Monday, January 28, 2008
WHE The Enemy
Anyway, the WHE of whom I speak (actually a would-be WHE who shall remain nameless out of Christian charity) was claiming that home prices wouldn't fall because interest rates were falling. I laughed pretty hard for a while, but then I sobered up and wondered if a new generation of bank presidents wasn't at risk. The reality is that when bank presidents go bad, your pension funds have a tendency to sour likewise.
So let me be the Big Bird for a new generation of Bagholders and Bank Presidents, and possibly even would-be homebuyers.
So let's start with an economic axiom:
Price(P) is a function of the relationship between Supply(S) and Demand(D). Generally, if Supply is rising at the same rate as Demand, Price will remain stable in real terms; if Supply is rising faster than Demand, Price will fall, and if Demand is rising faster than Supply, Price will rise.
And continue with another economic axiom:
Demand is dependent on Price to some extent under almost all circumstances. I'm not going to buy a Twinky for $20, unless I haven't eaten for two or maybe three days AND I can't find anything else to eat AND I have $20. But then, I hate Twinkies. However, if I were hungry and there was nothing but Twinkies to be had, but they were offered at $1, I probably would buy one after 10 or 12 hours.
Supply will often be dependent on Price and Demand. When Demand rises, it will generally trigger a rise in Price, which will create a strong incentive to increase Supply.
As a result of those axioms, it can be seen that economies don't work as straight line equations of the sort that WHE's(TM) propose. For that matter, no pricing relationships do. That is why CR spends so much time and detail analyzing various aspects of the situation, and why his forecasts are so good. You actually solve for any of these variables by iterating against current economic circumstances with expected movements of the underlying constants that affect the three variables for the commodity in question. This is not precise, but it is a whole lot more authoritative than the hilarious forecasts of WHEs.
Taking residential real estate prices in particular, they are somewhat complex. During most periods of our history, there has been an entry cost (downpayment, closing costs, and/or other reserves). Entry cost acted as a gatekeeper. If you could get through the gate, then the question was whether you could afford a P&I monthly payment, plus monthly carrying costs (taxes, property insurance, HOA, etc).
Therefore effective prices for homes are usually combination of a step function (how much of the population can come up with the entry cost at any given time) and a relatively smooth function of monthly P&I plus carrying costs. For an amortizing loan, the monthly P&I varies based on loan principal and interest. Property taxes vary somewhat based on price, but also according to local factors. Property insurance varies mostly on experience, and it so happened that we were in a period that had had an abnormally low rate of disasters on the coasts.
What happened in 2002 was that several trends coincided. First, interest rates were drastically lowered. Second, the trend for some years had been toward lowering the entry costs in relationship to price and indeed those had dropped more than 50% on average. To some extent, the entry cost barrier had been shifted to FICO score.
The combination of lower entry costs with low rates was that demand was sharply boosted in the short term, which created an environment in which price rose. Because prices were rising, the relative risk of lowering entry costs genuinely did reduce for a time, and entry costs basically dropped to an effective zero for at least half the population of would be buyers. Indeed, the previous financial entry cost was substantially replaced with a FICO gatekeeping function.
However, entry costs were related to price, whereas FICO had very littler relation to price. That removed a price-related drag on demand and tended to boost prices. Naturally this increased Supply. This cycle iterated a few times, with the more conservative lenders steadily losing market share. Eventually, they responded by loosening their credit terms, which caused those operating on the higher risk margins to reduce their credit underwriting standards again.
By 2005, virtually all potential demand had been consumed, and by the end of the cycle (2007) the only way to sustain demand was to steadily knock factors out of the Price equation given above, which was
- Entry Cost (Boolean Yes/NO) and
- P&I plus carrying costs.
- FICO (Boolean, but steadily decreasing so that at the end it hardly acted as a gatekeeper at all), and
- No principal payment, about 3.5% initial interest on principal, and no verification of ability to pay long term carrying costs plus even the long-term interest rate on the loan. These are known as funny-money loans, because they are granted on the theory that the Borrower will default to someone else, which will be funny.
- There was further awesome iteration at some institutions, at which they connived at faking appraisals for loans for their own portfolio, wrote loans which required less than the accrued interest on principal to be paid for a period as long as three to five years, and paid themselves big bonuses. These are Institutional Suicide Loans, which are not funny at all and cause a pronounced rise in early retirements among bank executives and the like, as well as vastly increased payments from financial companies to politicians.
There were good reasons for the earlier version of pricing: A downpayment has been traditionally required because it spreads the cost of default to both buyer and lender. This limits walkaways. Demanding payment of the true cost of interest on the loan amount supports cash flow in the absence of a Stupider Lender(TM), and it is foul and unpleasant to find yourself paying the property taxes and insurance on the collateral for your loan in order to protect yourself from total loss, even while your expected net loss on sale is rising. This causes great grief in the hearts of Bagholders and an abiding aversion to funny-money loans.
So in view of the fact that we are rapidly restoring the Entry Cost equation of 5-10% of purchase price, plus verification of ability to repay at least carrying costs plus the interest on the mortgage amount, a change in interest rates has very little effect on the Price equation. Real mortgage interest rates are not going to fall below the 3.5% average teaser for entry-level. They just won't. Therefore, the Demand portion of the equation has been constricted suddenly and will continue to constrict for some time. Since Supply must rise, and Demand must fall, Price must also fall.
The supply of declining markets is expanding.
The way I look at it is that we'll end up with standards only a little looser than the point at which you and I would lend money. We'd be a lot tighter than it is today, wouldn't we?
Likewise if the average wage cannot support the average house price in a given area, there will be limited sales.
In the bigger picture, if the income of the average consumer can only support a given level of consumption, then that will be the level consumed. This is the big one as far as I am concerned. Assuming that consumption was supported by credit with easy entry; credit cards, home equity loans and such then the level of consumption can only fall and that sets into a vicious cycle as incomes fall consumption falls; jobs related to consumption go away, income from those jobs goes away and consumption falls further.
Mix in a need to repay current loans without new funds coming in from new loans and things look bleak.
If you could comment: I see an obvious problem with home loans adequately pricing risk. Fannie and Freddie offer low rates, presumably due to the low cost of capital...but somewhere in that price of the loan over other sources of money is the cost of default (spread over all the loans) much like insurance fees are set.
Given that we are in an environment, as you point out, of dropping prices and increased default activity, making defaults much more common and expensive, both on an individual and cumulative basis, (add aslowing economy for good measure), can Fannie and Freddie lend at these artificially low rates for very long?
It seems, as you other posters point out, that to lend in this environment, rates will have immediately approach what you and I would charge (double digits?) to lend.
You are genius by the way.
The traditional conforming loan limits have spread GSE risk much more evenly over the country, and have created a situation in which, for example, Fannie's loss severity is much lower than for this higher-priced echelon of loans.
There are just way, way more couples who earn 60K and can reasonably vacuum up a house for $180,000 than those who can ante up for these $600,000 deals. Okay, so Fannie might be selling homes for $180,000 that originally sold for $200,000, and taking an average loss of $15,000. That's a heck of a way from having to count your losses like you would in these jumbo neighborhoods. Lower priced houses in flyover country just won't fall as far unless the local area is in a depression like Michigan, because there is a much greater pool of people who can afford to buy them.
This is why I am literally shaken and appalled by the deal to raise conforming loan limits. That's like a doctor coming along and telling me "We're going to call you healthy". Whatever my state of health is, it is. You can call it whatever you want - it's not going to change the outcome. Similarly, calling a jumbo "conforming" won't make it act like one.
Just so I'm clear, you point out the difference in loss severity in the mostly conforming flyover country (example of a $15,000 loss on a $180,000 home). I understand that losses are smaller on smaller dollar loans.
However, in the current unusual house market enviroment (nationwide flat or falling prices) you have several things increasing losses.
In your example: In the past, when a $180,000 loan defaults in a market of (even normal 1-3%) rising prices several things are true:
1) A default usually can be avoided by the person who needs to move, gets sick, divorced, etc, either through a sale or refinance. So overall defaults are fewer.
2) If a default does occur, the losses from the time it takes to reposses and then resell the property are partly (albeit barely) offset by rising asset prices. As opposed to the losses accelerated by a deteriorating market over time.
3) The default in a normal market is likely to get immediate attention by the REO department and thus the damage from time/neglect etc are mitigated. In this current market, where time is the enemy, the high numbers of defaults increases both loss through deterioration as well as price drops in the area.
So, not only do you have the loss per individual house greater in a falling house price market, but you also have many more homes defaulting...compounding the overall losses for a lender.
So, if the above is true, how do they price the risk of even conforming loans at the same rate as they were only a few years ago? They (FNM, Freddi) seem to respond in the rate they charge totally to the price of money and don't price in the added risk of this environment.
Its like pricing the risk of life insurance for a group of nuns living in a convent, and then shipping them all to the center of a war zone where casualties mount up. They are the same people, and the payouts are the same per casualty....just the casualties are greater in number and thus previous insurance premiums are totally out of whack.
Am I way off here?
Basically the low, low interest rates you hear of are for people who have the 20% down, whether in equity or in cash. Some lenders are now adding premiums for seconds, even if their first is only 80%. So in that case, investors believe there is relatively little risk.
If you compare MBA's weekly release (20% down) with HSH Trends you'll see the difference.
Note that points are coming back for the good rates.
There is another aspect to the jumbos. When demand for homes overall drops, the jumbo group is initially a narrower segment of demand and overall demand in that bracket drops more (some would-be jumbo buyers drop down to a lower bracket), so these loans get hit worse than lower pricing levels. There's more residual demand for those lower-priced homes overall.
My belief is that even 20% down won't do it unless you also drop back to more reasonable DTIs. The people who had major deposits down and will walk are those who have lost their stake and can't really survive in the home for years without another infusion of capital from equity. If they put major money down, they were serious and not speculating on the house in the first place. If they can, they will stay.
FHA loan defaults had gotten up to 12.58%, and they are careful on appraisals and documentation. But honestly, a 41% DTI is too high for most conforming.
From my point of view, it now comes down mostly to area risk. I expect low LTV high DTI prime mortgages in busting bubble areas to do worse than traditional subprime in areas that weren't heavily "creatively" financed. The correlation between excessive price/income ratios and creative financing is very strong.
Traditionally, conforming mortgage defaults are strongly correlated with adverse life events (borrower-specific) or economic downturns which occur before the borrower has built signficant equity and therefore margin.
So in terms of a borrower such as Tom's man, he is now moved to the traditional risk category of someone who put 5% down. Now it is just a matter of what happens to him. If he loses a job, gets ill, has a family emergency, gets in a serious car accident - he's got no margin in the home. Nothing to work with. And he won't probably for at least 5 years.
So the question there is can the man save and pay his mortgage? He needs a nest egg.
Unfortunately, when you get high DTI people in depreciating homes, you have a really nasty situation. Over 5 years they will have high defaults regardless, cause stuff is going to happen to a certain percent of the population.
Severe adverse life event is usually around 1% a year. In the first 7-10 years of the mortgage, that will most often cause a default. Over 5 years, about 3-4% of the population will have a relatively severe car accident causing injury, loss of income, unexpected medical bills. That will cause temporary default for borrowers who don't have any reserves.
And so it goes. The reason why Fannie can charge such comparatively low rates is that it has a huge and relatively even distribution, and economies vary by locality - so that when MI is down, TX may be booming! This means they get a much more even overall flow of defaults.
Let me know if this does not make sense. I am extremely tired and could do better tomorrow!
Also, if you're going to start off with economic axioms, I recommend focusing more on the Greater Fool effect. I may be a fool to buy a tulip bulb for fifty dollars, but I'm a shrewd fool if I have good reason to believe I can turn around and sell it to a greater fool for a hundred. But to the person selling it for fifty, I was the reason they were such a fool as to buy it for twenty-five. In a whole marketplace full of people reasonably fools, if the average guess about others' folly is even a little off, everyone will be (so to speak) more foolish than average and the price will go screaming off to infinity.
Mentioning funny-money loans with nothing more said is perfectly ok, of course, but why miss the opportunity to mention the most directly relevant axiom of the bunch?
When downpayments were eliminated, everyone who might have been saving for one bought in. So now we have the odd situation of a couple of dead years in which most would-be FTHB cannot qualify for mortgages.
This is a historical rarity. The last time it happened was in the 1920s. Since then we have never had a time when eligibility standards for mortgages loosened so greatly and then snapped back like this.
There have been regional bubbles, but not this 3/4 of the country type of thing.
For refinancing, the terms are tightening in various areas, but that doesn't change the home pricing situation at all. Home prices are really dependent on first time home buyers.
Same for the TLC program directors who keep running those Flip That House marathons day after day.
Links to this post: