When and how the Fed will depart the MBS market is no longer a mystery. Following last month’s FOMC meeting, it announced it would purchase the full $1.25 trillion of MBS previously announced, but that it would gradually slow the pace of those purchases “in order to promote a smooth transition in markets.” It “anticipated” the program would be completed by the end of the first quarter of 2010. The Fed’s withdrawal plan was very close to what many MBS analysts had been anticipating/recommending in their weekly reports for some time (indeed, market chatter says the NY Fed reached this decision after conferring widely among market participants). With the headline uncertainty resolved, the market has refocused on the question of how much might pass-throughs cheapen next year, in the Fed’s wake, as well as on more immediate issues of the Fed’s impact on relative value within the fixed-rate pass-through market. What follows is a mini-review of recent commentary on the Fed’s role in the MBS market from professional sell-side MBS analysts. It’s salted with my own thoughts and it’s not necessarily comprehensive. I may have overlooked or misfiled something among the dozens of “pdfs” accumulating on the hard drive. Also, a couple of the firms still trading MBS and producing quality MBS research (the number shrank significantly in the crisis) consider their analysts’ thoughts proprietary and reserve distribution to firm customers. Given the value of information, about mortgage markets, their participants and how they function, to policy makers, I think this is shortsighted. Tracking Every Trade The Fed’s activity in the MBS market is easy to track. The money managers acting on its behalf are required to indicate when they are trading for the Fed’s book, so that daily trading commentaries provide considerable Fed “color.” In addition, the Fed provides a weekly accounting for the MBS Purchase Program in the H.4.1 Statistical Release “Factors Affecting Reserve Balances.” Tracking the Fed’s purchases became even easier this month when the fed began releasing the pool numbers of its holdings (once they settle). Using this information, along with detailed pool level data from Fannie, Freddie and Ginnie, analysts tabulate, chart and analyze this information versus new, existing and tradable supply of pass-throughs. (For example, pass-through pools owned by CMO/REMIC trusts are not tradable.) In brief, these reports indicate that Fed purchases exceeded eligible gross issuance in the months of April, May, June and July. More telling, they exceeded net issuance (new supply minus prepayments) in every month this year (and are projected to do so through the first quarter of 2010). Altogether, they’ve purchased almost 25% of outstanding fixed-rate 30-year Fannie, Freddie and Ginnie pass-throughs. The Fed’s share of the outstanding bellweather GSE pass-through market is a bit higher – over 26% and of pools not pledged to CMOs its more like 30%. Wither Spreads? MBS research groups across the market are agreed that Fed purchases (actual or anticipated) have been the driving force behind the dramatic tightening of MBS spreads over the last year. And all are agreed that spreads will widen on the Fed’s departure. Before going into how much they can widen, a housekeeping note. When speaking of MBS as investments (rather than the conduit of capital market funding to mortgage borrowers), analysts prefer to speak of option-adjusted spreads (OAS). Simply put, OAS are calculated in models that calculate the long-term cut that borrowers’ rights to prepay can be expected to extract from nominal MBS yields. For example, on one proprietary daily report, the October 20 price on a theoretical par-priced 30-year Fannie (aka the current coupon) yields 119 basis points to swaps, but the comparable OAS is negative 14 basis points. (This example also should convince non-investor readers that MBS are very difficult investments. Skilled investors never expect that the yield assumed at purchase can be realized by holding the pass-through to maturity.) Since individual models produce different OAS for the same security at the same price, I’ll avoid quoting any more spreads. However over time the models track the same value trends, so most research desks observe that current coupon OAS have tightened by 90 to 100 basis points from the wides posted last fall to tights seen this summer. They’ve widened since, but still lie within the trading range that was defined 2005-06, before the crisis. That is a terrific amount of tightening, but not surprising given that Fed purchases have exceeded net new supply. The obvious question then is, Do MBS give back all the tightening after the Fed? Several MBS analysts at broker-dealers with large MBS trading operations don’t think the widening will be nearly as pronounced as the tightening was. They argue that investors were crowded out by the Fed, but they will return when the Fed leaves and spreads widen. Fannie and Freddie will not lead this return – as I’ve stressed before, they are scheduled, by law, to begin shrinking in 2010. But banks, foreign accounts and money managers – including indexed funds (not passive replicators, but active index beaters) – should come back and provide significant demand on the margin. Not all analysts writing about the Fed’s departure will quantify the widening, but a couple have screwed up their courage to put a number on it. One house expects OAS to widen about 20 basis points from current levels, while the other expects OAS to overshoot by about 30-40 basis points before stabilizing around 25 basis points above current levels. All else equal (a brave assumption), mortgage rates would adjust upward in a comparable manner (saying more would be to start down a slippery slope). Projecting a widening of 40 basis points (or less) may be a tad optimistic, especially at the outset, when the Fed shuts down. Buyers will step back until the inevitable widening appears to be over or close to it – a process that can feed on itself as investors recalibrate perceptions and rules or thumb re: what’s cheap enough. After all, these are unprecedented circumstances the market faces, and history is likely to be a faulty guide. Consider too that many market-value-sensitive investors will be trimming positions (lowering their mortgage weightings in favor of appealing corporate spreads, for instance) as the end of 1st quarter 2010 approaches in order to avoid the price impact of the inevitable widening. Strip-Mining the MBS Landscape Both the magnitude and style – concentrating on newly minted pools – of Fed purchases have profoundly altered the pass-through investing landscape. Above all, they’ve reduced the tradable supply of pools. For all practical purposes, the Fed and Treasury purchases don’t trade – we would say they’ve been removed from the tradable MBS “float.” That means the 30% of formerly tradable supply is now removed from tradable supply. Moreover, the Fed purchases, focused on new production, have had proportionately heavier impact on coupon “buckets” filled in the super-low mortgage rate environment created by the Fed: 4s, 4.5s and, to a lesser extent, 5s. (To maximize liquidity, originators vary the amount of servicing they “strip” from a pooled mortgage to create securities with whole and half coupons. That means loans made at 5% and 5.125% can be pooled in a 4.5% pass-through.) For example, almost all of the outstanding 30-year 4s were originated in 2009 and the Fed bought most of them. Likewise, the Fed owns about three-quarters of the 30-year 4.5s. Higher coupon 5s to 6.5s were produced in large amounts in earlier years, so, as we would expect, the Fed holds a smaller share of the tradable amount of these coupons. This changes the day-to-day trading environment for denizens of the MBS market. Some coupons have become difficult to trade and some susceptible to squeezes. Dollar rolls (like a repo, but conducted in TBA securities rather than specific pools with unique CUSIPs; commonly used both as a securities lending and a financing mechanism) have likewise traded rich. In higher coupons like 5s and 5.5s, where Fed purchases exceeded new issuance by wider margins (and the Fed was effectively removing paper from other investor’s holdings), normal value relationships between TBA and specified pools were distorted. That is, seasoned paper issued in, say 2005, which in “normal” market conditions would offer extra value for favorable prepayment characteristics or observed prepayment behavior was more valuable delivered into a TBA trade (TBA is “cheapest to deliver”). The Fed purchases also removed a disproportionate amount of duration from the market. That is, by loading up on low coupon 4s and 4.5s, the Fed took the lion’s share of the pass-throughs with the lowest prepayment risk and longest expected average lives and durations. This has a number of consequences for MBS market participants’ hedging practices. For example, these coupons have value as hedging instruments for servicers. Declining interest rates (increasing prepayment risk) decrease the value of interest-only servicing income. Low coupon MBS are a natural hedge for this risk. Another example would be the banks and other mortgage investors who historically have responded to interest rate rallies by “going down in coupon.” By taking up most of the low coupon MBS, the Fed has left these investors little room to shift if interest rates decline from current levels. They’ll be forced to hold a larger chare of premium-priced MBS than they would otherwise care to do. In other words, the Fed has not just out-bid traditional MBS players, it has also squeezed their opportunity set. Perhaps the most striking symptom is the small controversy that erupted in recent weeks over the MBS “index.” Some background: An important segment of the fixed income investment managers market their ability to improve on the performance of specified bond universes. In addition to these fund managers, many bank and insurance portfolio managers are evaluated internally versus an index. Indeed, as the GSE MBS market grew over the last decade or so, so did its weight in commonly referenced aggregate U.S. indices grow (it’s presently around 38% of the aggregate), making many bond generalists de facto MBS investors. It’s easy to see the conundrum Fed purchases have created for investors who are measured against MBS or aggregate bond indexes: the coupons trading best are the ones the Fed is buying. But the Fed has taken them out of circulation, so how do you capture that performance if you can’t own a proportionate share of the coupon? In June, Barclays Capital (acquirer of the universe of Lehman Indices) announced it would offer a US Aggregate Float-Adjusted Index, a benchmark of the investment grade bond market that would exclude Treasuries, agencies and MBS held in Federal Reserve accounts. It would also offer a float-adjusted MBS index. The change seems to have raised few eyebrows at the time. Nonetheless, when a big investor – Vanguard – did announce it would shift to float-adjusted indices for 12 of its bond funds, some analysts called the step a negative for the MBS market. Nay-sayers expected Vanguard to allocate away from MBS most or all of the float adjustment. By contrast, Barclays MBS analysts wrote they did not expect investors shifting to the float-adjusted indexes would have a big impact on the market, because the Fed already owned the paper being removed from the index. In other words, the investor community was already underweight the coupons dominated by Fed purchases and would not have to rebalance to achieve a desired stance versus the new indices.