Reviewing the risks of non-bank sellers and servicing mortgages



An article I wrote that recently appeared in Housing wire on the nature of the risks of a non-bank mortgage company and Jeannie MaeThe proposal to introduce risk-based capital and liquidity requirements for their issuers emphasizes that non-bank lighter regulation of safety and security is an issue associated with the higher risk profiles of these firms compared to depositories.

In light of this important topic for industry and politics, I found it helpful to dig deeper into the arguments in the previous article.

In a comprehensive study of the liquidity of the mortgage market The liquidity crisis in the mortgage market, several prominent academics have provided compelling evidence of “liquidity vulnerabilities associated with non-banking institutions.” In their research, these researchers examined the 2008 financial crisis and its impact on non-bank liquidity. They found that: “Of the 19 non-banks and depositories that financed their institutions using both warehouse lines and SIVs in the pre-crisis period, only two, Nationstar Mortgage and Suntrust, lived to 2017 “. They further state:

“The collapse of the structure of short-term financing of non-banking organizations and some depositories, such as Financial in the whole country, led to a rapid loss of liquidity and credit activity. Many of these firms have experienced bankruptcies and closings, as have New Age… ”

Suffice it to say that not much has changed since then in terms of the liquidity risk potential of non-bank mortgage companies. Denials that these institutions pose relatively no greater risk to the mortgage industry at some point in the future than depositories, reminds me – starting their careers during the savings and loan crisis with the savings industry regulator and ending with the 2008 financial crisis in Citigroup increase the risk to their North American consumer division – these memories tend to fade into the details of past crises.

Another major risk for non-bank service providers and service providers is operational risk in terms of the quality of loan preparation. In the research I have done for MBAwith Housing Research Institute of America I analyzed the impact of the quality of credit production on credit loss and its implications for the industry. It is very important to understand how operational risks can be closely related to other risks such as credit.

In my previous article, I referred to a statistical analysis that showed that the risk of non-banks was significantly higher than that of depositories, taking into account other risk factors. I applied multivariate statistical analysis that matched the one I used in developing the industry’s first FHA an automated underwriting scorecard (now known as TOTAL) and an industry first VA scorecard during my work in Freddie Mac

Far from being a simple assessment of non-bank risk, it analytically monitors many factors such as a borrower’s credit rating, LTV, DTI and many other factors to understand the credit risk profile of borrowers. The analysis period was 2000-2015. To account for the frequency of the loan, two models were evaluated: one defining a default when the loan is 90 days overdue or worse, and the other a loan that is 180 days overdue or worse. …

I have segmented the time periods into four machine learning underwriting modes, grouping loan performance based on year and quarter. These were 2000-2003, II quarter of 2004-2008, III quarter of 2008 and 2012-2015. I created separate indicator variables to determine if the loan was originated by a custodian or non-bank and included them in the models along with other variables of interest.

The generic models had a high level of out-of-sample discrimination (KS 66 and 69 for the 90DPD + and 180DPD + models, respectively). The results table of these models is shown below, which shows the odds ratios (risk factors) for these indicator variables.

Multipliers of non-bank risk versus depositories


The base or benchmark (odds ratio = 1) for each odds ratio is the risk of default on mortgage loans issued by depositories between 2000 and 2003. Any odds ratios above 1 in the table indicate a higher risk for nonbanks compared to the baseline. All of the estimated coefficients underlying the odds ratios were statistically significant at the 995 level, with the exception of the 180DPD + model results for the 2000–2003 non-banking period, which did not differ from the baseline. Please note that for the 180DPD + model, the period 2012-2015. Has been phased out from Q3 2008 based on machine learning results.

The results show that with the exception of the period 2000-2003. Loans issued or serviced by non-banks were more risky than depositories providing or servicing loans in the 2000-2003 period. In particular, the 1.9 risk factor that I quoted in the previous article was taken from a simpler version of this model.

The results are consistent with this analysis, where in 2012–2015, loans from non-banks were 2.2 times more likely to become 90DPD + than depositories from the more normal underwriting period (2000–2003). Similar results (odds ratio 2.57) persist for non-bank service organizations in the period 2012-2015, which analytically confirms GAO study non-bank mortgage companies and their recommendations for strengthening regulatory oversight.

In particular, they stated: “Banks and non-bank service providers are subject to different safety and security rules and different capital rules. As a result, mortgage market participants and others questioned the extent to which non-bank service centers could pose additional risk to consumers and the market, and whether the current supervisory system can ensure the safety and reliability of non-bank service centers. ”

The main findings are that non-banks, by virtue of their business model and regulatory oversight, carry more liquidity and operational / credit risks than depository institutions, all things considered.

Clifford Rossi is Professor of Practice and Executive Director of the Robert H. Smith School of Business at the University of Maryland. He has 23 years of industry experience and has held several senior risk management positions at major financial institutions.

This column does not necessarily reflect the views of the HousingWire editorial team and their owners.

To contact the author of this story:
Clifford Rossi

To contact the editor responsible for this story:
Sarah Wheeler at


Source link