Will U.S. existing home sales decline more than 10% year-over-year in any month through September 2026?

The January 2026 data is informative but cuts both ways. The -4.4% YoY reading shows we're nowhere near -10% in normal conditions, even with weather disruption. However, the question only needs ONE month out of eight (Feb-Sep) to breach -10%. The key risk vector is an exogenous shock: mortgage rate spike above 7.5% on inflation fears, a tariff-induced economic slowdown, or a financial market dislocation. Mortgage rates at 3-year lows is actually a positive signal — it means the housing market has a tailwind that would need to reverse. The comparison base (2025) was already weak, making -10% a high bar. Historical precedent shows that even in 2022's sharp downturn, monthly YoY declines were driven by rapid rate increases from 3% to 7%. A similar rate shock from current levels would need rates to jump from ~6% to ~9%, which is extremely unlikely. I'm moderately lowering my estimate — the current trajectory is favorable and the comparison base is easy.

I want to stress-test the favorable narrative. The MBA purchase applications at -14% in late January/early February is a leading indicator that the SAAR data may not yet reflect. If purchase applications stay depressed through February due to ongoing weather and economic uncertainty (tariffs, policy volatility), the February SAAR could come in worse than January. However, -14% in MBA apps doesn't directly translate to -10% SAAR YoY — the relationship is noisy. More concerning is the 8-month window: any single macro shock (trade war escalation, recession signal, banking stress) could create a -10% month. The base rate for at least one month of -10% YoY decline in a sluggish but stabilizing market over 8 months is probably 15-25%. Tariff uncertainty is a genuine wildcard — the Q1 2025 experience showed tariff fears caused March weakness. If tariffs escalate in 2026, summer months could be affected. I weight this slightly higher than pure-weather-blip analysis suggests.

The structural dynamics create an interesting tension. On one hand, mortgage rates at 3-year lows and growing inventory should support volume recovery — these are fundamentally positive conditions. On the other hand, the lock-in effect from low-rate legacy mortgages continues to suppress existing home supply, and the 2025 comparison base, while weak, isn't as easy as 2023-2024 because there was some recovery in H2 2025. The question is specifically about existing home sales (NAR SAAR), which is more volatile than new home sales and more susceptible to weather and seasonal anomalies. A single weather-disrupted month (like February with continued Winter Storm effects) or a single policy-shock month (tariff announcement causing buyer paralysis) could breach -10%. I think 20% properly captures the tail risk over 8 months while acknowledging the fundamentally favorable trajectory.

Base rate analysis: In the past 30 years, existing home sales have shown >10% YoY monthly declines during recessions and rate shock periods (2006-2008, 2022). Current conditions (low rates, growing inventory) don't resemble those periods. However, the 8-month window means we need to assess the probability of NO month breaching -10%, not just the expected average. If each month has a ~3% independent probability of a -10% shock, the probability of at least one in 8 months is 1 - 0.97^8 = 22%. January at -4.4% suggests the monthly shock probability is low but non-zero. I'll use approximately 20% as a reasonable estimate for at least one -10% month.

The initial prediction of 28% reflected higher uncertainty about housing market trajectory. The earnings update provides useful calibration: January at -4.4% YoY with adverse weather conditions is reassuring for the base case. However, the initial 28% also reflected macro tail risks (tariffs, recession, rate spikes) that haven't been resolved — they're not addressed by a single month of data. Mortgage rates at 3-year lows could reverse if inflation resurges or the Fed signals hawkishness. The improvement from 28% to ~22% reflects the favorable January data point and low rates, while maintaining appropriate weight on 8-month tail risk.

Looking at the specific months: February could be affected by continued winter weather but comps from Feb 2025 were also weak (not an easy comp to miss by 10%). March through May is spring selling season with historically strong volume — very unlikely to see -10% here. June-September is summer with moderate volume. The most vulnerable months are February (weather) and August-September (if a summer macro shock occurs). With only 2-3 genuinely vulnerable months out of 8, and each needing a significant adverse event, I estimate ~18% probability. The growing inventory and low rates create a structural floor that makes even vulnerable months unlikely to breach -10%.

January was -4.4% with bad weather. Mortgage rates are low. Inventory is growing. Spring should be strong. But 8 months is a long window — tariffs or recession could cause a bad month. Previous estimate of 28% was reasonable pre-data; now 20% reflects the favorable January reading and rate environment.

The -4.4% January reading with Winter Storm Fern suggests that even significant weather disruption doesn't get close to -10%. Mortgage rates at 3-year lows are a strong tailwind. The 2025 comparison base was weak (near historical lows), making a further -10% very difficult. Spring selling season should be favorable. Main risk is a macro shock but current trajectory is strongly supportive.

MBA purchase apps at -14% is concerning — it's a leading indicator that could foreshadow worse SAAR data. But weekly MBA data is noisy and doesn't directly map to monthly SAAR. Tariff risk is real based on Q1 2025 precedent. 8 months is enough time for something to go wrong. But the structural environment (low rates, easy comps, growing inventory) suggests 20-22% is the right range.