Mortgage-Backed Securities Overview

A single misjudgment in prepayment speed assumptions can swing a mortgage-backed securities portfolio by 50 to 150 basis points in yield—and most investors discover this only after the damage is done. The $9.1 trillion agency MBS market (Source: SIFMA, 2024) is the second-largest fixed-income market in the United States after Treasuries, yet the mechanics of how cash flows move through these structures remain poorly understood by investors who treat them like plain-vanilla bonds. They are not.

The practical antidote isn't avoiding MBS—it's understanding prepayment mechanics, tranche structures, and the agency guarantee well enough to position intentionally rather than accidentally.

How Mortgage-Backed Securities Actually Work (The Cash Flow Chain)

A mortgage-backed security takes a pool of residential (or commercial) mortgages, bundles them into a trust, and issues bonds backed by the monthly principal and interest payments from those borrowers. The basic structure is straightforward: homeowners pay → servicer collects → trust distributes to investors.

But simplicity ends there. Unlike a corporate bond where a single borrower makes predictable coupon payments, an MBS pool contains hundreds or thousands of individual loans, each with a borrower who can prepay, default, or simply keep making scheduled payments. Every month, the cash flow has three components:

1. Scheduled interest based on the pool's weighted average coupon (WAC)
2. Scheduled principal from loan amortization
3. Unscheduled principal from prepayments and recoveries on defaulted loans

The point is: that third component—unscheduled principal—is what makes MBS analysis fundamentally different from corporate bond analysis. You know you'll get your principal back (assuming no defaults), but you don't know when. And timing determines yield.

The Players in the MBS Chain

Understanding who does what eliminates confusion about where risk sits:

The guarantor role is the critical distinction between agency and non-agency MBS. Agency MBS carry a guarantee from a government-sponsored enterprise (GSE) or government agency, meaning investors face prepayment risk but essentially zero credit risk on the underlying loans. Non-agency (private-label) MBS have no such guarantee—you bear both prepayment risk and credit risk. (More on this distinction in Agency vs. Non-Agency RMBS Differences.)

The Agency MBS Market (Why $9.1 Trillion Matters)

As of December 2024, Fannie Mae and Freddie Mac collectively guarantee $6.6 trillion in agency MBS, representing roughly 50% of all outstanding U.S. mortgage debt. Ginnie Mae—the only issuer backed by the explicit full faith and credit of the U.S. government—accounts for another $2.5 trillion, or about 20% of the market (Source: Ginnie Mae Global Market Analysis, January 2025).

Why this matters: the agency market dwarfs the non-agency market by a factor of roughly 10:1. When practitioners say "MBS," they typically mean agency pass-throughs unless they specify otherwise.

The Three Agency Issuers

Ginnie Mae (Government National Mortgage Association): Guarantees securities backed by FHA, VA, and USDA loans. Carries the explicit full faith and credit of the U.S. government. Ginnie Mae's portfolio grew 25.4% ($547 billion) from 2021 to 2024, driven by rising government-insured mortgage volume.

Fannie Mae (Federal National Mortgage Association): Purchases and securitizes conventional conforming loans. Operates as a GSE under federal conservatorship since September 2008. Carries an implicit (not explicit) government guarantee—though investors largely treat it as equivalent.

Freddie Mac (Federal Home Loan Mortgage Corporation): Similar mandate to Fannie Mae, also under conservatorship. Together with Fannie Mae, sets the conforming loan limit ($766,550 in most markets for 2024) that determines eligibility for agency securitization.

The practical distinction: Ginnie Mae's explicit guarantee makes its MBS marginally tighter in spread terms (typically 5-15 basis points tighter than Fannie Mae or Freddie Mac MBS with similar characteristics). For portfolio construction purposes, all three are treated as having minimal credit risk.

Prepayment Risk (The Variable That Defines MBS Returns)

Every MBS investor confronts the same core problem: borrowers have the right to prepay their mortgages at any time, without penalty, and they exercise that right at the worst possible moment for investors.

When interest rates fall, borrowers refinance, returning your principal early—just when reinvestment yields are lowest. When rates rise, prepayments slow to a crawl, extending your duration—just when you'd prefer shorter-duration exposure. This is the essence of negative convexity, and it's the reason MBS trade at a spread over Treasuries even when backed by the full faith and credit of the U.S. government.

Rate drop → Refi wave → Early principal return → Reinvestment at lower yields → Yield compression

Rate spike → Prepayment slowdown → Duration extension → Locked into below-market coupons → Price decline

The PSA Prepayment Model

The Public Securities Association (now SIFMA) developed the standard prepayment benchmark in 1985. At 100% PSA, the model assumes:

• Prepayment rates start at 0.2% CPR in month 1
• Rates increase by 0.2% CPR each month for the first 30 months
• Rates plateau at 6% CPR from month 30 onward

Multiples of PSA scale everything proportionally:

The durable lesson: PSA is a benchmark, not a forecast. Real prepayment behavior is far more complex—driven by refinancing incentives, housing turnover, borrower demographics, loan seasoning, and the "media effect" (when news coverage of low rates triggers a refi wave). Use PSA to communicate relative speed; use a multi-factor prepayment model for actual pricing.

Worked Example: How Prepayment Speed Changes Your Return

Setup: You purchase a $10 million pass-through MBS at par with a 5.50% pass-through rate, backed by 30-year fixed-rate mortgages with a weighted average maturity of 354 months (6 months seasoned).

Scenario A: 100% PSA (Baseline)

• Plateau CPR: 6.0%
• Weighted average life (WAL): ~11.8 years
• Expected yield at par: 5.50%
• Monthly cash flow at plateau: approximately $72,500 (principal + interest combined)

Scenario B: 250% PSA (Rate Rally)

• Plateau CPR: 15.0%
• Weighted average life (WAL): ~4.9 years
• If you purchased at 101 (a 1-point premium), your yield drops to approximately 5.18% because principal returns faster at par, accelerating the premium amortization
• Monthly cash flow at plateau: approximately $158,000

Scenario C: 75% PSA (Rate Spike)

• Plateau CPR: 4.5%
• Weighted average life (WAL): ~14.6 years
• If you purchased at 99 (a 1-point discount), your yield increases modestly to approximately 5.57% because the discount accretes over a longer period
• Monthly cash flow at plateau: approximately $57,000

The point is: the same security purchased at the same price delivers materially different yields depending on prepayment speeds. The WAL shifts from 4.9 years to 14.6 years across these scenarios—a nearly 10-year swing in effective duration from a single variable.

Tranche Structure (How Risk Gets Redistributed)

Not all MBS investors want the same exposure. Tranching takes a pool of mortgages and carves the cash flows into slices (tranches) with different risk-return profiles.

Sample Tranche Structure: $500 Million Agency CMO

In this structure, the Planned Amortization Class (PAC) tranche (A-1) receives principal within a defined prepayment band (say, 100-300% PSA), giving it the most stable cash flows and shortest WAL. The companion tranche absorbs prepayment variability outside that band—when prepayments are fast, the companion gets paid down quickly; when prepayments slow, the companion extends dramatically.

Why this matters: the companion tranche is effectively a leveraged bet on prepayment behavior. It can shorten to a 2-year WAL or extend to 25+ years, depending on the rate environment. If you don't understand what tranche you own, you don't understand what bet you're making.

Private-Label MBS: Adding Credit Tranching

Non-agency (private-label) deals add credit subordination on top of prepayment tranching. A typical pre-crisis private-label RMBS structure:

The credit enhancement for the senior tranche equals the sum of all subordinate tranches below it. If subordination is 12%, the pool must lose more than 12% of principal before the senior tranche takes any loss. (The average private-label RMBS deal from 2002-2005 contained 90.4% AAA-rated tranches, 8.4% mezzanine, and 1.2% equity (Source: Ospina & Uhlig, 2018, "Mortgage-Backed Securities and the Financial Crisis of 2008," NBER Working Paper 24509).)

The durable lesson: during the 2007-2009 crisis, those subordination levels proved woefully insufficient for subprime pools. Cumulative loss rates on 2006-2007 vintage subprime RMBS reached 20-30%, blowing through the equity and mezzanine tranches and impacting senior holders who thought they were protected. The credit enhancement is only as good as the loss model behind it.

Historical Case Study: The 2007-2008 Subprime Meltdown

In April 2007, New Century Financial—then one of the largest subprime mortgage originators in the United States—filed for bankruptcy. This was the first domino.

The timeline of destruction:

• Early 2007: Subprime delinquencies begin spiking; Bear Stearns' two mortgage hedge funds collapse in June-July 2007, losing virtually all investor capital
• Late 2007: Over 100 mortgage lenders close; rating agencies begin mass downgrades of subprime MBS tranches
• March 2008: Bear Stearns collapses, sold to JPMorgan Chase for $2 per share (down from $172 eighteen months earlier)
• September 2008: Lehman Brothers files for bankruptcy; Fannie Mae and Freddie Mac are placed into conservatorship
• By mid-2008: Global financial institutions had written down approximately $501 billion in subprime-related securities (Source: Federal Reserve History, "Subprime Mortgage Crisis")

Rating agency performance: Between autumn 2007 and mid-2008, agencies downgraded nearly $2 trillion in MBS tranches. By end of 2008, 80% of CDOs by value that were originally rated AAA had been downgraded to junk (Source: Financial Crisis Inquiry Commission Report, 2011).

The practical antidote: the crisis didn't invalidate MBS as an asset class—it exposed what happens when originators abandon underwriting standards, rating agencies rely on flawed loss models, and investors chase yield without understanding collateral quality. Agency MBS, backed by government guarantees, never missed a payment. The catastrophic losses were concentrated in non-agency subprime and Alt-A securitizations where no government guarantee existed.

Key Metrics Every MBS Investor Must Track

The test: if you can't tell someone the current CPR, WAL, and OAS of every MBS position in your portfolio without looking it up, you're managing the position passively—which means prepayment surprises will find you before you find them.

MBS vs. Corporate Bonds vs. Treasuries (What You're Trading Off)

The point is: MBS offer a meaningful spread over Treasuries with minimal credit risk, but you pay for that spread with cash flow uncertainty and negative convexity. That trade-off is favorable in stable rate environments and punishing in volatile ones.

Practical Checklist: Evaluating an MBS Position

Essential (Do These First)

• Identify the guarantor (Ginnie Mae, Fannie Mae, Freddie Mac, or private-label)—this determines your credit risk
• Run prepayment scenarios at minimum 3 speeds (slow, base, fast) and calculate yield, WAL, and duration at each
• Calculate OAS using a recognized prepayment model—this is the only spread metric that accounts for the borrower's prepayment option
• Check the pool's WAC relative to current mortgage rates—if the WAC is 150+ bps above market, expect elevated prepayments

High-Impact (Do These for Significant Positions)

• Analyze the borrower composition (FICO distribution, LTV distribution, geographic concentration)
• Stress-test for extension risk by modeling a 200+ bps rate increase scenario
• Compare tranche type (sequential, PAC, companion, support) and understand its specific prepayment exposure
• Monitor monthly factor reports for actual vs. projected prepayment speeds

Optional (For Advanced Portfolio Management)

• Model convexity cost explicitly and compare to the spread earned
• Track specified pool pay-ups for favorable prepayment characteristics (low loan balance, high LTV, investor properties)
• Analyze roll-down return in the TBA market for carry optimization

Your Concrete Next Step

Pull up the largest MBS position in your portfolio (or a benchmark MBS ETF like MBB or VMBS) and run three prepayment scenarios: 100% PSA, 200% PSA, and 50% PSA. Calculate the WAL and yield under each scenario. If the WAL swing exceeds 5 years between scenarios, you're holding more duration risk than you probably intended—and that's exactly the kind of insight that prevents losses before they happen.