Mortgage Subsidies - Arguably Useless But Definitely Expensive

When Reuters yesterday released the rumor of the mortgage subsidy program, the market shot up to close unchanged after being down over 3%. But was the market, as usual, ahead of itself in its optimism? While the news of subsidies may result in some marginal benefits to the most at risk homeowners, it is at best a temporary solution as it does nothing to impact the real toxic asset overhang - the staggering notional size of unserviceable mortgage debt. Additionally, subsidies will obviously not come free. A full estimate of the program's cost can only be made once there is full disclosure on this effort (which aside from Reuters' piece, there has not been any additional confirmation from the administration), but this post attempts to give some rough preliminary indications of how much will need to be invested and whether or not the most recent plan by the administration will even be effective.

First, we present some interesting data from BAC's mortgage group, which tracks the most at-risk groups, and what current roll rates are for assorted cohorts (roll rate is the percentage of current mortgages of specific age in months that defaults within a month), emphasizing the mortgage types mostly associated with higher than usual foreclosure rates: Jumbo ARMs and Options ARMs. The conclusion is that while principal reductions have a favorable impact on reducing delinquency rates and foreclosures, the impact of interest rate reductions is much less certain and likely will have a very muted effect at best.

In the figure below, the dashed black line shows that 1.4% of 24-month Jumbo ARMs in which the borrower has little equity (>= 90% of the Loan To Value (LTV) and where monthly debt payments consumed a large proportion of income (Debt To Income, or DTI, of 45%) become delinquent within a month. The solid black curve shows that for comparable loans, but where the borrower can better afford mortgage payments (DTI of 35%) roll rates decline to 0.8%. The red lines represent borrowers with equity in the home (LTV <90%) - delinquency rates here are lower and much more insensitive to changes in monthly mortgage payments. This demonstrates a mortgage subsidy program would mostly benefit homeowners with Jumbo ARMs and little home equity. Comparing the data for borrowers who spend a large portion of monthly incomes on debt payments (dashed lines) indicates that the biggest benefit would come from principal writedown, as delinquency rates drop significantly for those homeowners who have equity in the home. The brief conclusion is that yesterday's news, even if ultimately confirmed, will be yet another step in the wrong direction by the government, which should be instead focusing on reducing the overall debt burden.

This relationship is even more pronounced when looking at the much more aggressive Option ARM mortgage data. Here the main difference is that in loans with little to no equity (DTI of 45%) there is virtually no roll rate improvement if monthly mortgage payments are made moderately more affordable (DTI dropping to 35%). Therefore, even more pronounced than in the Jumbo ARM scenario, there is no evidence that an interest rate modification program will prevent foreclosures for homeowners with Option ARM mortgages. The only sure way to reduce roll rates, delinquencies and foreclosures (at least based on empirical studies) is to enhance the equity in the mortgage, which can only be achieved via a program of debt principal reduction.

What will it cost?

Market estimates of first-lien mortgages expected to be liquidated in this economic cycle approach $2 trillion (out of $9.3 trillion in outstanding total first-lien mortgages). Assuming the new loan modification program can reach these borrowers, and the cost of lowering mortgage rates by 100 bps is 2-4 points, the total cost of the program would be $40-80 billion per 100 bps reduction. As the targeted group of homeowners currently likely has very high mortgage rates due to high DTV and LTV ratios, the program will likely need to lower mortgage rates by anywhere from 2-4%, suggesting the cost of the program could reach $300 billion (or higher), and there is no conclusive evidence it would even lower foreclosure rates. Alternatively, to effectively achieve the desired effect of lowering delinquencies, thereby lowering LTVs, the government would have to impair the $2 billion first-lien tranche, with a 50% reduction here costing $1 trillion.

Once again the government is faced with the dilemma: do the effective but expensive thing, or blow less capital on something that very well may not have an impact at all. And if the Reuters article is correct, the government will not disappoint in its continued course of consistently picking the latter option.