Deflating the Housing Balloon: Ka-Pow! or Pssssssss.
July 13, 2005--1st in a series.
Cumberland Advisors:
cumber.com
"When does interest rate risk become credit risk?" With that title Fannie Mae economist, David Berson, set forth the explanation of how an adjustable rate mortgage (ARM) structure can morph into increasing systemic delinquencies as interest rates rise. Readers may wish to print a copy of his essay cumber.com . I will be referring to the chart at the bottom of it during this commentary.
Berson's essay cites the excellent work of his colleague, Anton Haidorfer, We thank Anton for the help he has given us and the explanation of the data that we are using. The conclusions are mine as are any of the interpretative errors. The credit for the seminal research work and assistance provided to this writer rightfully go to Anton Haidorfer and David Berson.
There are about 45 million first lien single unit mortgages in the United States. They total about $6.6 trillion if you exclude the government guaranteed mortgages like those made to veterans. The majority are fixed rate; the rest are variable. We will jump right to the variable rate ones, the ARMs.
Of those ARMs about $1 trillion will reset their interest rate within the next 18 months. That is about 7 million households. See the chart at the bottom of Berson's essay.
Let's break down the categories in that chart. The prime conventional conforming loans (PCC) are gradually resetting and have been placed after comprehensive credit scrutiny of the borrower and appropriate down payments and appraisal of the property. As the chart shows, the phase in of a rate reset is slow and steady over a longer period.
The Alt-A category has a faster rate of reset. Alt-A mortgages are loans of presumed "A" credit quality but whose users have traded a higher mortgage interest rate or higher fees in return for a faster and less cumbersome approval process. They are not expected to become a big problem; however, some analysts worry that during the recent bubble period some of these Alt-As have morphed into "Alt-B" credit quality.
Jumbos are mortgages above the conforming loan limits. Conforming loans are the ones that can be securitized by Fannie and Freddie. That limit is currently $359,650 in the lower 48 states and $539,475 in Alaska and Hawaii. Jumbos are resetting even faster than Alt-As. Those $450 billion of borrowers are deemed to be wealthy. They won't like the increase they will pay but they probably will not be defaulting because of it. Rich folks don't often lose their houses because of a change in interest rates. One warning is in order: in high cost areas like California, New England or Washington D.C., many of the non-rich have been using jumbos because the house prices are so high. If we are in an a bubble and these folks encounter financial pressure, the problem loan characteristics in the jumbo category will reflect it in default and delinquency statistics.
The sub prime category concerns us. It is resetting rapidly. We estimate that the average reset adjustment will be an increase of 2 percentage points. In addition there will be the prospect of a future adjustment confronting the borrower in the next few years. These are the mortgages which were placed at a time when the Federal Reserve was maintaining a 1% policy-setting federal funds rate. Many of these mortgages were initiated at exceptionally low "teaser" rates. Furthermore, the loan-to-value ratios on those mortgaged properties are much higher than the conventional loans. And there was minimal or no documentation of income.
Is there a problem? If yes, how big is it?
We think the answer to the first question is yes. These mortgages average $150,000. They tend to be concentrated in urban areas and in the lower property value strata within them. They have a high number of first time home buyers. We estimate that the average annual income of that household in this group is about $50,000. That would be a consistent income needed to pass an affordability test for a $150,000 mortgage placed at a very low interest rate with minimal or no amortization of principal.
If we take our assumed average and raise the rate by 2 percentage points, we have just added $3000 a year to the mortgage payment in that household. $3000 to a $50,000 income is huge. The economic recovery underway in the last two or three years has not raised that household's income fast enough to overcome its higher household expenses and still absorb the shock inflicted by this rising larger payment. In addition, the homeowner will quickly be able to figure out that a future and second increase is coming. They know that higher rates will hit them more than once.
Okay. How big is the problem?
There is a total of $941 billion outstanding in sub prime fixed rate mortgages and ARMs. Of those, $543 billion sub prime ARMs will reset by yearend 2006. The rate resets are underway right now and will accelerate during the next year as the chart shows. That means nearly 4 million households are, or soon will be, feeling this impact. We think that is big.
Those households will face pressure and either cut back spending or make a default election. No one knows how much of either will occur. But remember we are dealing with a group of many first time home owners. They only qualified because of the very low interest rates that enabled them to buy in the first place. And they have very little equity in their house. Do they elect default? Do they allow the property to deteriorate because they cannot afford to do the maintenance? Are some financial institutions going to "get the keys back" and find themselves with a large number of workouts? Or, do these folks cut back on other spending and trigger a slowdown in the consumer sector of the economy. In our view the likely outcome as the bubble deflates is some combination of all of the above.
Two conclusions:
1. This is one aspect of the housing bubble story that has not yet made the front pages. So far, the media have only examined the rising prices in certain sectors and the amount of speculation underway in housing. We believe the ARM rate increase and default trigger is next year's front page.
2. Only time will reveal the degree of defaults originating from this ARM construction and rising interest rates. See notes below re: delinquency vs. default. However, the severity of the rate rise is estimable now. Fed Funds were 1% and teaser rates were used in abundance. Fed Funds are now 3 1/4% with an almost certain hike to 3 1/2% in August. Many forecasts have Fed Funds around 4% at yearend (Cumberland is among them.). This means the total $1 trillion in ARMs will experience a $20 billion increase in interest cost in 2005-6 and probably a like amount in the next reset cycle. Remember, this is a recurring and often the annual adjustment; the present value of that adjustment for long term planning purposes is enormous. We need to view this as a transfer of the spending power from the households to the lenders who hold the mortgages.
Cumberland continues its strategy of greatly under weighting the financial, home building and mortgage sectors in our stock and bond portfolios. We believe that risk is very high. We cannot project a calamity nor can we be sanguine about a gradual and painless adjustment. High risk says be careful. That sums up our strategy in the summer of 2005.
David R. Kotok
Some additional notes. In the Journal of Fixed Income (June 2005, pgs 28-39) SEC Financial Economist Michelle Danis and St. Louis Fed Senior Economist Anthony Pennington-Cross collaborated and examined the sub prime category in great detail and with significant statistical modeling. They sought the trade off between default and workouts which were a result of a sale of the property that led to a payoff of the mortgage. In the sub prime category they note that "the current equity status of the property is a key determinant" because " from a borrower's perspective, a positive equity position makes the borrower more likely to attempt to preserve" equity "by selling rather than letting the property go into foreclosure. From a lender's perspective, the opposite is true in the case of a property with positive equity."
They found that the "there is no reason to assume the relationship between delinquency and default is linear" and that house price "volatility" is important. Rising prices tend to bail out the borrower and the lender. Falling prices from a bubble collapse would work in the opposite way. They also note that sub prime "lending tends to be concentrated in low income and minority areas and in areas with troubled economic conditions." Also noted is that sub prime borrowers have "poor credit characteristics" and are "less knowledgeable about the mortgage process" and are "less satisfied with their mortgages."
In Cumberland's view this excellent technical study gives support to our conclusion that the risk is high and that the housing bubble may end badly. Readers who wish to read the study can email me with their address and I will snail mail a copy to them. |
