Case Study Six—How to Fudge a Little
A broker orders an appraisal indicating the target property value is $325,000. The borrower wants to refinance and get as much cash as possible. The appraiser’s initial look-up shows comparable sales (comps) in the area range from $278,000 to $362,000. At first glance, the target value appears reasonable.
An appraiser should use comps that best resemble the subject property in age, size, style, and location. While appraisers should use a minimum of three comps, they can have more depending on what has sold in the area over the last six months. If the selection is thin, the appraiser will search for the nearest comps that most closely resemble the subject. The appraiser modifies the value for each comp relative to the subject property by adjusting up or down for major factors—square footage, number of bedrooms and bathrooms, swimming pool, and so on. This helps the appraiser determine a fair market value for the property.
In this example, we’ll assume the property values have already been adjusted for these factors. For this loan, the appraiser finds four comps with the following values:
$290,000
$299,000
$308,000
$311,000
At first glance, it doesn’t appear the target value of $325,000 is supported using these comps. So the appraiser digs a little deeper and finds two additional comps a mile away with values of $335,000 and $343,000. He also realizes that just adding them into the mix won’t support the target value unless he removes the two lowest comps. After determining they weren’t as representative of the subject as he originally thought (too small, too old, wrong style), he feels justified in removing them. These are the four comps he uses for the appraisal:
$308,000
$311,000
$335,000
$343,000
This produces a final value of $325,000. Having followed his professional standards, he believes it accurately represents the property’s fair market value. This process shows that stretching an appraised value by 5 to 10 percent is not difficult. However, the difference between creativity and manipulation is a fine line that’s open to interpretation.
Knowing that property values are often stretched, lenders develop a routine to validate the appraiser’s work. Most start by checking county tax records and using one or more automated valuation models (AVMs). These programs provide basic information on properties in the area that sold, including square footage, date of sale, and distance from subject property.
The AVM shows 18 sales in the area over the last six months. However, only three properties sold above the $325,000 appraised value, two of which were used as comps. This is a telltale sign the appraiser is stretching the value. Sometimes looking at the property photos can provide more information. Not surprisingly, the two higher-priced comps appear nicer than the subject property. In fact, the subject property more closely resembles the other two comps ($308,000 and $311,000). If the lender’s assumptions are correct, the property is overvalued by $15,000, making it worth closer to $310,000. It’s clear the appraiser is using some of his fudge factor to reach the target value.
At this stage, the lender either accepts the value or does more research. If they believe the deviation is small, they might order a desk review. This means another appraiser reviews the original work without going to the property. As the name implies, he will conduct the review from a desk, only using whatever AVMs are at his disposal. If the lender is uncomfortable with the original appraisal, they’ll order a field review. This is more extensive and requires the new appraiser to visit the property, examine the original appraiser’s work, and provide additional comps if necessary.
In this case, the lender accepts the value without any additional reviews. The borrower refinances 100 percent of the appraised value, which means he’s probably borrowing $15,000 more than the property is worth. The lender and the investor sign off on the loan, believing the variance is reasonable. Although the industry had no published standard, five years of funding and selling subprime mortgages told us at Kellner that 10 percent was an acceptable deviation for our four investors: Countrywide, RFC, Household, and Citi. This means that if the investor’s underwriter thought the property was worth $310,000, they’d accept an appraised value up to $341,000. Anything above this threshold ran the risk of being declined. Ten percent may not sound like much, but as you’ll read in a moment, this variance was a major reason property values increased as much as they did.
How to Fudge a Lot
Let’s assume the scenario is different. In this case, the borrower has more debt to pay, which requires the broker to give the appraiser a higher target value. Since the borrower is trying to pay two collections, a federal tax lien, and get $25,000 in cash, the broker determines a $359,000 value is required to pay all the bills and pay his two-point commission. This figure is $34,000 higher than the original targeted value of $325,000. So instead of using the four comps, the appraiser keeps the highest one and discards the others. To increase the appraised value, he goes north three miles and finds two new comps, one for $359,000 and another for $363,000. He adds a fourth comp for $372,000, located three miles northwest. These are the four comps:
$343,000
$359,000
$363,000
$372,000
The appraiser submits a $359,000 property value to the lender. Compared to $310,000, which was the lender’s original estimated value, it’s a 15 percent deviation.
One look at the area map reveals the appraiser’s tactics. The subject property is located in a midsize city, just south of downtown. Getting from it to any of the new comps requires driving across a major highway and two sets of railroad tracks to a nicer part of town. The appraiser found more affluent neighborhoods and used these comps to stretch the value. The property photos confirm that the new comps are much nicer than the subject. Since the lender recognizes the appraiser’s strategy, he’ll order a field review. Not surprisingly, the reviewer cuts the value by $40,000, providing the lender with four new comps to support a value of $319,000.
Even a loan with a deviation this large could work, it just requires some creativity. While the lender believes the field reviewer’s assessment is accurate, this value won’t work because the collections and tax lien must be paid, and the broker has informed the lender that if the borrower doesn’t get some cash at closing, he won’t close the loan. So the lender counters with a value of $341,000, which is a full 10 percent deviation from the original value of $310,000. To get the borrower some more cash, the broker eliminates his two-point origination fee but increases the interest rate by 1.25 percent to make it back in a yield-spread premium.
The final value is not arbitrary. It’s derived by knowing exactly how far a value can be fudged and by working all the numbers down to the last penny. In the end, the borrower gets what he wanted, the lender and broker each earn a premium, and the investor received a performing loan.
There’s one problem with this example. The property must appreciate substantially for the borrower to break even when the time comes to sell. In an up market like 2001 to 2004, he might have gotten away with it. But if this loan closed in the last few years, especially in an inflated market like Las Vegas, Phoenix, or Miami, the borrower is in a world of hurt.
Understanding the Impact
Subprime lenders used to conduct a desk or field review for most broker-ordered appraisals because these appraisals were considered to be unreliable. For us, the number of unreliables reached as high as 80 percent. With the appraisal process highly susceptible to manipulation, lenders had to conduct business as though the broker and appraiser couldn’t be trusted. This meant that either the majority of appraisers were incompetent or they were influenced by brokers to increase the value. Brokers didn’t need to exert direct influence. Instead they picked another appraiser until someone consistently delivered the results they needed.
To put things in perspective, there’s nothing extreme about the example given in Case Study Six. During m
y company’s history, half of all the loans we underwrote were overvalued by as much as 10 percent. This meant one out of two appraisals was still within an acceptable tolerance for our end investors. Another quarter were overvalued by 11 to 20 percent. These loans were either declined or the value reduced to an acceptable level. The remaining quarter were so overvalued they defied all logic. Throwing a dart at a board while blindfolded would have produced more accurate results. The excessive inflation of property values largely explains why subprime loans experienced a high rate of decline.
The implications become evident from doing the math. If multiple properties in an area are overvalued by 10 percent, they become comps for future appraisals. The process then repeats itself. We saw it on several occasions. We’d close a loan in January and see the subject property show up as a comp in the same neighborhood six months later. Except this time, the new subject property, which was nearly identical in size and style to the home we financed in January, was being appraised for 10 percent more. Of course, demand is a key component to driving value, but the defective nature of the appraisal process served as an accelerant. In the end, the subprime industry’s willingness to consistently accept overvalued appraisals significantly contributed to the run-up in property values experienced throughout the country.
How is this possible when a home or any asset should be worth whatever the market will bear? The answer lies with the down payment. If similar homes in an area have sold for $350,000, and a seller gets a contract for $400,000, that’s a function of market demand. A home is worth whatever someone is willing to pay for it. The appraisal, however, should still show that the property is valued at only $350,000. Since there are no comparable sales to justify the higher sales price, lenders should base the loan amount on the $350,000 value, not the $400,000 purchase price. If a borrower wishes to buy the home at a premium price, he must bring an additional $50,000 in cash to closing, the difference between the purchase price and the appraised value.
If the process had worked correctly, a significant percentage of subprime borrowers would have been turned down due to a lack of funds. Inevitably, this would have forced sellers to drop their exorbitant asking prices to more reasonable levels. The rate of property appreciation experienced on a national basis over the last seven years was not only a function of market demand but was due, in part, to the subprime industry’s acceptance of overvalued appraisals, coupled with a high percentage of credit-challenged borrowers who financed with no money down.
Laying the Blame
So who is responsible for this fiasco? Everyone from the broker and the appraiser to the lender and the investor shares in the blame, even though no industry group is willing to take responsibility. Why should they? Doing so equates to self-incrimination.
For me, the issue of property valuation served as a major source of frustration. Sandwiched between brokers who had no liability, an appraisal industry that was inherently flawed, and investors who were willing to buy loans with inflated appraisals, there’s no question the entire process was defective. While every lender who closed a loan with an overvalued appraisal bears some responsibility, the issues we experienced led me to believe there was a much larger problem.
When Kellner first opened, we used several different appraisal evaluation companies to conduct our desk and field reviews. Since we didn’t order the initial appraisals, these reviews played an important part in helping us make our final lending decision. If a review appraiser confirmed our suspicion that a property was overvalued, we either cut the property value to an acceptable level or denied the loan. Conversely, if we suspected an appraisal was overvalued but the review appraiser confirmed it was correct, we took this into consideration when making our decision. After all, an appraisal is just an opinion. Just because we suspect a property is overvalued doesn’t mean we’re correct. If both the original appraiser and the review appraiser are in agreement, we’re going to rely on their collective judgment as licensed professionals to make the right call.
A year after we opened, we decided to order all appraisal reviews through Landsafe, a wholly owned subsidiary of Countrywide Credit Industries, because of the tie-in between their two companies. If a loan was sent to Countrywide for purchase and Landsafe performed the appraisal or the appraisal review, Countrywide agreed to accept whatever property value Landsafe determined was accurate. It was a brilliant business strategy. Since determining property value was so critical to the lending process, their commitment to stand behind Landsafe provided us with security. It didn’t matter if we thought an appraisal was overvalued by 20 percent—if Landsafe blessed it, Countrywide treated it as gospel. They never put it in writing, and sometimes it took a little arm-twisting when values were questionable, but they stood behind their commitment.
The problem started whenever Countrywide made Kellner repurchase a loan, usually due to fraud on the part of the broker. When an investor issued a repurchase request, either the lender paid cash for the note or both parties agreed to a settlement based on an estimated loss. This required taking into consideration the property value, loan amount, premium paid to the lender, as well as a host of other costs. Smaller lenders, like Kellner, preferred to create a settlement since it meant not having to use large sums of cash to buy back the loans.
Since the property value was supported by a Landsafe appraisal review, it seems reasonable that Countrywide would use the same value when calculating a settlement. Somehow logic never made it into the equation. Instead, Countrywide would order a broker price opinion (BPO)—an independent third-party appraisal to determine a new property value. In every instance, which amounted to more than a dozen repurchase requests over a four-year period, the BPO Countrywide received was substantially lower than Landsafe’s value, in some cases by as much as 20 to 25 percent.
Trying to reason with their repurchase department was pointless. Our conversations bordered on nonsensical as we listened to them justify why Landsafe’s opinion was not acceptable to them, even though they owned the company. The justification they used was that Landsafe’s value only applied to the front end of their business.
Tying the use of Landsafe’s appraisal services to Countrywide’s core business paid enormous dividends. They not only made money off the appraisal fee income, they bought more loans from lenders as a result of the tie-in. For Kellner, purchasing a Landsafe appraisal review was like buying an insurance policy that only paid when it served Countrywide’s financial interest. Once the loan became a liability, Landsafe went from being a member of the Countrywide family to being the redheaded stepchild.
By accepting the BPO instead of Landsafe’s opinion, Countrywide sent the message that Landsafe was wrong. If this happened once, twice, or even a few times, I’d chalk it up to an overzealous appraiser. But since the property value for every repurchase request was dramatically reduced, there are only three possible conclusions. First, every Landsafe appraiser was incompetent, which seems unlikely since all of them were licensed professionals. Second, Countrywide influenced the appraisers conducting the BPOs to produce overly conservative estimates in order to reduce their losses. This is possible but also unlikely since these appraisers worked independent of Countrywide. Third, Countrywide encouraged Landsafe appraisers to validate excessive property values.
To an outsider, the last option might not make any sense. Why would any lender, let alone the largest mortgage company in the United States, implement such a practice? Since all lenders have a financial interest in the loan’s performance, endorsing excessive property values is tantamount to playing with fire.
If you understand how the process worked, however, the logic behind the practice becomes clear. Even if a property was overvalued, there was no immediate concern to Countrywide as long as the loan paid on time. Once a loan became delinquent, Countrywide’s QC department tore it apart looking for any reason a lender should be required to repurchase the mortgage. Of course, just because a loan was delinquent didn’t mean it would become a repurc
hase, but many of them did. Once a repurchase request was triggered, the loss ultimately moved from Countrywide’s balance sheet to another lender.
If Landsafe endorsed overvalued appraisals on the front side, it drove more business, which created more profit. Disregarding Landsafe’s opinion on the backside enabled them to minimize some of the losses. The strategy worked well as long as the status quo didn’t change. When the meltdown that became the mortgage Armageddon of 2007 didn’t leave any subprime lenders to absorb the losses, Countrywide was left holding the bag.
The important question is just how much the issue of overvalued appraisals will impact Countrywide over the next few years. If my experience in any way resembles the standards used by Landsafe in Countrywide’s other divisions, the next 12 to 24 months will prove to be painful for America’s #1 home loan lender. It also means that their projected return to profitability in 2008 will be nothing short of fantasy.
CHAPTER 5
Wall Street and the Rating Agencies: Greed at Its Worst
Until now, we’ve focused on the front end of the business by discussing the people who made the loans and the tactics they used to qualify challenging borrowers. Let’s shift our attention to the back side of the industry and what happens to a mortgage once it has closed and funded. We’ll examine how mortgage securitization works and analyze the Wall Street investment firms that drove the process as well as the agencies that rated the securities. These participants, in my opinion, share the greatest blame for the current housing fiasco.
Confessions of a Subprime Lender Page 10