According to Fannie Mae, a limited cashout refi loan is a mortgage that meets a variety of criteria and which restricts of the amount the borrower can withdraw to not more than the lesser of 2% of the new refinance loan amount or $2,000. In addition, the funds must be utilized only for specific purposes including the payment of closing costs. As the GSE’s have never made a limited cashout flag available in its loan level files, we have never been able to look at the properties of these transactions.
It turns out however, that we can back into identifying some of these loans through the use of the Property Inspection Waver (PIW) flag. As we noted in a recent post, Fannie Mae sets a different standard for obtaining a PIW for a limited cashout refi than for other cashout refis. In fact, the criteria for a limited cashout is the same for a non-cashout refi.
Our Cohort Analyzer system contains the rules that Fannie Mae sets for obtaining an appraisal waiver based on the LTV-based criteria. The loan-level files we receive from Fannie Mae do not distinguish between limited cashout and other cashout refis, so the rules in the system to obtain a PIW are set to those for “normal” cashout refis (e.g LTV<= 60). Therefore, we surmise that cashout refi loans that have PIW’s with LTV>60 are in fact limited cashout refis.
Now Fannie Mae allows limited cashout refis to obtain an appraisal waiver based on the same rules as non-cashout refis so presumably they believe that the risk profiles between the two groups are similar. To test this, we first look at the different prepayment speeds between loans identified as limited cashout refis and noncashout refis with and without PIW’s, all in a range of LTV between 60 and 90.
The limited cashout loans prepay distinctly faster than non-cashout refis with LTV’s in the specified range.
Second, we compare the 30-day dq across this same classification, but this time the dataset used is the Fannie Mae Credit Risk Transfer (CRT) program Connecticut Avenue Securities (CAS) and not the loan-level tape.
The limited cashout refi cohort demonstrates a higher rate of delinquency than the other two since the onset of the Covid-19 crisis.
This analysis comes with a number of caveats including that limited cashout loans may have other distinguishing characteristics that drive their relative performance such as average loan size, coupon, cohort, etc. In addition, the limited cashout loans in CAS reference pools are a subset of the total and may not be a representative sample of the total universe of such loans.
This note demonstrates clearly however the power of digital big data tools to dig deeply into questions about loan performance that have previously been inaccessible to market analysts.
 Freddie treats limited cashout same as other cashout, so this analysis only limits to Fannie Mae.
 We have no way to identify limited cashout refis with ltv’s less than 60 or any limited cashout refis without a PIW.