As usual you do have your point of view, which sometimes agrees with mine while other times it doesn't and I must usually end up correcting your statements or observations. However, I am very glad that we both agree on the importance of change in insurance programs for the HECM and its consequence on the legistative actions we are seeing today. My comment was mainly directed as to get some feedback from the playmakers and why they missed such an important part of the law. Lessons learned are always an important aspect of keeping an industry alive.

As for your questions and comments lets address them:

“Can you justify your claim that rolling back to $417,000 would have any impact on the anticipated losses?”

I had a feeling you were going to ask that question and I did not want to get into the mathematical background of the HECM PL calculation details. The answer is probably two fold:

1) the more simplistic answer is that if you roll back the max claim amount to $417k then of course, there will be a reduction of number of borrowers willing to get a HECM either because their mortgage balances may not be covered or because they feel they are not getting enough, etc. If the total number of projected loans is reduced, and with the current assumption of house valuation trends they are showing a net loss of (expected claims – expected premiums) then the overall subsidy required should be less, as you add up all the numbers for each loan age interval and multiply by the reduced projections.

2) As for the more regorous answer you must look and understand how the expected value of the losses are calculated at the loan level at given time in the PL calculation algorithm. To make thing easier I would just present a very high overview(see the references in the link for details)
The expected value of a loss L at a given time t (which could be years or months) is given by:

E(L(t)) = Prob(loss at time t, for a borrower who took a loan at age x) x {Balance(t)-Beta x E[H(t)]}

Without going into too much detail the probability of a loss is calculated as the product of the survival probability (with a correction for mobility) times the probability that the property value on a terminated loan is less than the outstanding balance.
Balance(t) is the outstanding balance at some time t
E(H(t)) is the unconditional expected value of the future home appreciation model (based on a lognormal distribution) at some time t.
Beta is a factor between (0,1) that when multiplied by the unconditional expected value provides the conditional expected value of the Home at time (t) with the condition that the balance exceeds the home value at that time.
As you can see with higher claim amounts the Balance(t) will be usually higher, especially for the fixed rate close account HECM (lump sum). This will most likely make the difference between the outstanding balance and the home value larger at later time intervals. Remember, with the PL algorithm, for a give PL factor, age of the borrower at origination and an expected interest rate, the value of the expected losses minus expected premiums is calculated for every year from origination to termination (assumed to be age 100). The total is added and you converge to a PL when you have a break even point (ie present value of losses minus premiums is about 0).
Keep in mind the expected value of the home is computed assuming a home appreciation rate. As they use more pesimistic values this difference also increases.
Also notice that I am not even considering the effect of higher balances on the probability of the balance exceeding the home value. That would be another effect and certainly reducing the MCA will provide more room to operate or reduced the likelood of that happening, especially for those homeowners with house values exceeding the MCA limits which is more likely with lower MCA limits.

As we say in mathematics Q.E.D or in latin (as it was to be proven). This is why reducing the max claim amount would reduce the risk to the program.

As for your comment about the hybrid model I proposed, you referred to a “HECM lite” term. FHA doesnt have, as far as I know a HECM lite program.If you care to provide some links or reference or specs on such program please do so. Inquiring minds want to know.

As for your Hybrid idea, I would think your problem will be with the secondary market implementation. Keeping two rates and two types of accounts (open ended and close lended)at the loan level in a pool in a HMBS will be tricky. There is a reason why GNMA has been successfull with their HMBS because it is a closed ended account and lump sum so there is no money or fund to be set aside for future loan disbursements. That is key for secondary market issuers and for servicers. There would also be the regulations applying to close ended and open ended accounts. Does MDIA sound familiar? Mixing them both, is like mixing a HELOC and forward mortgage in one single loan. Dont know if it is feasible from a legal point of view. With the HECM lite program as you call it, there is no difference in implementation at that level. There would still be some implementation issues at the lender level.

As for the lowering of the MIP upfront, I still need to see an algorithm to arrive at such conclusions. The study by NRMLA, would you care to share? If it was done before the downturn then it needs to be redone with current assumptions. Increasing the periodic MIP rate and lowering the MIP upfront could increase the long term risk of the program if the same algorithm is used.However, studying some happy medium may be a good exercise.

I am all for lowering the cost of HECMS and for higher PLs as most folks do. The one thing to remember is that there would not be a reverse mortgage industry without the government insurance fund. As much as we may not want to admit, MI is a necessary condition for the industry. Without the Mortgage Insurance program there would not be RM loans originated with such PL profiles (remember the proprietary loans and what PL they provided??).

As for your comment as how they arrive at a 5%, I dont know. My point was about how they should have arrive at the number.

Lets hope that somehow NRMLA and AARP can stem the negative tide in Congress and keep the program alive and yet useful to our seniors.

Please see my response below.

As usual you are right on point on the most crucial issues (but not on others).

As to the most significant issue, you have hit the nail on its head. It is the shift (or recategorization) of the HECM program within the HUD insurance funds that is the primary reason for the need for any subsidy at this time. For all of the positives it was alleged to be, HERA created several negative outcomes for HECMs, one of which was the lowering of origination fees and another, the recategorization of HECMs.

The insurance fund category change can be found at Section 2118(b)(2) of the Housing and Economic Recovery Act of 2008 (“HERA”, P.L. 110-289) and the actual change itself is now codified as 12 U.S.C. 1715z-20(i)(2)(A). As you correctly stated the Fund change was from General Insurance (“GI”) to Mutual Mortgage Insurance (“MMI”).

When the category changed so did the requirements on how the HECM budget amount is computed. CBO is now responsible for that calculation. The category change also means that the program must be in balance fiscal year by fiscal year. It also means that CBO (no longer HUD) for budget calculation purposes determines: 1) the estimated home appreciation rates, 2) discount interest rates used in calculating present values of future cash flows, and 3) future interest rates for interest accrual purposes.

As to the budget request for the fiscal year ending September 30, 2010 (“FY 2010”), only the HECMs endorsed within that fiscal year are considered. All projected MIP revenues and all projected losses from the HECMs expected to be endorsed in FY 2010 are reflected. This means if some MIP revenues from the HECMs endorsed during the FY 2010 are projected as being generated on January 1, 2040, that MIP must be discounted back to October 1, 2009 and reflected in the HECM budget amount for FY 2010. No MIP revenues and no losses from HECMs endorsed in any fiscal year other than FY 2010 are reflected in the $798,000,000 HECM shortfall FY 2010. Alleged surpluses from prior fiscal years cannot offset it.

Can you justify your claim that rolling back to $417,000 would have any impact on the anticipated losses? No “accounting” can yield that information. Accounting is historical in nature and I am reasonably sure that very, very few HECMs that have had MCAs (Maximum Claim Amounts) over $417,000 have terminated as of August 13, 2009. I think what you are looking for is a breakdown on the budget projections.

As a real estate broker, it is clear that percentage losses to values are significantly different in different value ranges in different localities in the country. Does the CBO calculation take that into consideration? (A rhetorical question) Some argue that if the existing lending limit was at $417,000 the projected losses would have been much disproportionately lower because there would be less potential loss on homes with values over $417,000 to $625,500 due to lower principal limits and upfront costs. But here again, consumer borrowing habits in the various value ranges by region of the country would also become a significant factor which I doubt is reflected in the budget shortfall.

I know you want to believe that there was a very serious study leading to the conclusion that 5% is the right percentage to reduce the principal limit under the Senate Amendment and you could be right. But knowing something about legislation on the tax side, that is probably more of a wish than a reality. Political compromise is a major part of any budget process. It is not for lack of financial math reasoning or skills that my questions arises but more because of not knowing how much political compromise, if any, is reflected in that percentage. For example, I wonder if how they settled on $288,000,000 was simply by dropping the 3 on the bill number and multiplying the result by $1,000,000. Sillier basis for compromise have been agreed to before.

From a study that NRMLA commissioned reflecting historical data, if the periodic 0.5% rate is doubled, a significant part of the upfront MIP can be lowered without any significant loss to overall MIP revenues. However, it seems this will result in a shift of costs from lower risk pools to higher risk pools of HECMs, increasing the risk to HUD. The study was commissioned and completed before the current economic downturn.

Your idea of having another product with lower principal limits and thus lower MIP costs to the borrower actually has a nickname in the industry already called “HECM Light.” However, I personally favor a hybrid HECM with all moneys taken at funding being close end and having a fixed rate applied to that portion of the loan while all proceeds not taken at funding being accessible through a “growing” line of credit (open end) and proceeds taken from that pool being subject to adjustable rates identical to the annual adjustable rate HECM. Of course other options of taking payouts should be available to borrowers through the adjustable rate portion of the HECM.

Also MIP can be lowered upfront and the periodic rate held at or near its current rate, if the program charged MIP on proceeds as they were distributed. The rate would probably have to be around 6%. If there is a line of credit and the balance due is paid down, future borrowings up to the total amount of previous net pay downs would not be subject to the 6% MIP charge. Such a program would encourage those with higher equities to get a HECM since the upfront cost barrier would be greatly reduced and MIP revenues from fairly low risk HECMs would be a source of additional income to the program without taking on significantly more risk.

As an advocate of independent actuarial studies for well over five years now, this is a provision whose time is past due. For some time we have had a different view of the HECM financial (or math) model. Our views came closer together a few months back when you expressed your agreement on the need for periodic testing of the validity of assumptions underpinning the model. Despite the independent study, HUD internal testing of underlying assumptions seems very warranted.

While you blame much of the change in categories on the political ineptitude of AARP and NRMLA, it seems there is much more to the change and the required study. Some of us have been concerned about the inability of the HECM financial model to adjust adequately to significant increases in annual endorsements, shifts in borrower behavior in economic downturns, and a significant downturn in home prices for an extended period of time among other significant assumptions. It seems that if the existing HUD financial model had been adjusted to reflect some of these concerns along with reflecting a portion of anticipated surpluses from prior fiscal years and adjust MIP and principal limit factors accordingly, this terrible recategorization we are experiencing might have been avoided.

Congress's solution to lower the principal limit on reverse mortgages could not come at a worse time for senior homeowners. Once seniors are made aware that they may not be able to get the help they need, (and when they see a 5 to 10% decrease in available proceeds, they will be aware) could they answer back with their vote?

I am not against the reverse mortgage program being on sound financial footing. I do not think that the timing of this particular budgetary action is good for the senior homeowning population. If a senior asks me why a reverse mortgage cannot provide enough money for them to stay in their home and their difference is within the budgetary decrease percentage, I know what my answer will be.

Congress's solution to lower the principal limit on reverse mortgages could not come at a worse time for senior homeowners. Once seniors are made aware that they may not be able to get the help they need, (and when they see a 5 to 10% decrease in available proceeds, they will be aware) could they answer back with their vote?

I am not against the reverse mortgage program being on sound financial footing. I do not think that the timing of this particular budgetary action is good for the senior homeowning population. If a senior asks me why a reverse mortgage cannot provide enough money for them to stay in their home and their difference is within the budgetary decrease percentage, I know what my answer will be.

Congress's solution to lower the principal limit on reverse mortgages could not come at a worse time for senior homeowners. Once seniors are made aware that they may not be able to get the help they need, (and when they see a 5 to 10% decrease in available proceeds, they will be aware) could they answer back with their vote?

I am not against the reverse mortgage program being on sound financial footing. I do not think that the timing of this particular budgetary action is good for the senior homeowning population. If a senior asks me why a reverse mortgage cannot provide enough money for them to stay in their home and their difference is within the budgetary decrease percentage, I know what my answer will be.

First is the question of how they determined the subsidy rate. My guess here is that HUD knows the distribution by age of HECM loans and has projections for the total number of loans for the next few years. They probably ran the PL calculation analysis for each age interval (62-99) using a modified housing appreciation rate (to reflect curent conditions) and look at the deficit for the next year or so for each age interval. Then if you have a distribution of loans by age intervals and the projected number of loans for the next few years you can calculate the defict for each age interval and then add them up and determine the difference of expected present value of losses minus expected present value of premiums. That is probably how they arrived at the 780 number for 2010.

Now if they did things right, then all you have to do is for each year interval change or decrease the PL factor (using the modified housing appreciation rate) by a certain amount and recalculate the expected loss using the age distribution and the projected number of loan as above. You do this until you end up with a reduction that eliminates the shortfall (for 2010 or perhaps next few years??). Perhaps they probably padded the correction so that they may not have to go back and changed in the next few years. A good question here is, if they accounted for a possible reduction in the projected number of loans that will be made due to the PL reduction and also the impact on the distribution by age, that is, is more likely that there would a shift on the age of the borrowers getting HECMs due to the new rule? Hmm… I dont think so but they should have accounted for this in the analysis.