CONVEX
Glossary/Fixed Income & Credit/Mortgage-Backed Securities
Fixed Income & Credit
8 min readUpdated Apr 12, 2026

Mortgage-Backed Securities

ByConvex Research Desk·Edited byBen Bleier·
MBSmortgage-backed securitiesagency MBSRMBSpass-through securities

Bonds backed by pools of residential or commercial mortgages, held in massive quantities by the Fed as part of QE programs, their runoff is a key component of quantitative tightening.

Continue reading on Convex
Current Macro RegimeSTAGFLATIONDEEPENING

The macro regime is unambiguously STAGFLATION DEEPENING. The hot CPI print (pending event, 24h ago) is not a surprise — it is a CONFIRMATION of the pipeline signals that have been building for weeks: PPI accelerating faster than CPI, Cleveland nowcast at 5.28%, breakevens rising +10bp 1M across the …

Analysis from May 14, 2026

What Are Mortgage-Backed Securities?

Mortgage-backed securities (MBS) are bonds created by pooling thousands of individual home mortgages into a single tradeable security. Homeowners' monthly mortgage payments, interest and principal, flow through to bondholders as cash flows. The US agency MBS market is approximately $9 trillion outstanding, making it the second-largest fixed-income market in the world after US Treasuries.

MBS are the backbone of the American housing finance system: over 70% of US mortgages are securitized into MBS, which are then sold to investors (pension funds, insurance companies, banks, the Fed, and foreign governments). Without the MBS market, mortgage rates would be significantly higher and home ownership rates significantly lower.

MBS are also one of the most complex instruments in fixed income because of their unique risk characteristics, particularly negative convexity and prepayment risk, which create hedging dynamics that amplify interest rate moves and directly influence the Treasury market, mortgage rates, and financial conditions.

How MBS Are Created: The Securitization Chain

Step 1: Mortgage Origination

A homebuyer takes out a 30-year fixed-rate mortgage from a bank (Wells Fargo, JPMorgan, etc.) or a non-bank lender (Rocket Mortgage, United Wholesale Mortgage).

Step 2: Sale to GSE

The lender sells the mortgage to a government-sponsored enterprise (Fannie Mae or Freddie Mac) or to Ginnie Mae (for FHA/VA loans). The GSE pays the lender par value, freeing up the lender's capital to make more loans.

Step 3: Pooling

The GSE pools thousands of similar mortgages (same approximate rate, region, loan type) into a single security, an MBS "pass-through" bond.

Step 4: Guarantee

The GSE guarantees timely payment of interest and principal to MBS holders, regardless of whether individual homeowners default. This guarantee converts credit risk into prepayment risk.

Step 5: Sale to Investors

The MBS is sold to institutional investors. Monthly payments from homeowners flow through the GSE to bondholders.

Agency vs. Non-Agency MBS

Feature Agency MBS Non-Agency MBS
Issuer Fannie Mae, Freddie Mac, Ginnie Mae Banks, financial institutions
Credit guarantee Yes (GSE or government) No
Credit risk to investor None (GSE absorbs defaults) Full (defaults reduce cash flows)
Market size ~$9 trillion ~$500 billion (shrank post-GFC)
Primary risk Interest rate + prepayment Credit + interest rate + prepayment
Liquidity High (TBA market is extremely liquid) Low to moderate
Role in 2008 crisis Survived (government backstop) Epicenter of the crisis

Negative Convexity: The Defining MBS Risk

Negative convexity is what makes MBS fundamentally different from every other bond type. It means that MBS holders are systematically disadvantaged in both rising and falling rate environments:

When Rates Fall (Prepayment Risk)

Homeowners refinance their mortgages at lower rates, returning principal to MBS holders at par. The MBS investor receives money back when they don't want it, in a low-rate environment where reinvestment options are poor. The bond's duration shortens (less sensitivity to further rallies) just when you'd want longer duration to capture gains.

When Rates Rise (Extension Risk)

Nobody refinances. The MBS investor is stuck holding below-market-rate bonds for much longer than expected. Duration extends (more sensitivity to further selloffs) just when you'd want shorter duration to limit losses.

The Hedging Feedback Loop

This negative convexity forces MBS holders to hedge dynamically:

  • Rates fall → MBS duration shortens → hedge by buying Treasuries (to add duration back)
  • Rates rise → MBS duration extends → hedge by selling Treasuries (to reduce duration)

These hedging flows are pro-cyclical: they amplify rate moves in both directions. When the $9 trillion MBS market collectively adjusts hedges, the forced Treasury buying/selling creates significant price pressure. This is why large rate moves tend to accelerate rather than mean-revert, MBS hedging creates momentum.

Prepayment Speeds: The Key Variable

CPR (Conditional Prepayment Rate) Environment Investor Impact
< 5% High rates, lock-in effect Extension risk, duration grows, stuck in low-coupon bonds
5-15% Normal, moderate rates Moderate, manageable prepayment and extension
15-30% Falling rates, active refi market High prepayment, principal returned early, reinvestment risk
> 30% Rapid rate decline, refi boom Extreme prepayment, bonds "called away" at the worst time

The 2020-2021 Refi Boom

When 30-year mortgage rates fell to 2.65% (January 2021), prepayment speeds surged to 40%+ CPR for some pools. MBS investors faced massive reinvestment risk: they received billions in prepaid principal that they could only reinvest at near-zero yields.

The 2022-2024 Lock-In Effect

The mirror image: with mortgage rates above 6.5% and most outstanding mortgages originated at 2.5-4.0%, prepayment speeds collapsed to historic lows (CPR ~4-6%). Homeowners with 3% mortgages will not refinance to 7%. This created:

  • Extended MBS durations, making MBS more rate-sensitive
  • Slower Fed QT MBS runoff, only $15-20B/month vs. the $35B cap
  • Housing market freeze, homeowners won't sell (lose their low rate), reducing housing inventory

The Fed and MBS: $2.7 Trillion of Influence

How the Fed Accumulated Its MBS Portfolio

QE Round Period MBS Purchases Purpose
QE1 2008-2010 $1.25 trillion Stabilize housing market post-GFC
QE2 2010-2011 Treasuries only ,
QE3 2012-2014 $40B/month initially, up to $85B (MBS + Treasuries) Lower mortgage rates, stimulate housing
COVID QE 2020-2022 $40B/month MBS + $80B Treasuries Crisis response
Peak Early 2022 ~$2.7 trillion MBS on balance sheet ,

The QT MBS Runoff Challenge

The Fed's QT plan allows up to $35 billion/month in MBS to mature or prepay without reinvestment. However, actual runoff has been much slower (~$15-20 billion/month) because:

  1. Prepayment speeds are at historic lows (lock-in effect)
  2. The Fed does not actively sell MBS (only passive runoff)
  3. The remaining portfolio is dominated by low-coupon bonds that nobody is refinancing

This slow MBS runoff is a key reason the Fed's balance sheet is declining more slowly than planned, and why some FOMC members have discussed actively selling MBS to accelerate the process.

MBS Spread Impact

The Fed's MBS purchases compressed the "primary mortgage spread", the gap between the 30-year mortgage rate and the 10-year Treasury yield, from a normal 170-180 bps to as low as 100 bps during peak QE. As the Fed withdraws, this spread has widened back to 250-300+ bps, adding approximately 70-130 bps to mortgage rates beyond what Treasury yields alone would dictate.

MBS and the 2008 Financial Crisis

The subprime MBS crisis was the proximate cause of the worst financial disaster since the Great Depression:

The Chain of Failure

Phase Period Event
1. Origination boom 2003-2006 Subprime/Alt-A mortgages issued to unqualified borrowers; "NINJA" loans (No Income, No Job, No Assets)
2. Securitization machine 2004-2007 $2+ trillion in non-agency MBS issued; AAA ratings on fundamentally risky pools
3. Housing peak Q2 2006 Home prices stop rising; adjustable-rate mortgages begin resetting higher
4. Default cascade 2007-2008 Subprime default rates surge from 5% to 20%+; AAA-rated MBS tranches start losing value
5. Financial system seizure Sep-Nov 2008 Bear Stearns collapses (Mar 2008), Lehman Brothers bankrupt (Sep 2008), AIG bailout ($182B), money market fund "breaks the buck"
6. Global recession 2008-2009 GDP -4.3%, unemployment 10%, $2+ trillion in global losses from MBS-related exposures

Why Rating Agencies Failed

Rating agencies (S&P, Moody's, Fitch) assigned AAA ratings to senior tranches of subprime MBS pools based on models that assumed: (1) housing prices could not decline nationally, (2) mortgage defaults in different regions were uncorrelated, and (3) historical default rates (from a period of rising home prices) were representative. All three assumptions were catastrophically wrong. The result: institutional investors who thought they held AAA bonds (equivalent to Treasuries in safety) actually held instruments that would lose 30-90% of their value.

MBS Trading and Investment

Key Instruments

Instrument Ticker/Type Exposure Liquidity
TBA (To-Be-Announced) OTC market Forward delivery of agency MBS Extremely high (~$300B/day)
MBS ETFs MBB (iShares), VMBS (Vanguard) Agency MBS portfolio High
CMO tranches Various Structured MBS with specific risk profiles Low-moderate
Non-agency RMBS Various Private-label MBS with credit risk Low

The TBA market is the primary mechanism for MBS trading, it allows forward delivery of generic MBS pools without specifying the exact bonds. This forward mechanism is what makes the MBS market so liquid despite the heterogeneity of the underlying mortgages.

MBS as a Spread Product

MBS investors earn a spread over Treasuries that compensates for:

  1. Prepayment risk (~30-50 bps in normal environments)
  2. Negative convexity (~20-40 bps)
  3. Liquidity premium (~10-20 bps for less liquid pools)

Total MBS spread over Treasuries: typically 60-120 bps in normal markets, 150-300+ bps during stress. For income-focused investors, MBS offers a yield pickup over Treasuries with agency-guaranteed credit quality, the trade-off is the complexity of prepayment and convexity risk.

Why MBS Matter for Every Investor

Even if you never buy an MBS, the market affects you through:

  1. Mortgage rates: MBS yields directly determine the 30-year fixed mortgage rate that most Americans pay
  2. Housing affordability: MBS spread widening raises mortgage rates independent of Fed policy
  3. Treasury market dynamics: MBS hedging flows amplify Treasury rate moves
  4. Fed policy: The pace of MBS QT affects financial conditions and the liquidity cycle
  5. Systemic risk: The MBS market's size ($9T) and complexity make it a perpetual source of systemic risk during stress

Frequently Asked Questions

What is the difference between agency and non-agency MBS?
Agency MBS are guaranteed by government-sponsored enterprises (Fannie Mae, Freddie Mac) or the full-faith-and-credit government agency Ginnie Mae. This guarantee means the bondholder bears no credit risk — if homeowners default, the GSE covers the payments. Agency MBS therefore trade based purely on interest rate risk and prepayment risk, like a slightly more complex Treasury. The agency MBS market is approximately $9 trillion outstanding — the second-largest fixed-income market after Treasuries. Non-agency (or "private-label") MBS are issued by banks and financial institutions without government backing. The bondholder bears both credit risk (homeowner defaults reduce cash flows) and prepayment risk. Non-agency MBS were at the center of the 2008 financial crisis: subprime and Alt-A non-agency MBS experienced default rates of 20-40%, causing trillions in losses. The non-agency market has shrunk dramatically since the GFC (from ~$2 trillion to ~$500 billion outstanding) as regulatory changes and investor wariness reduced issuance. For most market analysis, "MBS" refers to agency MBS unless specified otherwise — and agency MBS is what the Fed holds on its balance sheet.
What is negative convexity and why does it make MBS uniquely dangerous?
Negative convexity is the defining characteristic that makes MBS fundamentally different from Treasuries. In a normal bond (positive convexity), price rises accelerate when yields fall and price drops decelerate when yields rise — a favorable asymmetry. MBS exhibit the opposite: when yields fall, homeowners refinance their mortgages, returning principal early. This "prepayment risk" shortens the bond's effective duration just when you want it longer (to benefit from the rally). When yields rise, nobody refinances, and the bond's duration extends — getting longer just when you want it shorter (to limit losses). The result: MBS holders lose in both directions. They capture less upside when rates fall (prepayments shorten the bond) and suffer more downside when rates rise (extension lengthens the bond). This negative convexity requires constant hedging: MBS holders must sell Treasuries when rates rise (to reduce duration) and buy when rates fall (to increase duration). These hedging flows are pro-cyclical — they amplify rate moves in both directions. When the $9 trillion MBS market hedges collectively, the forced Treasury selling/buying can move the entire rate market, creating a feedback loop that is one of the most powerful forces in fixed income.
How does the Fed's MBS portfolio affect mortgage rates?
The Fed accumulated approximately $2.7 trillion in agency MBS through successive rounds of QE (QE1, QE2, QE3, and COVID-era QE). By absorbing a huge share of MBS supply, the Fed compressed the "MBS spread" — the yield premium MBS trade above Treasuries — from a typical 100-150 bps to as low as 50-70 bps. Lower MBS spreads translate directly to lower mortgage rates because mortgage rates are approximately: Mortgage rate = 10Y Treasury yield + MBS spread + origination margin. The Fed's $2.7 trillion MBS holding reduced mortgage rates by an estimated 50-100 bps below where they would have been without QE — roughly the difference between a 6.5% and 7.0-7.5% mortgage rate. As the Fed allows MBS to run off through QT (capped at $35 billion/month), the MBS spread gradually widens, putting upward pressure on mortgage rates independent of Treasury yield movements. However, QT MBS runoff is much slower than expected because with mortgage rates above 6.5%, almost no one refinances — the "lock-in effect" means actual MBS prepayments are running at historically low speeds, and the Fed's MBS portfolio is declining by only $15-20 billion/month vs. the $35 billion cap.
Why did MBS cause the 2008 financial crisis?
The 2008 crisis was fundamentally an MBS crisis. The chain of causation: (1) From 2003-2006, banks originated trillions of dollars in subprime and Alt-A mortgages to borrowers with weak credit, low documentation, and adjustable rates. (2) These mortgages were securitized into non-agency MBS, then sliced into tranches: senior tranches rated AAA by rating agencies, mezzanine tranches rated A/BBB, and equity tranches (unrated). (3) Rating agencies dramatically underestimated default correlations — they assumed housing defaults in Florida, Nevada, and California were independent events, when in reality they were driven by the same national bubble. (4) When housing prices stopped rising in 2006 and adjustable-rate mortgages reset higher, defaults surged — first in subprime, then spreading to Alt-A and prime. (5) MBS tranches that were rated AAA suffered catastrophic losses. Investors who thought they held safe bonds realized they held toxic waste. (6) Banks, insurance companies (AIG), and money market funds that held these MBS suffered enormous losses. Lehman Brothers held $85 billion in MBS-related assets. Bear Stearns' two MBS hedge funds collapsed in June 2007. (7) The entire financial system seized as no one could assess the MBS exposure of their counterparties — credit markets froze, banks stopped lending to each other, and the global economy entered the worst downturn since the Great Depression. Total estimated losses: $2+ trillion globally.
How do MBS prepayment speeds affect bond investors?
Prepayment speed — the rate at which homeowners pay off their mortgages early (through refinancing, home sales, or extra payments) — is the key variable in MBS investing. It is measured in CPR (Conditional Prepayment Rate): the annualized percentage of the remaining mortgage pool that prepays each month. When prepayment speeds are fast (CPR > 30%): homeowners are refinancing aggressively (because rates have fallen). MBS investors receive their principal back early, at par, when they would prefer to hold higher-coupon bonds in a low-rate environment. This is the "reinvestment risk" — you get your money back when yields are low and must reinvest at worse rates. When prepayment speeds are slow (CPR < 5%): rates are high, nobody refinances. MBS investors are stuck holding below-market-coupon bonds with extending durations. As of 2024-2025, prepayment speeds were near historic lows (CPR ~4-6%) because the vast majority of outstanding mortgages were originated at 2.5-4.0% rates during 2020-2021 — no rational homeowner would refinance from 3% to 7%. This "lock-in effect" has dramatically changed MBS market dynamics: duration is extended (making MBS more rate-sensitive), Fed QT MBS runoff is slower than planned, and the housing market has frozen (homeowners won't sell because they'd lose their low rate). The prepayment environment directly affects MBS investors' returns, duration risk, and hedging requirements.

Mortgage-Backed Securities is one of the signals monitored daily in the AI-driven macro analysis on Convex Trading. The platform synthesises data across monetary policy, credit, sentiment, and on-chain metrics to generate actionable trade recommendations. Create a free account to build your own signal layer and see how Mortgage-Backed Securities is influencing current positions.

ShareXRedditLinkedInHN

Macro briefings in your inbox

Daily analysis that explains which glossary signals are firing and why.