Mortgage Rates End Week at Highs

Mortgage rates, while technically the lowest since early 2023 if excluding the past five days, rose modestly day-over-day and are slightly higher versus last week after an earlier drop. Last week's move was driven by Fannie Mae and Freddie Mac announcing roughly $200 billion of MBS purchases, which sharply lowered average mortgage rates but had largely run its course by Monday. Offsetting that support, Treasury yields and Fed rate expectations have been climbing, with the 10-year Treasury breaking a four-month range today — a development that would normally push mortgage rates higher absent the GSE MBS intervention.

Analysis

Market structure: Fannie/Freddie’s $200B MBS purchases are the marginal buyer that compresses agency MBS yields and directly benefits agency-MBS holders (MBB, VMBS) and leveraged mortgage REITs (AGNC, NLY) through tighter spreads and higher carry. Losers are long-duration Treasuries and mortgage originators whose margins compress when rates fall; the recent 10yr breakout removes a tailwind and implies position fragility if GSE flows abate. Supply/demand: the GSE bid temporarily shifts the balance from supply-dominant to demand-dominant up to the $200B cap, but the effect is finite and front-loaded into weeks–months rather than permanent.

Risk assessment: immediate (days) risk is continued Treasury-driven repricing—watch a >10–20bp move above the 4-month range high as a trigger for mortgage rates to re-align with Treasuries. Short-term (weeks–months) tail risks include FHFA/GSE policy rollback or political pushback that curtails purchases; long-term (quarters) the dominant risk is Fed terminal rate surprise that re-prices duration across MBS and Treasuries. Hidden dependencies: heavy GSE buying reduces liquidity and increases convexity/prepayment exposure for holders, amplifying downside if purchases stop.

Trade implications: primary tactical trade is long agency MBS exposure via MBB/VMBS sized 2–4% of portfolio funded by a short 10yr Treasury futures position to isolate spread compression (target net duration ~0). For mortgage REITs (AGNC, NLY) prefer protected exposure: buy 6–8 week put spreads (e.g., AGNC SEP month 2–4% OTM) sized 1–2% notional to hedge duration and prepayment shocks. Rotate modestly into homebuilders (PHM, DHI) on a confirmed 30yr mortgage <6% for 4–8 week window; trim long-duration sovereigns (TLT, IEI) if 10yr holds above the recent breakout by >15bp.

Contrarian angles: consensus treats GSE buying as a permanent backstop; history (2013 taper tantrum, 2020 QE front-loading) shows market reaction is asymmetric once purchases slow—MBS can underperform quickly. The market may be underpricing the risk that a Treasury-driven rate selloff will overwhelm GSE buys; this creates a mispricing to short agency MBS or sell mortgage REIT rallies if FHFA hints at tapering. Unintended consequence: sustained buybacks inflate prepayment risk, starving long-term carry and punishing leveraged REITs when rates move up, so convexity-aware positioning is essential.

Key Decisions for Investors

• Establish a 2–4% portfolio long in agency MBS via MBB or VMBS over the next 2 weeks to capture spread compression, funded by a 1–2% short position in 10‑year futures (or equivalent Treasuries) to target neutral net duration and isolate MBS carry.
• Initiate a hedged 1–2% tactical long in AGNC or NLY using buy-write or put-spread protection: buy 6–8 week OTM put spreads (e.g., 3–5% OTM) to limit downside if 10yr rises >15bp from current breakout levels.
• Pair trade: long PHM and DHI (1–2% each) vs short regional mortgage originators or bank mortgage pipelines (e.g., a basket of regional banks) if 30‑yr fixed reverts to <6% within 4 weeks; exit if 10yr > prior range high +20bp or mortgage 30‑yr >6.5%.
• Prepare a tactical short of agency MBS (sell MBB) sized 1–2% if FHFA/GSE communications indicate tapering or if 10‑year Treasury rises >20bp from the post-breakout level within 10 trading days—this is the trigger to monetize the crowding risk.
• Reduce long-duration sovereign bond exposure (TLT, IEI) by 3–5% and reallocate to short-dated cash/floaters if 10yr remains above the breakout threshold (defined as >15bp above the 4‑month range) for more than 2 weeks.