This piece follows the article "Uncertainty in Federal Intervention" in the September 1996 issue of The Region that analyzed the difficulty policy-makers face in determining if federally subsidized interventions achieve public policy goals. That article used Fannie Mae and Freddie Mac as a case study because the uncertainty about their effectiveness is pervasive and inherent in the way the federal government subsidizes them. This follow-up discusses some of the potential and possibly obscured costs of Fannie and Freddie's technological initiatives.
Owning a home is the American dream. Getting a mortgage, in contrast, often involves providing a mountain of documentation and devoting a considerable amount of time, all the while suffering anxiety about the outcome. Loan underwriters, who must determine whether a loan application meets the standards of the firm bearing the risk of the loan, face the laborious task of sorting through this unwieldy paper flow. At the same time, underwriters are under tremendous competitive pressure to keep costs low, volume high and defaults under control.
A business process critical to a huge number of financial transactions with acknowledged weaknesses provides an opportunity for firms who can cost-effectively improve it. A number of firms over the last several years have created automated underwriting (AU) systems to accomplish that goal.
The story of AU would simply be another example of technological progress were it not for the fact that two firms with special status and indirect subsidies from the federal government, the Federal National Mortgage Association (Fannie Mae) and the Federal Home Loan Mortgage Corp. (Freddie Mac), are also offering AU systems [see "A Brief Recap of Fannie and Freddie's Activities"]. There are two main concerns with Fannie and Freddie's AU initiatives. First, these new technologies in the hands of Fannie and Freddie could lead to less competition in the market for AU systems, which could increase the cost of home buying above what it would be otherwise. In this case, the unique advantages conferred to Fannie and Freddie by the federal government could actually make home buyers worse off.
Second, Fannie and Freddie's use of their AU systems could lead to less competition in the secondary mortgage market, which is the national market where residential mortgages are securitized, that is, pooled and sold as securities called mortgage-backed securities (MBS). (The payments made on the underlying pooled mortgages flow to the investors who own the MBS.) Because of their government-provided advantages, Fannie and Freddie control virtually 100 percent of the market for securitized mortgages that meet their legal and underwriting standards and are "conventional" (not insured by the federal government). Securitized mortgages that meet Fannie and Freddie's qualifications represent about 65 percent of the total secondary mortgage market. The other roughly 35 percent of the secondary mortgage market is characterized by more competition than Fannie and Freddie's area of control. The automated underwriting systems will now allow Fannie and Freddie to capture some of the mortgages that would have been sold into the more competitive parts of the secondary mortgage market, or that lenders would have not sold at all.
Furthermore, the increase in Fannie and Freddie's secondary market activity will enlarge their indirect subsidization. Yet, as was described in the previous Region article, the public can be more confident that homeowners benefit from federal subsidies if they are provided directly to targeted populations instead of circuitously and implicitly through Fannie and Freddie.
More generally, this article provides another example of why policy-makers must carefully examine all the implications of federal intervention. Surely, policy-makers would not want to indirectly subsidize an activity that could actually lead to higher prices and worse use of our resources. Yet, that could be the outcome in this case.
AU: The generic benefits
There are several reasons why a variety of firms have developed AU systems for their own use or for sale:
First, these systems allow mortgage providers, using a portable computer, to take a borrower's application and provide an initial review in a matter of minutes instead of days. Final loan approval also comes much faster.
Second, some of these systems rely on new types of underwriting procedures that require less documentation, speed up the process and make better loan decisions. Using statistical procedures, underwriters have determined which borrower characteristics (e.g., a borrower's history of repaying loans) and features of the loan (e.g., the amount of the loan relative to the value of the property being purchased) are the most important for determining the probability that the loan will be repaid. The focus on fewer variables, some of which are based on credit report data, allows the underwriter to work with fewer supporting documents from the borrower. In addition, a computer using a statistical model can review a loan application much faster than a human. Finally, focusing on a limited number of highly predictive variables increases the lenders' ability to accept loans that will pay off and reject loans that will go into default.
Third, the automated systems allow lenders to make better use of their staff and lower their costs. The automated systems can process the applications that will clearly be accepted or rejected, thus allowing the human underwriters to spend more of their time on the applications that fall into the borderline area. The systems also assemble the information into a near-complete electronic package that the underwriter previously gathered manually. This process eliminates time-consuming tasks, such as manually keying the same information several times. Speeding the process and removing "choke points" allows firms using automated systems to originate more loans. Less paperwork, less human review and higher credit quality should reduce the cost of originating a mortgage, although some borrowers could face higher borrowing rates. [The effect of AU on mortgage costs and industry profitability is discussed in "AU, Borrowing Costs and the Profitability".]
How Fannie and Freddie's use of AU could make society worse off
While AU systems offer generic benefits, there are two ways in which Fannie and Freddie's AU initiatives could make society worse off. First, Fannie and Freddie could gain dominance over the AU market and lower the level of competition that would otherwise have occurred. As a result, home buyers could pay more to buy a home than they would have otherwise. Second, the systems could lead to less competition in the secondary mortgage market. This, in turn, increases Fannie and Freddie's level of indirect federal subsidization at a time when other methods of passing a subsidy provide greater certainty that home buyers will benefit. [The additional concern that federal subsidies to Fannie and Freddie will lead to greater than optimal investment in mortgage technology is discussed in "Can Society Ever Invest Too Much in Technology?"]
AU market dominance
As discussed in the September Region article, some economists have argued that
Fannie and Freddie do not operate in fully competitive markets.
Indeed, two economists wrote in Studies on Privatizing Fannie
Mae and Freddie Mac, a report commissioned by the Department
of Housing and Urban Development, that they were, "... confident
that Fannie Mae and Freddie Mac have been engaging in tacit collusion
in the conforming [secondary mortgage] market." (Conforming mortgages
fall below a legal limit making them eligible for purchase by Fannie
and Freddie.) In studying the structure of the mortgage market,
the authors found that Fannie and Freddie were able to have a controlling
position in this secondary market because they are given special
powers and indirect subsidies by the federal government.
Over the last two years, Fannie and Freddie have tried to become leaders in automated underwriting by introducing systems called Desktop Underwriter and Loan Prospector, respectively. As with the secondary market, Fannie and Freddie could take advantage of their special status and indirect subsidies to gain control over the AU market. Fannie and Freddie could try to initially gain controlling market share by offering their automated products at a lower price than would competitors who are not subsidized, all else being equal. Indeed, several industry observers believe that Fannie and Freddie will give the systems away as a means of increasing market share. In this strategy, Fannie and Freddie could use their subsidized status to keep prices low in order to gain market dominance and then raise prices in order to exploit their market position.
Fannie and Freddie have another extraordinary tool to gain market share in the AU market, namely, Fannie and Freddie can take advantage of their secondary market control. Indeed, many lenders depend on selling their mortgages to Fannie and Freddie in order to obtain the funds to make new mortgages. How could Fannie and Freddie use their control of this secondary mortgage market to gain AU market share? Fannie and Freddie, for example, have agreed to waive certain financial commitments normally required of lenders if the loan sold to them has been underwritten by their automated systems. No other competitor to Fannie and Freddie could offer this advantage. Fannie also offers volume discounts and rebates for loans underwritten by Desktop Underwriter that it acquires.
Finally, Fannie and Freddie have made their systems more attractive to originators by modifying them so they can underwrite all mortgages, including mortgages Fannie and Freddie will not finance because their credit quality is too low or because their value exceeds the conforming limit. These systems can also underwrite mortgages insured by the federal government even though Fannie and Freddie buy relatively few such mortgages.
Fannie and Freddie already report significant market penetration. In July 1996, Freddie reported that Loan Prospector underwrote 80 percent of all the loans underwritten by commercially available AU services. A stock analyst forecast that Fannie and Freddie's automated systems will review more than half of the nation's mortgage applications by 1998. More recently, Freddie reported that 350 lenders have adopted Loan Prospector and half of the mortgages it finances will be underwritten via the automated system by 1997. In addition, an analyst at the Wall Street investment firm Goldman Sachs has already increased his estimate of Freddie's earnings to take account of fees generated by Loan Prospector.
Dominance of the AU market by Fannie and Freddie could reduce society's well-being. If Fannie and Freddie dominate the market for AU systems, they would be in a position to raise the price for automated services above what the market would have otherwise charged. In addition, society would receive fewer benefits from innovations in AU if firms cannot effectively enter the AU market over time.
Furthermore, it will not be easy to dislodge Fannie and Freddie after they gain a controlling share in the AU market. Most importantly, Fannie and Freddie will always be able to use their control over the conforming, conventional secondary mortgage market to offer the users of their AU systems cost-saving opportunities and access to funds that no other firms could provide. In addition, it may be costly for originators to switch to a competitor's system after they incorporate Fannie or Freddie's system into their loan approval process; there could be very high transaction costs associated with switching systems. Finally, once a firm makes the significant fixed investment in AU technology, it would face little cost, over some range, for each additional loan it processes. Up to some point, the more loans it processes, the lower its average cost. This could make it difficult for new competitors to gain market share because their costs may be higher than those already supplying AU services.
It is probably true that the AU market will never be characterized by hundreds of firms providing AU systems. However, home buyers would almost certainly be better off with at least a limited number of firms (three or five, for example) competing to provide AU systems than they would be with a market dominated by two firms with a history of tacit collusion.
The Office of Federal Housing Enterprise Oversight (OFHEO), Fannie and Freddie's financial regulator, noted the power that AU systems offered Fannie and Freddie. OFHEO argued that the automated systems fundamentally change Fannie and Freddie's role in mortgage origination by allowing them to: (1) directly evaluate a loan application, (2) directly gather the information needed to underwrite a mortgage and (3) obtain complete data on all loan applications submitted to the systems. These changes could allow Fannie and Freddie to centralize, perform and profit from almost all the tasks previously undertaken by originators or firms, such as appraisers, that provided services to originators. In effect, the systems give Fannie and Freddie the potential to dominate aspects of mortgage production to which they previously did not have access.
Reduced competition in the secondary market
Gaining a dominant position in the AU market would allow Fannie
and Freddie to reduce the level of competition in the secondary
mortgage market. AU market dominance would allow Fannie and Freddie
to capture mortgages in the part of the secondary mortgage market
that they control that may have otherwise been sold in more competitive
parts of the secondary mortgage market or that would not have been
sold at all. The automated systems, for example, seem to encourage
sales of mortgages to Fannie and Freddie by placing them one step
closer to mortgage production and providing easier access to the
secondary market.
Fannie and Freddie's AU systems could also increase the number of mortgages that are purchased in the less competitive part of the secondary market by underwriting mortgages that Fannie and Freddie cannot, or normally do not, finance. Freddie, for example, has established a program to use Loan Prospector to underwrite subprime mortgages that do not meet its credit standards and jumbo mortgages that exceed the legal loan limit. Up to 25 percent of the loans that originators submit to Loan Prospector believing they are unqualified for financing by Freddie may actually be found acceptable by the underwriting system. As such, additional mortgages will be captured in the less competitive part of the secondary mortgage market.
Some observers also believe Loan Prospector will lead Freddie's effort to purchase, or guarantee pools of, mortgages it currently cannot finance. In fact, Freddie's chairman called for the elimination of rules that limit the size of the loans it can finance.
Increasing the number of mortgages Fannie and Freddie could finance would also raise their indirect subsidies by expanding the amount of outstanding securities implicitly guaranteed by the federal government. This increase in activity and indirect subsidies has real costs because there are alternative ways of providing subsidies to home buyers, such as a direct subsidy program, that would provide more certainty that home buyers benefit from the subsidy. The uncertainty associated with the indirect Fannie and Freddie subsidy were described in the previous Region article.
Weighing the costs and benefits of policies
Policy-makers will make best use of society's resources if they carefully weigh the benefits and costs of their policies. A cursory review of costs will, however, not suffice. The success of Fannie and Freddie in selling their AU systems, for example, could appear to be market driven. Indeed, the federal government has never explicitly subsidized the development of Fannie and Freddie's AU programs. However, both firms receive difficult-to-evaluate, indirect subsidies that could, in some years, significantly reduce their expenses. The fungible nature of the implicit subsidy means that any of their operations and investments, including the AU initiatives, could be subsidized.
Through the use of this indirect subsidy and their dominant position in the secondary mortgage market, Fannie and Freddie could gain a controlling market share in the automated mortgage market and reduce competition below what it would be otherwise. In turn, the automated systems could reduce competition in the secondary mortgage market and increase the implicit subsidies that Fannie and Freddie receive when more certain alternatives to this subsidy exist.