Will Carvana's repo financing collateral cushion on beneficial interests remain above 15% as of Q1 2026?

The balance sheet improvements since the original prediction are uniformly positive for maintaining the repo collateral cushion above 15%. Net debt/EBITDA declined from 1.5x to 1.3x, cash increased to $2.327B, and loan purchase capacity expanded from $14B to $18B across 5 partners. PIK interest elimination means all $505M in interest is cash-pay, confirming real cash generation. Interest coverage at 4.3x is strong. The original 24% cushion at Q3 2025 was already well above the 15% threshold, and nothing in the Q4 data suggests deterioration. The key unknown — actual Q4/FY2025 repo and beneficial interest figures — is the main source of residual uncertainty. However, the broad balance sheet strengthening makes cushion compression very unlikely absent a systemic ABS market shock. With debt maturities pushed to 2028-2031, there is no near-term pressure on financing structures. Moving from 86% to ~89% reflects the modest positive signal without over-adjusting given the missing granular repo data.

The updated data confirms and strengthens the base case from the original prediction. The original 86% was driven by the 24% cushion (well above 15%), substantial cash reserves, and 1.5x leverage. All three factors have improved: cushion presumably stable or better (no negative signals), cash up $227M to $2.327B, leverage down to 1.3x. The expansion from 3 to 5 loan purchase partners ($18B capacity) directly reduces the counterparty concentration risk that was the main downside scenario. PIK elimination is a strong positive signal — it means the company is generating enough cash to service all debt in cash, removing the concern about paper-only profitability. The stress-scanner's STRETCHED classification is notable but focuses on structural vulnerabilities (sub-IG pricing, short track record) rather than near-term repo cushion risk. The Q1 2026 resolution window is only ~6 weeks from now, limiting the time for adverse developments.

I weight the balance sheet improvements more heavily than the other runs. The key insight is that the cushion dropping below 15% would require either (a) a ~38% decline in beneficial interest fair values while repo borrowings stay constant, or (b) a large increase in repo borrowings without corresponding collateral growth. Scenario (a) requires a massive increase in discount rates — from 9.7% to well above 12-13% — which would imply a severe credit market dislocation. With interest coverage at 4.3x, $2.3B cash, 1.3x leverage, and no near-term maturities, Carvana is better positioned to withstand such stress than at any point in its history. Scenario (b) would require management to deliberately increase repo leverage, which contradicts the DISCIPLINED capital deployment pattern. The calibration data showing 0.0324 Brier on the OCF/NI market suggests the ensemble was well-calibrated for CVNA, and the original 86% was already appropriate — the update warrants a modest increase.

The Q4 2025 earnings data is broadly positive for the repo collateral cushion remaining above 15%. The original 24% cushion provided a 9pp buffer above the threshold, and the balance sheet improvements (lower leverage, more cash, expanded loan partners) reduce the probability of cushion compression. The elimination of PIK interest is particularly significant — it removes any question about whether the company's improved financial position is paper-only. With $2.327B in cash and 4.3x interest coverage, Carvana has substantial resources to maintain or strengthen its repo facilities. The main residual risk is a systemic ABS market dislocation that increases discount rates significantly, but the expanded partner base ($18B from 5 partners) provides diversification against this. The STRETCHED classification from the stress-scanner reflects structural concerns about the business model's dependence on securitization markets, not an imminent risk to the repo cushion specifically.

I am slightly more cautious than some runs. While the balance sheet improvements are real, the specific repo/beneficial interest data from Q4 2025 is missing. The original prediction was already built on Q3 figures, and the Q4/FY2025 10-K could reveal changes to the repo facility structure, new beneficial interest valuations, or modified discount rate assumptions. The STRETCHED classification from the stress-scanner, while not directly about repo cushion, highlights that Carvana's sub-investment-grade debt pricing and short profitability track record mean the market still views the credit as risky — which could translate to wider spreads on repo facilities. Additionally, Q4 EBITDA margin declined to 9.1% from 10.1%, suggesting reconditioning headwinds that could affect the quality of underlying auto loans in ABS structures. That said, the 24% starting cushion and improved overall financial position still make a drop below 15% quite unlikely.

The resolution date for this market is Q1 2026 (March 31), disclosed in the Q1 2026 10-Q (expected ~May 2026). This means we are assessing whether the cushion will stay above 15% roughly 6 weeks from now. The extremely short time horizon is critical — there is very little time for a significant deterioration. The balance sheet is in better shape than at the original prediction date, and there are no signals of ABS market stress, rising interest rates, or deteriorating auto loan performance that would compress the cushion. The expanded loan purchase capacity and partner diversification reduce the tail risk of counterparty-driven margin calls. The Q4 margin decline is a mild negative but relates to reconditioning costs, not credit quality. At the prior 86% this was already well-calibrated; the modestly positive update supports ~89%.

The balance sheet improvements are clear and directionally positive: lower leverage, more cash, more partners, PIK eliminated. The original 24% cushion was well above 15%, and nothing in the Q4 data suggests deterioration. With only ~6 weeks until the Q1 2026 measurement date, the probability of a shock large enough to compress the cushion by 9pp is very low. Interest coverage at 4.3x and no near-term maturities further reduce risk.

The update is modestly positive. Key improvements include cash increasing to $2.327B, leverage declining to 1.3x, loan purchase capacity expanding to $18B, and PIK interest fully eliminated. These factors strengthen the overall financial position and reduce the likelihood of repo cushion stress. However, the actual Q4 repo and beneficial interest figures are unknown, which introduces some uncertainty. The original 86% was already high confidence; an increase to ~88% reflects the positive signal without overreacting to incomplete data.

With the balance sheet strengthened across all dimensions and the Q1 2026 resolution window only weeks away, a cushion compression from 24% to below 15% would require a rapid, severe shock — a sudden ABS market freeze, dramatic interest rate spike, or unexpected deterioration in auto loan collateral quality. None of these appear imminent based on available data. The expanded loan partner base and elimination of PIK interest further insulate against tail risks. Adjusting modestly upward from 86% is warranted.