The subprime origination market has been transformed in the past six months. Housing price declines, tightening underwriting standards from Wall St conduits, and rising foreclosures led to an evaporation of the subprime securitization. With the securitization execution drying up, subprime originations have declined from roughly 30% to 5% on total originations. What are the implications for the market valuations of subprime MSRs and Residual Interests? These mortgage assets are valued on projected and market consensus collateral behavior. How significant is historical or regression analysis of prepayments and foreclosure frequency if the underlying subprime origination market has been nearly eliminated? Clearly voluntary prepayments or refi’s will decline relative to their historical behavior and, as we are all seeing, involuntary prepayments (foreclosures) and loss severities are dramatically increasing. What are the projected or market consensus for future voluntary and involuntary prepayment vectors? Let’s examine the effect on subprime MSRs and Residual Interests valuations and market prices as projected and consensus assumptions as voluntary prepayments decline and involuntary prepayments increase.
As all mortgage market participants know, subprime and Alt-A MSRs and Residual Interests are priced on projected and market consensus collateral behavior. At present, MIAC provides third-party independent valuations and/or licenses our subprime MSR and Residual pricing model (MIAC Analytics) to the nine largest subprime servicers. Because MIAC is the leading provider of independent third-party subprime asset valuations for subprime and Alt-A mortgage assets, MIAC has been able to develop market consensus foreclosure frequency curves for the standard subprime and Alt-A loan product types.
The key question facing subprime and Alt-A asset holders is to what degree will voluntary prepayments will be adjusted downward with the elimination of the subprime origination. Borrowers will still voluntarily prepay for interest insensitive motives such as housing turnover but the question is how much slower will the voluntary prepays of different subprime mortgage products be. With the subprime market turmoil, MIAC has revised our market consensus prepayment and foreclosure vectors, but because of MIAC’s widespread market share, MIAC is able to establish market consensus vectors with a recognizable creditability.
As with previous mortgage product cycles, the strength of housing led to many new mortgage products in the subprime arena. The influx of highly leveraged loans; ARMS with below market initial rates and potentially severe reset rates; interest only products; longer amortization term loans; and limited and no-doc loans led to 2006 and 2007 subprime performance being the worse on record. MIAC compared subprime foreclosure performance from the recessionary times of 2000 - 2001 to 2006 - 2007 originations. On an age-adjusted basis, subprime delinquency/foreclosure performance is worse on 2006 and 2007 originations than any vintage since 2000. By combining the more accommodating underwriting standards of the past few years coupled with the limited equity buildup with slowing to negative home price appreciation of the last two years, one can see why we are experiencing such broad performance deterioration.
Part of the problem was higher degree of leverage of the newer originations coupled with deteriorating underwriting standards that fed the origination platform in its quest to continue/increase market share. The flight to quality into the safe haven of Treasuries and Agency product has had broad liquidity implications on the nonagency markets. Warehouse lines have been curtailed, margin calls have been fierce, secondary pricing execution for all products outside of the agency products have been crushed. Spreads between product classes and even mortgage market rates of top tier credit for Jumbo loans have widened considerably. Alt-A product rates have widened to previous subprime market rates. Subprime underwriting has tightened to be primarily LTV driven and spreads have significantly widened.
Loan Sector Spreads are Dramatically Wider:
Spreads between Prime and Jumbo have widened from 25 bps to 75 –100 bps
Spreads between Prime and ALT-A have widened from 75 bps to 100 –150 bps
Spreads between Prime and Subprime have widened from 100-150 bps to 250 –300 bps
CMO Spreads between Credit Grades have widened Significantly:
Fixed-Rate CMO Spreads to 1-Mth LIBOR:
2 Yr AAA From 40 bps to 150 bps
3 Yr AAA From 65 bps to 175 bps
5 Yr AAA From 85 bps to 200 bps
7 Yr AAA From 105 bps to 225 bps
5 Yr A From 160 bps to 600 bps
5 Yr BBB From 350 bps to 1500 bps
Floating-Rate CMO Spreads t0 1-Mth LIBOR:
3 Yr AAA From 20 bps to 250 bps
5 Yr AA From 30 bps to 750 bps
5 Yr A From 50 bps to 1400 bps
5 Yr BBB From 250 bps to 2200 bps
The majority of recent underperformance of subprime product has been concentrated in the adjustable rate product through all vintage years. Impacts of leverage (High CLTV, Seconds, High DTI’s) and lower credit tiers have also impacted performance and MSR valuations substantially. When reviewing subprime collateral performance, the percentage of foreclosures in lower tiered credit is substantially greater with great impact on values.
To demonstrate the impact on MSR pricing of the migration from voluntary prepayments to involuntary prepayments, we used a benchmark subprime MSR portfolio in which we modified the previous market consensus foreclosure curves to project higher foreclosure probabilities over the lifetime of the MSR portfolio asset. In our sensitivity analysis, we observed changes in the values at different levels of foreclosures. We adjusted our base foreclosure rate in increments of 120%, 150% and 200% to see the impact on the MSR asset. We isolated the major cash flow components and their respective impact on servicing values. Additionally, we held the total prepayment projection constant with a re-allocation of the voluntary and involuntary prepayments. In the increased foreclosure scenarios, involuntary prepayments went up at the expense of voluntary prepayments.
From the revenue side, increases in the foreclosure probability resulted in lower servicing income streams and lower prepayment float income. This is due to the lower percentage of performing loans generating service fee income. Additionally, with fewer voluntary prepayments, prepayment float income is reduced. In our various sensitivities, we witnessed a reduction in income streams of approximately 2 to 12 basis points on the revenue streams due to reduced service fee and prepayment float income in the 120% to 200% foreclosure scenarios.
The MSR expense side had greater sensitivity to the higher projected foreclosures. In the sensitivity analysis, only modifications to foreclosure frequencies and not delinquency rates were shocked. In general, we would expect some increases in the 30-day, 60-day and 90-day delinquency buckets to be somewhat offset by increased late fee income streams. By holding these delinquency buckets constant and only modifying the foreclosure probabilities, the major components impacted on the expense side centered around the increase in foreclosure costs and the increase in interest advances on principal and interest. The MSR prices decreased from 5 to 16 basis points due to the foreclosure scenarios. In the scenario in which foreclosure rates increased twofold, the total impact on MSR servicing values was greater than 25 basis points of value. This is quite a dramatic effect being that subprime MSR’s trade much lower than prime agency MSRs. The valuation depreciation was most pronounced on the ARM products that historically have much higher foreclosure rates versus fixed-rate products.
Component MSR Price Impact: