ByAnthony Pennington-Cross , Giang Ho
Following the lead of federal regulations, numerous states, counties and cities have enacted laws designed to reduce predatory lending. There is at least anecdotal evidence that predatory or abusive mortgage lending is primarily concentrated in the subprime market. However, the impact of these local predatory lending laws on the subprime mortgage market is unknown. The primary questions we examine are: do these laws affect the supply and flow of subprime mortgage credit and does the experience in North Carolina, the first state to enact a local predatory lending law, apply to other local laws?
As discussed in a Housing and Urban Development (HUD)-Treasury report, defining predatory lending can be problematic. This difficulty arises because predatory lending depends on the inability of the borrower to understand the loan terms and the obligations associated with them. For example, some borrowers might be willing to accept a prepayment penalty in exchange for lower interest rates or fees because they do not expect to move in the near future. Or, the borrower might plan to diversify his or her portfolio away from a home and therefore would like an interest-only loan with a balloon payment in 10 years. However, interviews conducted by HUD, the Treasury Department and the Federal Reserve Board indicate that some, perhaps many, borrowers using high-cost loans might not have understood that the loan had a prepayment penalty or that it did not amortize through time, leading to a balloon payment.
At the national level, the Home Ownership and Equity Protection Act (HOEPA) and the regulations promulgated under it define a class of loans that are given special consideration because they are more likely to have predatory features and require additional disclosures. Loans covered under HOEPA include only closed-end home equity loans that have an annual percentage rate (APR) and/or finance fees exceeding a certain threshold. Specifically, the APR trigger is 8 percent and 10 percent above the Treasury rate for first and second lien loans, respectively. The fee trigger is inflation-adjusted and includes dollars paid at closing for optional insurance programs, such as health, credit life, accident, loss of income and other debt protection programs. Home purchase loans and other types of lending backed by a home, such as lines of credit, are not covered by HOEPA.
Local authorities have gone beyond HOEPA by introducing their own predatory lending laws that extend the restrictions on credit to an even broader class of mortgages. These restrictions include limits on allowable prepayment penalties and balloon payments, prohibitions of joint financing of various insurance products with the mortgage (such as credit, life and unemployment) and requirements that borrowers participate in loan counseling.
For example, North Carolina—the first state to enact predatory lending restrictions—expands the coverage of HOEPA by including both closed-end and open-end mortgages. However, reverse mortgages are not included and loan size is limited to the conventional conforming limit (loans small enough to be purchased by Fannie Mae and Freddie Mac and therefore not considered part of the jumbo market). North Carolina did leave the APR triggers the same as the HOEPA triggers, although the points and fees triggers were reduced from the HOEPA 8 percent of the total loan amount to 5 percent for loans under $20,000. For loans $20,000 or larger, the same 8 percent trigger is used or $1,000, whichever is smaller. The North Carolina law also prohibits prepayment penalties and balloon payments for most covered loans. The law prohibits the financing of credit life, unemployment, disability or other life and insurance premiums, while HOEPA includes them only as part of the trigger calculation.
Variation in the strength of local predatory laws typically comes from two sources. The first is the extent to which the law extends coverage beyond HOEPA. The second is the extent that the law restricts or requires specific practices. Law coverage is defined typically in terms of loan purpose, loan limit, APR and points-and-fees triggers. Broader coverage strengthens a law. On the other hand, the extent of a law's restrictions is typically defined by prepayment penalty and balloon restrictions, counseling requirements, restrictions on mandatory arbitration, and other factors. Local laws, such as in Chicago and Cook County, Ill.; Colorado; and Washington, D.C., have relatively broader coverage than others, while Cleveland, Georgia and New Mexico laws can be said to be more restrictive.
The widespread adoption of state and local predatory lending laws raises a natural question: What are the potential impacts of the laws on the subprime mortgage market? Unfortunately, no research to date (to our knowledge) has measured the costs and benefits of HOEPA and the state and local predatory lending laws. However, researchers have been able to measure how the volume of loans reacts to the introduction of a law. This helps answer the question of whether the laws reduce the supply of credit. Prior research has found convincing evidence that the North Carolina predatory lending law did reduce the supply of high-cost or subprime credit. There was some initial evidence that laws passed in Chicago and Philadelphia also had an impact. The laws can also specifically impact the prevalence of targeted loan types or loan-related characteristics, such as balloon payments and prepayment penalties. Balloon payment loans and prepayment penalties tended to become a smaller portion of the market after the law in North Carolina was introduced. Other potential impacts include substitution by lenders from one product type to another and reduced liquidity in the secondary market.
By introducing geographically defined predatory lending laws, policy-makers have effectively conducted a natural experiment with well-defined control and treatment groups. Since state boundaries reflect political and not economic regions, we can compare mortgage market conditions in states with a law in effect (the treatment group) to those in neighboring states currently without a predatory lending law (the control group). Specifically, using the treatment and control group framework, we tested to see whether local predatory lending laws affect the application and origination of subprime loans. We also tested to see the rates at which subprime loan applications are rejected. If volume is unaffected, then the flow and supply of credit to potential consumers has not been affected in the aggregate.
We extended prior research by examining the impacts in a variety of locations to see if the North Carolina experience is representative or typical for other states. Using publicly available Home Mortgage Disclosure Act (HMDA) data, we examined the change in subprime originations in each state before and after the law became effective. The loan samples were reduced by applying the prescribed loan limit (if any) under each law. Growth rates were calculated for loans associated with a list of subprime lenders as identified in the HUD subprime lender list. In an attempt to create as similar comparison groups as possible, we sampled only counties that border each other across state lines. Thus, a typical treatment group includes border counties in a state with a law in effect, and the corresponding control group includes border counties in neighboring states that do not have a law in effect during the observed time period (the year before and the year after the introduction of the law). This contrasts with other studies (see footnote 3) that have used whole neighboring states or regions to define both control and treatment groups. Our approach should help to increase the comparability of the treatment group and the control group because they are geographically closer and, as a result, likely to be more economically similar than full state and region comparisons. This approach and HMDA availability reduce the sample to 10 state predatory lending laws (California, Connecticut, Florida, Georgia, Maryland, Massachusetts, North Carolina, Ohio, Pennsylvania and Texas).
Using North Carolina as an example, the results show that from the year before to the year after the law becomes effective, subprime originations decreased by 35.8 percent in the treatment counties compared with 18.9 percent in the control counties. In other words, consistent with previous research on the North Carolina predatory lending law, subprime originations decreased substantially more than would be expected given the performance of the control counties. This finding also holds in four other states: Florida, Georgia, Massachusetts and Ohio. However, in the remaining five states—California, Connecticut, Maryland, Pennsylvania and Texas—we found that subprime originations increased more in the treatment locations. These results indicate that the experience in North Carolina might not extend to all other predatory lending laws, and that there might be sufficient variations in the laws that induce different responses in the flow of high-cost credit.
The relative changes in both subprime application and rejection rates are also examined. Again, the application results are mixed and very similar to the origination results. For example, four state laws—California, Maryland, Pennsylvania and Texas—experienced a relative increase in applications and six state laws—Connecticut, Florida, Georgia, Massachusetts, North Carolina and Ohio—experienced a relative decrease in applications. However, the rejection rates tell a much more consistent story. In most states, rejection rates declined more in the treatment locations than in the control locations, indicating that the introduction of predatory lending laws was associated with a disproportionate reduction in the rate that subprime applications were rejected. For example, California, Florida, Georgia and North Carolina experienced a relative decrease in rejection rates of at least 14.9 percentage points. At the other extreme, Pennsylvania and Connecticut experienced almost no relative change.
These results do not provide any indication that predatory lending laws systematically reduce the flow of subprime credit. However, the results do show that predatory lending laws tend to be associated with lower rejection rates of subprime mortgage applications. It can be expensive just to apply for a mortgage: the nonrefundable application fee usually runs from $200 to $300, not to mention other unobserved or nonpecuniary costs. Thus, while reducing rejection rates might not have been the primary purpose of the laws, a reduction in rejections can represent substantial savings to consumers and potentially lenders, too.
Starting with North Carolina in 1999, states and other localities across the United States have introduced legislation intended to curb predatory and abusive lending in the subprime mortgage market. These laws usually extend the reach of HOEPA by including home purchase and open-end mortgage credit, lowering the APR and fees-and-points triggers, and prohibiting or restricting the use of balloon payments and prepayment penalties on covered loans.
Using HMDA data on subprime loans and a sample of state laws, we found that the typical law has little impact on the flow of subprime credit as measured by loan originations, but is usually associated with lower rejection rates. In particular, local predatory lending laws can be associated with either increases or decreases in applications and originations for subprime loans. Earlier research on North Carolina law had found that the supply and flow of credit was reduced when the law became effective. We replicated this finding but did not find any evidence that the North Carolina experience applies to all other local predatory lending laws. It is likely that the exact nature of the law will impact the supply and flow of credit differently. For example, some laws are designed to provide broad coverage of the mortgage market (Chicago and Cook County laws) while other laws are more restrictive (Georgia and New Mexico laws) in terms of prohibiting or requiring certain practices.
To help identify why fewer subprime loans are being originated under some laws but not others, future research needs to examine how the coverage of the law and the restrictions imposed by the law impact the flow and cost of credit. Analysis should attempt to control for not only time and location but also law characteristics, borrower and loan characteristics, and economic conditions in both the control group and the treatment group. In addition, research should examine to what extent there is a regulatory cost associated with the laws that is passed on to borrowers through higher fees or interest rates.