Securitization and Distressed Loan Renegotiation: Evidence From the Subprime Mortgage Crisis

1
Securitization and Distressed Loan Renegotiation
Evidence From the Subprime Mortgage Crisis

• Tomasz Piskorski, Amit Seru and Vikrant VigThe
  University of Chicago Working Paper (December
  2008)

Amy Younts April 26, 2009 MBAD 6160
2
Securitization Distressed Loan Renegotiation

• This paper investigates the effect of
  securitization on the activity of loan servicing.
• Several economic arguments suggest that
  securitized loans might be serviced differently
  from those directly held on the banks' balance
  sheets. The authors of the paper believe that
  seriously delinquent loans are foreclosed at a
  higher rate if they are securitized as compared
  to loans that are held by the bank (portfolio
  loans).

3
Loan Servicer Activities
Has your mortgage been sold to an outside
servicer?

• Loan originations and monitoring (to include
  servicing and restructuring) were once primarily
  conducted by loan officers inside the bank.
• Through securitizations, loans are now originated
  by the bank, then packaged and sold to investors
  who delegate the monitoring to a servicer.
  Therefore the servicing of the loans is no longer
  done by the final bearer of the risk.
• Approximately 60 of outstanding US mortgage debt
  is traded in mortgage-backed securities.
• Loan servicing not only includes the activity of
  processing the cash flow from the borrowers to
  the lending entity but also making the crucial
  decision either to foreclose a delinquent loan or
  prompt a workout/modification of the mortgage
  terms with the borrower.

4
Possible reasons for differing treatment

• Agency and incentive conflict brought about by
  the separation of ownership and control. This
  leads to less of a relationship and more of an
  arms length contracting which can make it more
  difficult to renegotiate debt contracts.
• Servicers may have greater incentive to foreclose
  delinquent loans than do bank-held loans due to
  the shift from concentrated debt to dispersed
  debt.
• Due to the securitizations bringing together
  multiple creditors with different classes of
  debt, it makes it harder to renegotiate debt
  contracts.

5
Possible reasons for differing treatment Legal
Liabilities

• Servicers of securitized loans are typically
  bound by Pooling and Servicing Agreement (PSA).
  Some PSA restrict the extent of loan
  modifications allowed.
• Servicers may be concerned about legal
  liabilities if modifications are not done
  properly and could benefit some borrowers and
  harm others. Modifications in the past were
  infrequent, so there was little guidance on this
  process on what is legally permissible to do.

6
Possible reasons for differing treatment Cost

• Both foreclosure and loan modifications involve
  cash outflows. The costs of loan modification can
  be substantial and servicers may not respond
  appropriately due to this.
• Pros
• Loan modification would help the servicer by
  extending the time period in which the servicing
  fee would be collected, so this would be
  preferred if they expect the cash flows from the
  borrower to exceed the cost of the loan
  modification
• Cons
• There are concerns of negative incentive effects
  and the fear modifications may trigger strategic
  defaults by borrowers.
• Cons- In most cases loan modifications/work outs
  do not enable servicers to funnel the costs
  through to borrowers or receive compensation for
  their expenses with securitized loans.
• In times with substantial house price risk,
  bringing greater possibly of refinancing and high
  likelihood of the borrower defaulting again it
  would be less attractive for servicers to choose
  the option of loan modification.

7
Analysis

• Examine differences in servicing of securitized
  loans at risk of foreclosure compared to loans
  held on the banks balance sheets
• Database of delinquent mortgages (both
  securitized and held of banks balance sheets) is
  used to track whether or not the delinquent loans
  from 2005 and 2006 were foreclosed. The McDash
  Analytics database is used and has less
  availability of data prior to 2005
• Logit regression conducted based on several loan
  characteristics. Regression is estimated
  separately for each quarter to lessen concern
  over changing macroeconomic conditions.

8
Data

• Loan level data having detailed information on
  the loan at the time of origination, such as loan
  amount, term, LTV ratio, type of loan, purpose of
  loan, credit score, geography where the loan is
  located and interest rate type are used in the
  analysis.
• Fully documented loans which provide verification
  of income as well as assets
• Focus on loans with purpose of home purchase
• First lien non-agency mortgages originated in
  2005-2006
• Loans with 15, 20 and 30 year maturities
• FICO scores were mostly between 550 and 800
• Criteria above left 6.2 million unique mortgages.
  About 75 were portfolio loans (25
  securitized).
• Analysis looked at 60 days delinquent (two
  mortgage payments behind) this left 330,000
  delinquent loans of which 11.3 were portfolio
  held.
• Also reviewed sub-sample of high-quality loans
  with FICO scores of at least 680.
• For high-quality sub-sample approximately 16,500
  were delinquent, with 20.4 being portfolio held
  at time of delinquency

9
Key Variables-Data Analysis

• Differences in the proportion of loans that are
  securitized depending on the riskiness of the
  loan. Approximately 11.2 of delinquent loans
  held on portfolio approximately 20.3 of high
  quality loans held

10
Other Impacts House Price Appreciation

• During periods of house price appreciation
  borrowers had a chance to build up additional
  equity and could exit delinquency by refinancing
  their loans or by selling their homes. This
  could have alleviated liquidation bias generated
  by failure to renegotiate securitized loans.
• Data for house prices and prepayment was reviewed
  to see the impact. Overall results showed
  borrowers were able to undo some of the
  liquidation bias generated by failure to
  renegotiate these loans during the time when
  house prices increased.

11
Other Impacts Varying Creditor Foreclosure Laws

• It was found foreclosure rate is much higher for
  loans originated in states with creditor-friendly
  laws that allow for quick foreclosure.
• The regression analysis was applied
  geographically . A state is defined as tough or
  creditor friendly if the historical average
  number of days to process a foreclosure is less
  than or equal to the median of about 117 days.
• On average delinquent loans in states with tough
  liquidation laws are about twice as likely to
  default as those states with weak liquidation
  law.
• This again shows servicers have a stronger
  incentive to foreclose a securitized loan in
  states where it is legally easier/faster to do
  so.

12
Other Impacts Lenders hold better quality loans

• It is a question whether lenders keep the better
  quality loans on their balance sheets based on
  unobservable characteristics they found.
• To alleviate this concern, the analysis on the
  subset of high quality loans was also conducted.
• The analysis is restricted to a sample of loans
  that have high quality of hard information such
  as credit score because soft info about borrowers
  is less likely to be valuable for borrowers on
  these type of loans at time of origination.
  (Studies are cited providing evidence for this.)
  
• The analysis on the high-quality loans had the
  same results that they were also serviced
  differently based on the securitization status of
  the loan.

13
Other Impacts Less information on securitized
loans after origination

• It is possible worse information is obtained for
  securitized loans between origination and
  delinquency.
• The analysis alleviates this concern by
  conditioning on the credit score of the borrower
  at the time of delinquency (since updated about
  every 3 months). This captures some of the
  information regarding quality of the borrower
  that may have changed from time of origination.

14
Data Analysis-Logit Regression
15
Results

• Results show the coeffient on Portfolio is
  consistently negative and significant for all
  quarters
• Other coefficient results are as anticipated
• Exception is FICO-conditional on being
  delinquent, loans with a lower FICO default less.
• Likelihood of a portfolio loan default is lower
  by 3-7.
• Being on the portfolio reduces the likelihood of
  default for a delinquent loan by at least 20
  relative to the mean.
• Results are the same in all quarters, but get
  stronger in quarters when house prices declined
  significantly.
• Results support view of substantial differences
  in the way delinquent loans are serviced
  depending on the securitization status.
• High quality loan analysis produced the same
  results
• This suggests conditional on being delinquent, a
  loan held on the lenders balance sheet is less
  likely to be foreclosed than a securitized loan.

16
Final Thoughts

• Authors feel their research suggests a loan held
  on the lenders balance sheet is less likely to
  be foreclosed than a securitized loan they do
  note they believe securitizations do have merit.
• They suggest that actions by the servicers of
  securitized mortgages may have a role in the
  recent wave of foreclosures.
• Authors feel their findings support ideas
  suggesting the government should take steps to
  address the lack of loan modifications on
  securitized loans. Rising foreclosures may lead
  to increased risk to the stability of the
  financial system so debt renegotiation might be
  necessary from a public policy perspective.