does not realize this gain; it must still deliver the same par amount of MBS, despite their increase in value. However, the borrower's option not to close reduces the number of mortgages the originator will be able to deliver in the forward sale. If, as a result of the higher fallout rate, the originator is short the mortgages required for delivery, the liquidity in the TBA market allows the originator to purchase and substitute agency MBS for the securitized mortgages originally promised to the buyer or buyers of the pool. These agency MBS will be more costly than the price it received for the MBS delivered in the forward sale, and that will reduce the originator's profit on the transaction as a whole, but the liquidity of the TBA market at least allows the originator to buy substitute agency MBS to cover its short position.
The above description is simplified in order to demonstrate principles. In practice originators manage their pipelines on a daily or even hourly basis.
Observers of the TBA market believe that the interest hedging opportunity provided by the TBA market reduces interest rates by reducing originator risks, just as the forward sale of corn or wheat—by reducing farmers' risks—probably reduces market prices. At the same time, it should be noted that there may be other ways to hedge if the TBA market did not exist. For example, it may be possible to purchase protection against an interest rate movement loss in the credit default swap market.
Finally, because Fannie and Freddie securities are exempt from registration under the securities laws, forward sales can be accomplished even though the mortgages underlying the TBA securities have not yet been originated and without the additional expense of SEC registration.
Accordingly, the TBA market as it functions today for agency MBS provides many advantages for hedging the risks of originators, including the opportunity to sell mortgages forward before a complete pool has been assembled; the ability to buy or sell agency MBS that are considered in the market to be identical with the agency MBS that have been sold forward; and the ability to offer and sell securities forward without SEC registration.
A TBA Market without Fannie and Freddie
Could a TBA market function for PMBS? It is clear that such a market could develop, although at least at the outset it will not be as efficient as the TBA market for agency MBS. In addition, a change in the Securities Act of 1933 may be necessary to assure that pools of private MBS will be eligible for what is called "shelf registration," where securities already registered are placed "on the shelf" ready for issuance when market conditions are deemed propitious. There are good policy reasons why this should be done.
Hedging interest rates before the PMBS market matures. In order to consider how a TBA market in PMBS would work, let's consider the interest rate hedging problem discussed above. To do so, we will assume that at some point in the future it will be established—by legislation or SEC ruling—that PMBS are eligible for shelf registration. In addition, we will assume that only the AAA and AA-rated tranches (the "A" Tranches) of a pool are securitized, and that the lower-rated tranches are privately placed so that no SEC registration is necessary (the "B" Tranches). Finally, we will assume—as the AEI plan proposes—that the GSEs will be wound down gradually over a period of five years.
At least at its outset, there will be relatively little liquidity in the PMBS market. Over time, of course, as more PMBS enters the market, this will change, but for purposes of this memorandum we'll first discuss how an originator would hedge its interest rate risk in the most challenging circumstances—that is, when there is little liquidity in the TBA market for PMBS.
In this case, the originator of a pool of mortgages will effect a short sale of agency MBS in an amount equal to the amount of PMBS the originator expects to sell in the future when the pool is complete. By doing this, the originator puts itself in the same position described above in the discussion of the how the TBA market currently functions. Everything described above works the same way. If interest rates rise, the originator stands to gain from the rise in the value of the agency MBS that were sold short. The originator buys in agency MBS at the lower rate and delivers them to cover the short position. The gain covers the decline in value of the PMBS the originator will eventually sell, and the originator takes a loss on the excess PMBS now on hand because of fallout risk.
The opposite is true again if interest rates fall. In that case, the PMBS the originator will sell in the future will become more valuable, but the originator will have to buy more expensive agency MBS to cover its short position. The important point here is that as long as the agency MBS market exists, it is possible for originators of PMBS to use that for interest rate hedging purposes. The match between the agency MBS and the PMBS that the originator will be selling is not exact—it's possible that the if interest rates rise or fall they will do so slightly differently for agency MBS than for PMBS—but the match is close enough to cover most of the originator's interest rate risk. The balance of any losses would have to come out of the originator's potential profits on the loans.
Hedging interest rates after the PMBS market matures. As the GSEs are gradually wound down over a five year period, the agency market will still likely be liquid enough to support an agency TBA market, while the PMBS market becomes more liquid. At some point, there will be enough liquidity in the PMBS market so that PMBS can be directly hedged by selling these securities forward. As discussed above, in addition to the size of the market, there are several reasons for the liquidity of the agency MBS market, including the government guarantee (i.e., the absence of credit risk) and compliance with the Good Delivery Guidelines of SIFMA. The Good Delivery Guidelines assume that the mortgages underlying the pools fall within certain underwriting parameters, and this allows agency MBS to act as though they are fungible, even though there are differences among the loans that are actually in the thousands of agency MBS pools.
The AEI plan would require that securitized MBS meet certain statutory underwriting standards for a prime loan and also have credit support from two factors—the collateral in the "B" tranches and mortgage insurance coverage down to a 60 percent loan-to-value ratio. These requirements should enhance the fungibility of the PMBS issued under the AEI plan, and thus eventually—as the size of the market increases over time—produce substantially the same liquidity benefits as are now achieved by the TBA market for agency MBS. While the combination of a AAA or AA rating and mortgage insurance is unlikely to be a complete substitute for a government guarantee, the credit performance of PMBS should be stable enough over time so that the fungibility factor will allow a TBA market in PMBS to function much as the TBA market for agency MBS functions today.
In that case, several years down the road, as the PMBS market has grown, an originator should be able to sell the Senior Tranches forward to a buyer at the price required by the market interest rate on the date of the forward sale. If interest rates rise, the value of the pooled mortgages will decline and the number of borrowers who want to close at the lower locked in rate will rise. The originator will be hedged by the original forward sale against the losses associated with the increase in interest rates, but will have to sell the excess PMBS (or whole mortgages) at a reduced value because more borrowers than anticipated chose to go through with the closing. This sale should be possible, as it is in the TBA market for agency MBS today, because there will be a stable market price for all PMBS backed by mortgages that meet the statutory requirement for a prime mortgage under the AEI plan.
Similarly, if interest rates fall, the originator will not realize a gain in value of the PMBS, since it has already sold the pool at a lower price, but the number of borrowers who want to close—exercising their option not to close—will fall. The originator could thus be in a position where it will have to buy in additional PMBS to cover its short position on the delivery in the forward sale. If there is a liquid market for PMBS at that point, the originator should be able to acquire the PMBS for that purpose in the open market, but of course at a lower price than the higher yielding MBS it originally sold forward. This will reduce the originator's profit on the whole transaction, but will avoid the significant losses that might have occurred if there was no market in which the originator could find the PMBS to cover its short.
There is no reason in principle why a TBA market for PMBS that conform to the AEI plan would not eventually function as nearly as effectively as the current TBA market in agency MBS. The elements necessary for a TBA market do not include a government guarantee. Instead, all that is required for a TBA market is (i) PMBS eligibility for shelf registration with the SEC, (ii) substantial after-market size, producing a high degree of liquidity, (iii) a robust level of credit enhancement, and (iv) a narrow set of performance and underwriting parameters, so that a set of Good Delivery Guidelines would allow pools with differing characteristics to be treated as fungible.
All of these requirements can be met under the AEI housing finance plan. Legislation that authorizes the AEI plan can also provide that the higher quality tranches in any PMBS pool will be eligible for shelf registration with the SEC. Until a PMBS market reaches sufficient size to provide the necessary liquidity, originators will be able to hedge their interest rates effectively by using the liquidity of the agency MBS market. After the PMBS market achieves sufficient size it will become the venue for direct hedging of PMBS. The AEI housing finance plan would produce prime mortgages that meet the standard for the narrow set of performance and underwriting parameters that permit agency MBS to be treated as fungible.
Edward Pinto is a resident fellow at AEI. Peter J. Wallison is the Arthur F. Burns Fellow in Financial Policy Studies at AEI.