Refinancing Frictions, Mortgage Pricing and Redistribution

Homeowners often refinance their mortgages to take advantage of more favorable terms like lower interest rates. Even though these changes often lead to significant savings over the duration of a home loan, not all homebuyers take equal advantage of them. In this paper, the authors distinguish between “fast” borrowers, who refinance frequently, and “slow” borrowers, who do not. They study whether these divergent refinancing behaviors have consequences for the broader market for mortgages. 

Regulations in the United States require mortgage lenders to offer fast and slow borrowers the same interest rates, a practice that results in what the authors refer to as “pooling.” While fast borrowers typically score a lower interest rate during the life of their home loan, slow borrowers are stuck with their original mortgage rate (or a higher rate based on less refinancing) until they pay it off. By the time they both pay off their loans, slow borrowers have often paid more towards their mortgages than fast borrowers have. This means, in effect, that slow borrowers subsidize fast borrowers.

The authors design a model that mimics this process. They apply their framework to data on actual mortgages in the United States and paint a systematic picture of how different refinancing behaviors impact the market. They find the following:

• Fast and slow borrowers receive the same rates at the outset of their mortgages. The data confirm, however, that the fastest borrowers indeed refinance more frequently and thus ultimately pay interest rates that are 100 basis points below those of the slowest borrowers. 

The authors use their framework to study what would happen to the market for mortgages if lenders were no longer required to pool fast and slow borrowers together. Their model allows them to capture both the direct effects of alternative mortgage scenarios, as well as indirect consequences. For example, even though a new policy might eliminate interest rate disparities between fast and slow borrowers, it could lead lenders to charge higher interest rates overall. The authors find the following:

• The authors first test what would happen under an alternative scenario where fast borrowers and slow borrowers are “separated,” such that fast borrowers receive higher rates and slow borrowers receive lower rates. They show that, on average, the fastest borrowers would end up paying interest rates that are 125 basis points higher. Notably, this 125-basis-points change is larger than the 100-basis-points difference that currently emerges between fast and slow borrowers, demonstrating the importance of capturing the indirect impacts of policies.
• The authors also use their model to assess what would happen under a scenario where mortgages refinanced automatically, with no active borrower intervention when rates decline. As one might expect, this scenario leads to much more refinancing for slow borrowers. It also leads to higher average mortgage rates at the outset, however, which may impact peoples’ homebuying decisions. The authors document an increase in mortgage rates at origination of about 110 basis points, which they estimate would force almost 20% of borrowers to select smaller homes. 
• The authors also study the effects of an alternative mortgage contract that prevents refinancing during the first few years of a mortgage. This policy eliminates the repeated “churning” of mortgages through refinancing, which reduces the dead-weight costs associated with mortgage origination. These savings are ultimately passed through into mortgage rates at origination, benefitting all borrowers. In addition, if such a policy were pursued under the current “pooling” arrangement, it would lead slow borrowers to pay less and fast borrowers to pay more towards their mortgages. 
• Finally, the authors test the effect of providing information about the economy to borrowers. They find that increasing attention about the economy would lead mortgage rates to rise by about 30 basis points. 

Approximately 20% of unconstrained US borrowers who would benefit financially from refinancing fail to do so. It is natural to think that policies leading to more frequent refinancing would improve borrower welfare and reduce inequality. The research presented here shows that while some policies can reduce the transfers associated with disparate refinancing behavior, they can often come at the costs of higher interest rates for borrowers, underscoring the importance of accounting for policy’s unintended consequences.