oversight

HUD Did Not Always Provide Adequate Oversight of Its Section 203(k) Rehabilitation Loan Mortgage Insurance Program

Published by the Department of Housing and Urban Development, Office of Inspector General on 2015-07-31.

Below is a raw (and likely hideous) rendition of the original report. (PDF)

     U.S. Department of Housing and
   Urban Development, Office of Single
    Family Housing – Washington, DC
         Section 203(k) Rehabilitation Loan Mortgage
                      Insurance Program




Office of Audit, Region 5       Audit Report Number: 2015-CH-0001
Chicago, IL                                           July 31, 2015
To:            Kathleen Zadareky, Deputy Assistant Secretary for Single Family Housing, HU

               //signed//
From:          Kelly Anderson, Regional Inspector General for Audit, 5AGA

Subject:       HUD Did Not Always Provide Adequate Oversight of Its Section 203(k)
               Rehabilitation Loan Mortgage Insurance Program


Attached is the U.S. Department of Housing and Urban Development (HUD), Office of Inspector
General’s (OIG) final results of our review of HUD’s oversight of its Section 203(k)
Rehabilitation Loan Mortgage Insurance program.
HUD Handbook 2000.06, REV-4, sets specific timeframes for management decisions on
recommended corrective actions. For each recommendation without a management decision,
please respond and provide status reports in accordance with the HUD Handbook. Please furnish
us copies of any correspondence or directives issued because of the audit.
The Inspector General Act, Title 5 United States Code, section 8M, requires that OIG post its
publicly available reports on the OIG Web site. Accordingly, this report will be posted at
http://www.hudoig.gov.
If you have any questions or comments about this report, please do not hesitate to call me at
312-353-7832.
                    Audit Report Number: 2015-CH-0001
                    Date: July 31, 2015

                    HUD Did Not Always Provide Adequate Oversight of Its Section 203(k)
                    Rehabilitation Loan Mortgage Insurance Program




Highlights

What We Audited and Why
We audited the U.S. Department of Housing and Urban Development’s (HUD) oversight of its
Section 203(k) Rehabilitation Loan Mortgage Insurance program as part of the activities in our
fiscal year 2014 annual audit plan. Our audit objective was to determine whether HUD had
adequate oversight of its Section 203(k) program.

What We Found
HUD needs to improve its monitoring of lenders for compliance with the Section 203(k) program
requirements because lenders did not always ensure that (1) borrowers or contractors obtained
required building permits to rehabilitate properties and (2) contractors were licensed or certified
to perform rehabilitation work. In addition, lenders did not always ensure that contractors’ cost
estimates contained clear descriptions of the proposed repairs to determine eligibility for the
Streamlined (k) program. As a result, HUD lacked assurance of the soundness of the repairs,
thus potentially impacting the safety of the borrowers and increasing the risk to the Federal
Housing Administration’s (FHA) Mutual Mortgage Insurance Fund by more than $1.2 million.
Further, HUD did not always ensure that (1) loan-to-value ratios were correctly calculated when
determining borrowers’ monthly mortgage insurance premiums and (2) lenders properly entered
borrowers’ loan information into FHA Connection. As a result, HUD lacked assurance that it (1)
properly managed the risk to FHA’s Mutual Mortgage Insurance Fund and (2) protected the
interests of borrowers due to the overpayment of mortgage insurance. We estimate that nearly
28,000 borrowers had overpaid their premiums by more than $3.2 million as of December 31,
2014, and will continue to overpay their premiums by more than $1.9 million over the next year.

What We Recommend
We recommend that HUD’s Deputy Assistant Secretary for Single Family Housing require
lenders to (1) support or indemnify HUD for any future losses on the 40 loans with estimated
losses totaling more than $1.2 million and (2) support or reimburse HUD for the actual losses
incurred on two loans totaling $83,322. We also recommend that HUD (1) strengthen its
controls over Section 203(k) program requirements, (2) adjust the formula for calculating the
loan-to-value ratio, (3) determine the overpaid mortgage insurance premiums for loans with
incorrect loan-to-value ratios, and (4) credit the accounts of active borrowers who overpaid their
mortgage insurance premiums and refund overpaid premiums to borrowers for terminated loans.
Table of Contents
Background and Objective......................................................................................3

Results of Audit ........................................................................................................5
         Finding 1: HUD Needs To Improve Its Monitoring of Lenders For Compliance
         With the Section 203(k) Program Requirements ........................................................... 5

         Finding 2: HUD Did Not Always Ensure That Loan-to-Value Ratios Were
         Properly Calculated When Determining Borrowers’ Mortgage Insurance
         Premiums ......................................................................................................................... 11

Scope and Methodology .........................................................................................15

Internal Controls ....................................................................................................23

Appendixes ..............................................................................................................25
         A. Schedule of Questioned Costs and Funds To Be Put to Better Use ...................... 25

         B. Auditee Comments and OIG’s Evaluation ............................................................. 26

         C. Criteria ....................................................................................................................... 33

         D. Schedule of Lender Compliance Deficiencies ......................................................... 38

         E. Schedule of HUD Review Deficiencies .................................................................... 40

         F. Estimated Losses to HUD From Deficiencies ......................................................... 41

         G. Schedule of Overcalculated Loan-to-Value Ratio Deficiencies ............................ 42

         H. Schedule of Undercalculated Loan-to-Value Ratio Deficiencies .......................... 45




                                                                     2
Background and Objective
The Federal Housing Administration (FHA) provides mortgage insurance on loans made by
FHA-approved lenders throughout the United States and its territories. It insures mortgages on
single-family and multifamily homes, including manufactured homes and hospitals. FHA is the
largest insurer of mortgages in the world, having insured more than 34 million properties since
its inception in 1934. Mortgage insurance provides lenders protection against losses as a result
of homeowners defaulting on their loans. The lenders bear less risk because FHA will pay a
claim to the lender if a homeowner defaults. Loans must meet established requirements to
qualify for insurance.

Congress established the Section 203(k) Rehabilitation Loan Mortgage Insurance program in
1978. It is the U.S. Department of Housing and Urban Development’s (HUD) primary program
for the rehabilitation and repair of single-family properties. The program is an important tool for
community and neighborhood revitalization and expanding home-ownership opportunities
because it allows the purchase or refinance of a single-family property and the cost for repairs
and nonluxury improvements to be included in the loan amount.

A 203(k) loan may be used to rehabilitate an existing one- to four-unit dwelling by (1)
purchasing a structure and the land on which the structure is located and rehabilitating it; (2)
purchasing a structure on another site, moving it onto a new foundation on the mortgaged
property, and rehabilitating it; (3) refinancing the existing indebtedness and rehabilitating such a
structure; or (4) rehabilitating such a structure.

HUD offers two Section 203(k) Rehabilitation Loan Mortgage Insurance programs, the Standard
(k) and Streamlined (k). The Streamlined (k) program is used for property repairs or
improvements that cost $35,000 or less. The Standard (k) program is used for properties that
require extensive repairs, including major additions and structural changes.

Under both programs, borrowers are required to obtain all licenses and permits required by local
governmental authorities. Funds from borrowers’ escrow accounts cannot be released until the
local HUD field office or direct endorsement underwriter is assured that these requirements have
been satisfied and the fees have been paid. For the Streamlined (k) program, HUD requires
lenders to ensure that contractors meet licensing requirements, and the repairs and improvements
must comply with any local codes and ordinances. The borrower or contractor must obtain all
required permits before starting the work.

In accordance with its requirements, FHA establishes and collects a single premium payment
(upfront mortgage insurance premium) and annual mortgage insurance premium based on the
loan-to-value ratio of a loan. The annual mortgage insurance premium is determined by
calculating the loan-to-value ratio, which is the mortgage insurance amount (excluding the
upfront mortgage insurance premium) divided by the appraised value. If the loan-to-value ratio
is less than 90 percent, FHA will collect monthly premiums for the first 11 years of the mortgage


                                                  3
term. If the loan-to-value ratio is greater than or equal to 90 percent, FHA will collect monthly
premiums for the lesser of the mortgage term or the first 30 years of the mortgage term.
However, for loans closed on or after January 1, 2001, with terms of more than 15 years, FHA’s
annual premium is automatically canceled when the loan-to-value ratio reaches 78 percent,
provided that the borrower paid the annual premium for at least 5 years.

HUD’s Office of Single Family Housing is responsible for the overall management and
administration of the FHA single-family mortgage insurance programs and provides guidance for
and oversight of the lenders that participate in its mortgage insurance programs. Its oversight
authorities include HUD’s Homeownership Centers, which are located in Philadelphia, PA,
Denver, CO, Santa Ana, CA, and Atlanta, GA. Within the Homeownership Centers are the
Processing and Underwriting and the Quality Assurance Divisions.

The Processing and Underwriting Division performs postendorsement technical reviews on
selected FHA-insured loans to evaluate the risk that loans represent to FHA’s insurance funds
and lenders’ compliance with FHA’s requirements. The Quality Assurance Division engages in
routine and continual monitoring of FHA-approved lenders to identify errors and noncompliance
and mitigate risks to the FHA insurance funds. The Division targets lenders for review using a
centralized and coordinated targeting methodology, which adapts to changes in business
practices and market conditions.

Our objective was to determine whether HUD had adequate oversight of its Section 203(k)
program. Specifically, we wanted to determine whether HUD ensured that (1) lenders endorsed
loans that complied with program requirements and (2) loan-to-value ratios were properly
calculated when determining borrowers’ monthly mortgage insurance premiums.




                                                4
Results of Audit

Finding 1: HUD Needs to Improve Its Monitoring of Lenders For
Compliance With the Section 203(k) Program Requirements

HUD needs to improve its monitoring of lenders for compliance with the Section 203(k) program
requirements because lenders did not always ensure that (1) borrowers or contractors obtained
building permits to rehabilitate properties and (2) contractors were licensed or certified to
perform rehabilitation work. Lenders also did not ensure that contractors’ cost estimates
contained clear descriptions of the proposed repairs to determine eligibility for the Streamlined
(k) program. These weaknesses occurred because HUD lacked adequate procedures and controls
to ensure that lenders complied with its requirements. As a result, HUD lacked assurance of the
soundness of the repairs, thus potentially impacting the safety of borrowers and increasing the
risk to FHA’s Mutual Mortgage Insurance Fund by more than $1.2 million.

FHA Insured Section 203(k) Loans That Did Not Always Comply With Program
Requirements
Using HUD’s Single Family Data Warehouse,1 we identified 70,196 loans insured under HUD’s
Section 203(k) program that closed from December 1, 2008, through July 31, 2013. We
reviewed 106 of the 70,196 loans totaling more than $16 million for compliance with the
program’s requirements.

Of the 106 loans reviewed, 37 (35 percent) did not comply with the program’s requirements.
Specifically, FHA insured loans without support that (1) borrowers or contractors obtained
building permits required by local government authorities to rehabilitate properties and (2)
contractors were specialty licensed2 or certified to perform rehabilitation work that involved
plumbing, mechanical, or electrical repairs or the disturbance of painted surfaces.3 Additionally,
for HUD’s Streamlined (k) program, lenders did not always require contractors to provide clear
descriptions of the proposed work4 to ensure that (1) the work would not involve structural
repairs and (2) all required repairs were included.5 The table below shows the deficiencies for
the 37 loans.6


1
  Single Family Data Warehouse is a large, extensive collection of database tables organized and dedicated to
support the analysis, verification, and publication of the Office of Single Family Housing’s data.
2
  HUD Handbook 4240.4, REV-2, section 4-9 and appendix 2, and Mortgagee Letter 2005-50. See appendix C for
details on related criteria.
3
  As of April 2010, the Environmental Protection Agency requires that contractors that renovate, repair, or prepare
surfaces for painting in properties built before 1978 be certified. HUD included this requirement in its Frequently
Asked Questions’ Valuation Protocol in January 2013.
4
  Mortgagee Letter 2005-50
5
  If structural work was involved, loans would not be eligible under the Streamlined (k) program.
6
  Fifteen of the thirty-seven loans contained two deficiencies.



                                                          5
                                      Type of deficiencies                             Count
                  Lacked support that building permits were obtained before
                  the repairs started                                                        32
                  Lacked support that contractors had a specialty license or
                  lead-based paint certification                                             17
                  Lacked clear descriptions to determine whether the
                  proposed repairs were structural or nonstructural                            2
                  Lacked support that required repairs were completed                         1
                            Total                                                            52

The table in appendix D of this report shows the loans with the deficiencies cited above.
Additionally, appendix C contains the related criteria.

HUD Reviewed Its 203(k) Loans
Using HUD’s Single Family Data Warehouse system and data provided by HUD’s Quality
Assurance Division, we determined that HUD reviewed 2,453 loans insured under its Section
203(k) program during the period January 1, 2011, through December 31, 2012. We reviewed
86 of the 2,453 loans to determine whether HUD adequately identified and mitigated lenders’
noncompliance with the program’s requirements.

HUD did not always identify lenders’ noncompliance with the program’s requirements.
Specifically, of the 86 loans reviewed, 187 (21 percent) contained deficiencies that were not
identified by HUD.

       For 12 loans, there was no evidence that borrowers or contractors obtained required
        building permits in accordance with local building codes.
       For six loans, there was no evidence that contractors were licensed or certified to perform
        rehabilitation work that involved plumbing, mechanical, or electrical repairs or the
        disturbance of painted surfaces.
       For five loans, the contractors’ cost estimates did not provide clear descriptions to
        determine whether the repairs involved structural or nonstructural rehabilitation work.8
       For two loans, the repairs were conditioned on the direct endorsement underwriter form
        HUD 54114, or appraisal report was not sufficiently addressed by the contractors.
       For one loan, the cost estimate did not contain a breakdown of the costs for labor and
        materials to ensure that the borrower was not reimbursed for labor.

The table in appendix E of this report shows the loans with the deficiencies cited above.
Additionally, appendix C contains the related criteria.




7
  Seven of the eighteen loans contained more than one deficiency.
8
 A total of seven loans (2 from the Section 203 (k) loan review + 5 from HUD’s review of Section 203(k) loans)
lacked clear descriptions to determine whether the repairs were structural or nonstructural.



                                                        6
HUD Lacked Adequate Procedures and Controls
HUD relied on the lenders to ensure that borrowers complied with the Section 203(k) program
requirements. For instance, borrowers were expected to determine the soundness of the property
before and after rehabilitation, including the value, cost estimates, and ability of the contractor to
complete the rehabilitation in a satisfactory, workmanlike manner in compliance with all
accepted exhibits and local codes and ordinances as outlined in the Section 203(k) borrower’s
acknowledgment form. According to HUD, although the program had requirements for the
borrowers, lenders had the overall responsibility to ensure that borrowers or contractors obtained
all permits. However, several lenders relied on the contractor or borrower to obtain the
necessary permits or licenses and did not verify these permits or licenses with the responsible
party or the local government authority.

HUD also did not always perform adequate monitoring and oversight of lenders’ compliance
with program requirements. According to HUD, it lacked the resources to contact the many
local governments in the United States to review their guidelines and obtain required
documentation since there was no central repository. Therefore, during a postendorsement
technical review, a HUD reviewer would review only the documentation that the lender was
required to maintain in the FHA case binder, which did not include evidence of building permits
and contractor licensing. However, according to HUD, the number of loans insured under its
Section 203(k) program was small compared to the number of loans insured under the Section
203(b) program. Therefore, its reviewers may have reviewed only one or two 203(k) loans per
week, if any. Specifically, a HUD reviewer stated that the 203(k) loans usually made up only 1
percent of the number of loans reviewed during a postendorsement technical review. For
instance, for every 10 loan files reviewed, only 1 may have been for a 203(k) loan. Therefore, it
would be unlikely that a HUD reviewer would have to contact many local jurisdictions when
reviewing a 203(k) loan as part of a postendorsement technical review.

In June 2013, HUD developed a standardized supplemental review checklist to assist reviewers
with reviewing loans insured under the Section 203(k) program. Both HUD’s Standard (k) and
Streamlined (k) programs require contractors to be licensed and obtain building permits as
applicable.9 However, the supplemental review checklist did not include a review for permits
under the Streamlined (k) program. Further, HUD’s Post Endorsement Technical Review Desk
Guide did not address how the checklist would be used to determine whether the loan was
deficient or unacceptable.

Additionally, according to the Director of HUD’s Single Family Home Mortgage Insurance
Division, HUD had not formally adopted the Environmental Protection Agency’s renovation,
repair, and painting rule for properties that were not HUD’s real estate-owned or federally owned
and targeted housing that receives Federal assistance. It also had not issued guidance regarding
lead-based paint remediation requirements for non-real estate-owned single-family properties.
However, in January 2013, HUD included this requirement in its Frequently Asked Questions’
Valuation Protocol. Therefore, it had guidance requiring lenders to comply with the
Environmental Protection Agency’s requirement.
9
  HUD Handbook 4240.4, REV-2, section 4-9 and appendix 2, and Mortgagee Letter 2005-50 - See appendix C for
related criteria.



                                                      7
Further, contractors’ cost estimates were not always clear to sufficiently determine whether the
loan was eligible for the Streamlined (k) program. According to the Director of HUD’s Home
Mortgage Insurance Division, lenders should determine whether the repairs would be structural.
However, HUD did not require lenders to support their determination that the work would not
involve structural repairs. For instance, for FHA case number 137-5861827, the contractor’s
cost estimate included the demolition of walls in two bedrooms, the dining room, the attic, the
kitchen, and the bathroom. However, the cost estimate did not specify whether any of the repairs
were for load-bearing structural walls. Further, the lender’s loan file did not contain
documentation to support how the lender determined that the demolition of the walls was eligible
under the Streamlined (k) program.

Conclusion
HUD lacked adequate procedures and controls to ensure that lenders complied with its
requirements. As a result of these deficiencies, HUD lacked assurance regarding the soundness
of the repairs, thus potentially impacting the safety of the borrowers and increasing the risk to
FHA’s Mutual Mortgage Insurance Fund by more than $1.2 million for 40 active loans.10

Recommendations
We recommend that HUD’s Deputy Assistant Secretary for Single Family Housing require the
lenders to

         1A.      Support that the repairs to the properties associated with the 32 loans without
                  evidence of permits complied with local code or reimburse HUD $792,837 for the
                  escrow repair funds.

         1B.      Support that the repairs to the properties associated with the six loans were not
                  structural repairs11 or indemnify HUD for the four active loans with a total
                  estimated loss of $222,07312 and reimburse HUD for the actual loss of $83,322
                  incurred on the sale of two properties associated with FHA case numbers 052-
                  4308836 and 034-8239100.




10
   Of the 55 deficient loans cited in the finding (37 + 18), 46 were active, 7 had been paid in full, and HUD paid
claims on the remaining 2 loans according to HUD’s Single Family Data Warehouse as of January 1, 2015. Of the
46 active loans, we did not question the cost for 4 loans. For these loans, even though there was no evidence that the
contractors were licensed or required permits were obtained after the repairs were completed, the properties passed
their building inspections. Further, we did not question the costs associated with two loans that were cited only for
potentially lead-based paint issues because the loans closed before HUD included the Environmental Protection
Agency’s requirement in its Frequently Asked Questions in January 2013. Therefore, only 40 (55 – 7 – 2 – 4 – 2)
active loans were cited for questioned costs in this finding. See appendix F.
11
   A total of seven loans were cited in this finding for unclear cost estimates concerning whether repairs involved
structural or nonstructural rehabilitation work (see appendixes D and E). Six of the seven loans were reported in
recommendation 1B, and the remaining loan was reported in recommendation 1D.
12
   This amount was based on the loss severity rate of 50 percent of the total unpaid principal balances of $444,145
for the four loans as of January 29, 2015.



                                                          8
      1C.    Support that the borrower for FHA case number 451-1165810 was not reimbursed
             for the cost of labor or indemnify the loan with an estimated loss amount of
             $83,715, based on the loss severity rate of 50 percent of the unpaid principal
             balance of $167,429 as of January 29, 2015.

      1D.    Support that the repair conditions and comments indicated in the direct
             endorsement underwriter form, form HUD-54114, were satisfied for FHA case
             number 501-8198149. If the repair conditions and comments were not properly
             addressed, the lenders should indemnify the loan with an estimated loss amount of
             $39,367, based on the loss severity rate of 50 percent of the unpaid principal
             balance of $78,733 as of January 29, 2015.

      1E.    Support that the required repairs to the property associated with FHA case
             numbers 241-9513470 and 277-1438986 were sufficiently addressed and
             complied with local codes or indemnify HUD for the estimated loss of $97,355,
             based on the loss severity rate of 50 percent of the unpaid balance of $194,709 as
             of January 29, 2015.

We also recommend that HUD’s Deputy Assistant Secretary for Single Family Housing

      1F.    Revise HUD’s policies and procedures, including the supplemental 203(k) review
             checklist and review rating guidelines, to ensure that HUD reviewers properly and
             consistently identify and resolve deficiencies.

      1G.    Ensure that postendorsement technical reviewers receive sufficient training to
             understand the scope of the repairs and improvements by contractors.

      1H.    Provide clarification to lenders regarding their responsibility to ensure that (1)
             repairs or improvements comply with local building codes, (2) contractors that
             perform specialized work are appropriately licensed, and (3) contractors provide
             clear descriptions of the work to be performed to sufficiently determine whether
             the repairs or improvements are structural.

      1I.    Revise its existing policies governing the 203 (k) program or implement a new
             policy requiring lenders to review the scope of the repairs or renovations and
             determine whether (1) building permits are required for the work and (2)
             contractors meet jurisdictional licensing and bonding requirements. The policy
             should also require lenders to maintain documentation supporting their review and
             determination.

      1J.    Develop and implement a complete training program for lenders that participate
             in the 203(k) programs. The training program should include but not be limited to
             ensuring the lenders understand their responsibilities to communicate FHA’s
             203(k) requirements to their client borrowers and contractors concerning




                                               9
      contractor licensing and building permits as required by local government
      authorities.

1K.   Communicate with HUD reviewers and lenders about HUD’s inclusion of the
      Environmental Protection Agency’s renovation, repair, and painting (lead safe)
      rule in HUD’s Valuation Protocol Frequently Asked Questions.




                                       10
Finding 2: HUD Did Not Always Ensure That Loan-to-Value Ratios
Were Properly Calculated When Determining Borrowers’ Mortgage
Insurance Premiums

HUD did not always ensure that (1) loan-to-value ratios were correctly calculated when
determining borrowers’ monthly mortgage insurance premiums and (2) lenders properly entered
borrowers’ loan information into FHA Connection.13 These weaknesses occurred because HUD
lacked adequate procedures and controls to ensure that (1) the formula for calculating the loan-
to-value ratio in its Computerized Homes Underwriting Management System14 was accurate and
(2) it provided lenders with complete and accurate information and adequately monitored lenders
to ensure that they entered accurate data into FHA Connection. As a result, HUD lacked
assurance that it (1) properly managed the risk to FHA’s Mutual Mortgage Insurance Fund and
(2) protected the interests of borrowers due to the overpayment of mortgage insurance. We
estimated that nearly 28,000 borrowers had overpaid their premiums by more than $3.2 million
as of December 31, 2014, and will continue to overpay their premiums by more than $1.9 million
over the next year.

HUD’s System Did Not Properly Calculate Borrowers’ Mortgage Insurance Premiums
Using HUD’s Single Family Data Warehouse, we identified 70,196 loans insured under HUD’s
Section 203(k) program that closed from December 1, 2008, through July 31, 2013. We selected
150 loans totaling more than $22 million to determine the accuracy of borrowers’ mortgage
insurance premiums.

HUD did not ensure that the formula in its Computerized Homes Underwriting Management
System correctly calculated loan-to-value ratios15 when determining borrowers’ mortgage
insurance premiums. Specifically, the loan-to-value ratios for 92 (61 percent) of the 150 loans
reviewed were either overcalculated or undercalculated.

        The loan-to-value ratios were overcalculated for 80 loans; thus, the borrowers would pay
         the monthly premiums for a longer period over the life of their loans. Further, borrowers
         for 61 of the 80 loans had overpaid their monthly insurance premiums by $12,576 as of
         September 30, 2014, and will continue to overpay.16 Although the borrowers for the
         remaining 19 loans had not overpaid their monthly premiums, they will pay their
         insurance premiums for additional months due to the overcalculated loan-to-value ratios.


13
   FHA Connection is an Internet-based system that allows FHA-approved lenders to have real-time access to
several of FHA’s systems over HUD’s Internet system for the purpose of originating and servicing FHA loans.
14
   The Computerized Homes Underwriting Management System automates the single-family mortgage insurance
application process.
15
   HUD’s Computerized Homes Underwriting Management System manual, dated November 2011. See appendix C
for details on related criteria.
16
   Borrowers will continue to overpay or underpay their premiums for at least 5 years and until their loan-to-value
ratio reaches 78 percent or the loan is no longer insured by HUD.



                                                         11
        The loan-to-value ratios were undercalculated for 12 loans; thus, the borrowers would
         pay the monthly premiums for a shorter period over the life of their loans. Further,
         borrowers for 2 of the 12 loans had underpaid their premiums by $158 as of September
         30, 2014, and will continue to underpay. Although the borrowers for the remaining 10
         loans had not underpaid their monthly premiums, they will pay their insurance premiums
         for fewer months due to the undercalculated loan-to-value ratios.

We estimated that nearly 28,000 borrowers had overpaid their premiums by more than $3.2
million as of December 31, 2014, and will overpay their premiums by more than $1.9 million
over the next year.17

The tables in appendixes G and H of this report represent the loans with the deficiencies cited
above.

Lenders Did Not Properly Enter Loan Data Into FHA Connection for the Computation of
Loan-to-Value Ratios
HUD did not always ensure that lenders correctly and consistently entered loan data into FHA
Connection. Of the 92 loans with inaccurate loan-to-value ratios, for 65 loans,18 lenders
incorrectly entered (1) property appraised values for 37 loans, (2) contract sale prices for 15
loans, or (3) repair escrow amounts for 31 loans.19

HUD Lacked Adequate Procedures and Controls
HUD lacked adequate procedures and controls to ensure that its formula for calculating the loan-
to-value ratio in its System was accurate. To calculate the ratio, a lender would enter a
property’s appraised value into HUD’s System. The System would then use the value to
calculate the (1) maximum mortgage amount and (2) loan-to-value ratio for determining the
borrower’s premium. According to the System’s manual, for loans that involve the purchase of a
property that was endorsed on or after November 12, 2008, to calculate the loan-to-value ratio,
the System would use the original mortgage amount without the upfront premium divided by the
lesser of (1) the appraised value or (2) the sale price plus the repair escrow amount.20 However,
according to HUD, to determine the loan-to-value ratio for calculating borrowers’ premiums, the
System should divide only the original mortgage amount without the upfront premium by the
appraised value.21



17
   See Scope and Methodology.
18
   Fourteen loans contained more than one deficiency. See appendixes G and H.
19
   The property appraised value, contract sale price, and repair escrow amount data fields were needed to compute
borrowers’ mortgage insurance premiums in HUD’s System.
20
   For Section 203(k) loans that were refinanced and closed from November 13, 2008, through September 21, 2009,
the formula for calculating the loan-to-value ratio was the mortgage without the upfront premium divided by the
unpaid principal balance plus any escrow funds. For Section 203(k) loans that were refinanced after September 21,
2009, the formula is the mortgage amount without the upfront premium divided by the appraised value.
21
   We obtained a legal opinion from the Office of Inspector General’s Office of Legal Counsel regarding the
definition of appraised value for determining borrowers’ monthly mortgage insurance premium. Counsel opined
that the appraised value is the after-improved value of the property. See appendix C for details.



                                                         12
In addition, HUD did not always provide lenders with complete and accurate information. In
May 2012, HUD posted instructions on FHA Connection’s message board requesting that
lenders enter the lesser of the total sale price plus the repair escrow amount or 110 percent of the
after-improved value from the maximum mortgage calculation worksheet, form HUD-92700,
into the appraised value field in FHA Connection to calculate the loan-to-value ratio. These
instructions resulted in some lenders contacting HUD’s homeownership centers seeking
additional clarification or guidance. However, the clarification or guidance that the lenders
received was not consistent. Specifically, a lender disagreed with HUD’s posted instructions
regarding the data that should be entered into FHA Connection as a property’s appraised value
because (1) the lesser of the total sale price plus the repair amount or 110 percent of the after-
improved value would result in the loan-to-value ratio always being 96.5 percent for all loans
and (2) repairs to rehabilitate properties do not always result in a dollar-for-dollar return.
Therefore, using this approach to determine the value of the property was not accurate.

In 2012, HUD created a working group to consolidate all FHA mortgagee letters and update its
handbook. In addition, the group worked on resolving the inconsistencies regarding the
calculation of the loan-to-value ratio. However, until these changes are fully implemented and
communicated to the lenders, the process for underwriting, regarding the computation of the
loan-to-value ratio for borrowers’ mortgage insurance premiums, will not be consistent among
lenders participating in the Section 203(k) program.
Conclusion
The deficiencies described above occurred because HUD lacked adequate procedures and
controls to ensure that (1) the formula for calculating the loan-to-value ratio in its System was
accurate and (2) it provided lenders with complete and accurate information and adequately
monitored lenders to ensure that they entered accurate data into FHA Connection. As a result,
HUD lacked assurance that it (1) properly managed the risk to FHA’s Mutual Mortgage
Insurance Fund and (2) protected the interests of borrowers due to the overpayment of mortgage
insurance. We estimated that nearly 28,000 borrowers had overpaid their premiums by more
than $3.2 million as of December 31, 2014, and will continue to overpay their premiums by more
than $1.9 million over the next year.

Recommendations
We recommend that HUD’s Deputy Assistant Secretary for Single Family Housing

        2A.      Reimburse or apply $10,552 in credit to borrowers’ future premiums for the 54
                 active loans22 with overpaid premiums and refund $2,024 to the borrowers of the 7
                 terminated loans.




22
  Of the 61 loans cited for overpayment of mortgage insurance premiums, 54 were active in HUD’s Single Family
Data Warehouse as of January 1, 2015. The remaining seven loans had been terminated. The total premium
overpayment for the seven terminated loans was $2,024. Thus, the total amount of the overpaid premiums for the
remaining 54 active loans was $10,552 ($12,576 – $2,024), as calculated through September 2014.



                                                       13
           2B.      Determine the overpaid mortgage insurance premium for the 69 active loans23
                    after September 2014 for the life of the loans, and reimburse or apply the
                    overpayments as credits to borrowers’ future premium payments.

           2C.      Determine the number of 203(k) loans24 impacted by the incorrect loan-to-value
                    ratio for mortgage insurance premium calculations and when applicable,
                    reimburse borrowers or apply the overpaid premiums as credits toward borrowers’
                    future premium payments.

           2D.      Issue clarification to lenders regarding the property value that should be used to
                    calculate loan-to-value ratios for determining borrowers’ premiums, which is
                    different from the value used to determine the maximum mortgage amount under
                    the program.

           2E.      Change the loan-to-value ratio calculation in HUD’s System to reflect the issued
                    clarification in recommendation 2D. This correction to the loan-to-value ratio
                    calculation should result in $1.91 million in funds to be put to better use.

           2F.      Update the FHA Connection user manual by providing clear descriptions of and
                    instructions for the data fields to ensure that lenders understand and enter the
                    correct loan data into FHA Connection for computing borrowers’ premiums.




23
     Eighty loans with overcalculated loan-to-value ratios less 11 terminated loans. See appendix G.
24
     Excluding the 150 loans reviewed during the audit.



                                                            14
Scope and Methodology
We performed our audit work from February 2013 through October 2014 at our offices located
in Chicago, IL, Columbus, OH, and Detroit, MI. The audit covered the period December 1,
2008, through July 31, 2013.25 To accomplish our objective, we

        Reviewed relevant background information and applicable HUD handbooks, mortgagee
         letters, the Code of Federal Regulations, the United States Code, FHA’s Post
         Endorsement Technical Review Desk Guide (effective October 2010), the Quality
         Assurance Division Desk Guide, the Computerized Homes Underwriting Management
         System manual (effective November 2011), the FHA Connection message board, and
         HUD’s Web site.
        Communicated with HUD staff, lenders, consultants, and local government authorities as
         applicable.
        Reviewed applicable documentation in the loan files, including but not limited to the
         maximum mortgage worksheet, rehabilitation agreement, 203(k) borrower’s
         acknowledgement, homeowner-contractor agreement, and borrower’s identity-of-interest
         certification as well as the contractors’ licensing, building permits, cost estimates, draw
         requests, appraisal reports, and settlement statements.
        Downloaded and analyzed loan-level data from HUD’s Single Family Data Warehouse.
        Reviewed loan-level data from HUD’s Neighborhood Watch Early Warning System.26
        Selected and reviewed statistical and random samples of loans related to the 203(k)
         program.
        Reviewed Accurint27 information for the selected loans.

Statistical Samples

Lender Compliance With Section 203(k) Program Requirements

For the survey, using HUD’s Single Family Data Warehouse system, we identified 552 loans that
closed between January 1, 2011, and December 31, 2012, and the borrowers of which defaulted
in their mortgage payments within the first 12 months as of January 2013. Using the U.S. Army
Audit Agency Statistical Sampling System software, we selected a statistical random sample of
41 loans, using a 90 percent confidence level with an expected error rate of 20 percent and a


25
   Initially, our audit period covered January 1, 2011, through December 31, 2012. However, due to the change in
the loan-to-value ratio calculation formula for loans with FHA case numbers assigned on and after November 13,
2008, we adjusted our audit period to include 203(k) loans that closed between December 1, 2008, and July 31,
2013, for our loan-to-value ratio calculation testing.
26
   Neighborhood Watch is a Web-based software application that displays loan performance data for lenders using
FHA-insured single-family loan information. The system is designed to highlight exceptions so that potential
problems are readily identifiable.
27
   The Accurint database is an online resource that provides information on legal and public records.



                                                         15
sampling precision of 10 percent. We reviewed the first 11 loans to determine whether an audit
was warranted.

Based on the results and the survey, we expanded our testing universe to include all loans that
closed during our audit scope. Using HUD’s Single Family Data Warehouse system, we
identified 70,196 203(k) loans, valued at more than $10.7 billion that were closed between
December 1, 2008, and July 31, 2013. We selected a statistical sample of 95 across 6 cost strata.
To control for accuracy, we excluded loans with mortgage terms of less than 15 years and 775
loans that exceeded $540,000 in value28 before we arrived at this total. Therefore, we reviewed
106 (11 + 95) loans to determine whether lenders endorsed loans that complied with HUD’s
203(k) requirements, in particular contractor licensing, building permits, and the eligibility of the
work items under the Standard (k) and Streamlined (k) programs.29

HUD’s Review of Section 203(k) Loans

Using HUD’s Single Family Data Warehouse system, we determined that HUD’s Processing and
Underwriting Division reviewed 2,022 loans30 insured under HUD’s Section 203(k) program that
closed from January 1, 2011, through December 31, 2012. Using the U.S. Army Audit Agency
Statistical Sampling System software, we randomly selected 69 of the 2,022 postendorsement
technical reviews to determine whether (1) lenders complied with HUD’s 203(k) requirements
and (2) HUD identified material deficiencies and required appropriate corrective actions, when
applicable, during its review of the 203(k) loans. Our sampling criteria used a 90 percent
confidence level and a sample precision level of 10 percent.

During the survey, we received the listing of 550 loans reviewed by HUD’s Quality Assurance
Division between January 1, 2011, and December 31, 2012. Of these, we excluded 116
terminated loans from our audit universe because they would not increase the risk to FHA
insurance funds. The remaining 434 contained 409 active loans and 25 loans for which FHA had
paid a claim. Of the 409 active loans, we statistically selected a sample of 59 loans with an
initial survey sample of 14 loans for the file reviews. We reviewed only nine of these selected
loans. Additionally, we selected and reviewed the first 8 of the 25 loans for which FHA had paid
a claim to determine whether (1) lenders complied with HUD’s 203(k) requirements and (2)
HUD identified material deficiencies and required appropriate corrective actions, when
applicable, during its review of the 203(k) loans. Therefore, we reviewed 86 (69 + 9 + 8) loans
to determine whether HUD adequately identified and mitigated lenders’ noncompliance with the
program’s requirements.31


28
   We excluded the loans with a mortgage term of less than 15 years and loans that exceeded $540,000 because they
represented less than 1 percent of the universe of 203(k) loans; thus, these loans were considered outliers. In
statistical sampling, an outlier is an element of a data set that distinctly stands out from the rest of the data.
29
   Since the costs were unsupported, we did not project them to the universe.
30
   These loans exclude 53 203(k) loans that had been either referred to or indemnified by HUD’s Quality Assurance
Division as of March 7, 2013.
31
   We reviewed only 9 of the 14 active loans and 8 of the 25 loans that had claim insurance status during the survey
phase to accomplish our objective (17 loans). During the audit phase, we did not review the remaining 22 sampled
loans. The loans reviewed by HUD’s Quality Assurance Division represented less than 1 percent of the 203(k) loans



                                                         16
Loan-to-Value and Mortgage Insurance Premium Reviews

Using HUD’s Single Family Data Warehouse system, we identified 70,196 203(k) loans that
were closed from December 1, 2008, through July 31, 2013, valued at more than $10.7 billion.
To control for accuracy, we excluded 613 loans with mortgage terms of less than 15 years, and
775 loans that exceeded $540,000 in value were excluded as outliers before we arrived at this
total. We selected a statistical sample of 150 loans to determine whether the loan-to-value ratios
were properly calculated for the borrowers’ monthly mortgage insurance premiums.

The sample was designed as a highly stratified systematic sample and to control for variance
resulting from different sizes of loans as well as the impact different ages of loans would have on
the amount of mortgage insurance premium payments incurred by loans. We stratified the
sample by 6 cost groups and 6 age divisions within each cost group for a total of 36 strata. The
details of these strata are as noted in the sample design table below. The cost groupings were
created by ranking loans in order of their unpaid balance and dividing them into six cost tiers by
percentile within this ranking. Age groupings were created according to periods when mortgage
insurance premium amounts were changed by policy and periods when HUD had issued new
guidance to banks on how to enter the information used to compute mortgage insurance premium
amounts. The sample design was stratified as shown in the table below.

                                                     Sample design

                                       Cost group
                                                                Rehabilitation                        Sampling
        Stratum name                                                                 Sample size
                                                                   loans                               weights
                                Rank         Lower bound
      Tier1_0-6 mos.           0-10pct              0.00                    560           2            280.000
      Tier1_07-16 mos.                                                    1,228           3            409.333
      Tier1_17-28 mos.                                                    1,950           4            487.500
      Tier1_29-34 mos.                                                      648           2            324.000
      Tier1_35-48 mos.                                                    1,723           4            430.750
      Tier1_48+ mos.                                                        987           2            493.500
      Tier2_0-6 mos.          10-30pct           67,322                   1,283           3            427.667
      Tier2_07-16 mos.                                                    3,000           6            500.000
      Tier2_17-28 mos.                                                    3,942           8            492.750
      Tier2_29-34 mos.                                                    1,253           3            417.667
      Tier2_35-48 mos.                                                    3,154           7            450.571
      Tier2_48+ mos.                                                      1,558           3            519.333
      Tier3_0-6 mos.          30-50pct           97,886                   1,541           3            513.667
      Tier3_07-16 mos.                                                    3,261           7            465.857
      Tier3_17-28 mos.                                                    3,588           8            448.500
      Tier3_29-34 mos.                                                    1,315           3            438.333
      Tier3_35-48 mos.                                                    3,082           6            513.667




that closed from January 1, 2011, through December 31, 2012. Since the costs were unsupported, we did not project
them to the universe.



                                                           17
                                                Sample design

                                  Cost group
                                                           Rehabilitation                 Sampling
       Stratum name                                                         Sample size
                                                              loans                        weights
                           Rank         Lower bound
     Tier3_48+ mos.                                                 1,404       3          468.000
     Tier4_0-6 mos.       50-70pct         130,602                  1,677       3          559.000
     Tier4_07-16 mos.                                               3,363       7          480.429
     Tier4_17-28 mos.                                               3,500       7          500.000
     Tier4_29-34 mos.                                               1,272       3          424.000
     Tier4_35-48 mos.                                               3,092       6          515.333
     Tier4_48+ mos.                                                 1,286       3          428.667
     Tier5_0-6 mos.       70-90pct         174,202                  1,923       4          480.750
     Tier5_07-16 mos.                                               3,622       8          452.750
     Tier5_17-28 mos.                                               3,247       7          463.857
     Tier5_29-34 mos.                                               1,348       3          449.333
     Tier5_35-48 mos.                                               2,944       6          490.667
     Tier5_48+ mos.                                                 1,106       2          553.000
     Tier6_0-6 mos.       90-100pct        273,100                  1,082       2          541.000
     Tier6_07-16 mos.                                               1,695       3          565.000
     Tier6_17-28 mos.                                               1,329       3          443.000
     Tier6_29-34 mos.                                                 636       2          318.000
     Tier6_35-48 mos.                                               1,209       2          604.500
     Tier6_48+ mos.                                                   388       2          194.000
          Totals            N/A                N/A                 70,196      150          N/A

To design a sample that is sensitive enough to detect this kind of error, we needed to reconstruct
the types of overpayments and underpayments in mortgage insurance premiums that would likely
be found in the audit. Based on a sample of 60 rehabilitation loans that underwent a
postendorsement technical review, we found that the primary driver of mortgage insurance
premium error was an incorrect calculation of the loan-to-value ratio. The loan-to-value ratio
was miscalculated about 72 percent of the time in the postendorsement technical review sample.
When the loan-to-value ratio was miscalculated, the error amount ranged from a slight
underestimate of 14 percent or more to a larger overestimate of 34 percent or more, and the
probability distribution of this error followed a Weibull distribution.

To recreate these typical error amounts, we applied randomly assigned loan-to-value ratio errors
that followed the same Weibull distribution and estimated the effect on the mortgage insurance
premium. We calculated both the cost of errors in monthly payment amounts and the cost of
failing to end the mortgage insurance premium when the true loan-to-value ratio dropped below
the level at which it was required. These calculations included monthly estimates of interest,
principal, and loan balances during the history of the loan through September 2013.

A sample size of 150, as was recommended for this audit, was randomly selected with the
number of samples in each stratum being proportionate to that found in the population, with
minor rounding adjustments as needed to specify whole-number sample counts within each
stratum. The audit sample survey was selected by means of computer routines written in SAS®,



                                                     18
using the survey select procedure and a seed of 7. We arrived at this number by simulating the
performance of various sample designs.

Having established the typical cost impact from a loan-to-value ratio error, we tested the ability
of our stratification design to pick up various rates of error and various sample sizes. We did this
by means of computerized, replicated sampling to reproduce the true precision and reliability that
an audit finding would have with various sample sizes and rates of error. We varied the possible
rate of loan-to-value ratio error from 40 to 80 percent (our test sample had an error rate of 72
percent), and sample sizes ranging from 70 to 150 samples were tested. The resulting audit
findings from these simulations were compared with actual dollar amounts in a given error
scenario to verify how accurate an audit would be using these estimating methods. The
recommended sample size was found to be effective in preventing false errors, and the rate of
accuracy for probabilistic statements made with this sample design was better than the stated
confidence interval.

We found that borrowers for 62 of the 150 rehabilitation loans were billed and paid an incorrect
monthly amount for their mortgage insurance premium (either too much or too little). This
amounted to a weighted average of 41.7 percent of our sample. In projecting the results of our
sample to the universe of 70,196 loans, we can say, with a one-sided confidence interval of 95
percent, that at least 24,762 borrowers, or 35.28 percent of our audit sample, were billed and paid
the wrong monthly amount for their premium.

           Percentage of loans: 41.7% – 1.658 ⨉ 3.87% ≈ 35.28%LCL
           Count of loans:      70,196 * (41.7% – 1.658 ⨉ 3.87%)32 ≈ 24,762LCL

Further, the weighted average for the overpaid premiums was $55.96 per rehabilitation loan. In
deducting for a statistical margin of error, we can say, with a one-sided confidence interval of 95
percent, that the average amount per rehabilitation loan was $41.55. Projecting this amount to
the audit universe of a total of 70,196 rehabilitation loans yields at least $2.9 million in overpaid
premiums. Looking forward 1 year, if this behavior continues, the annualized projected amount
of overpaid premiums for rehabilitation loans will be $1.7 million in funds to be put to better use.

             Sample projection:                                 $55.96 – 1.658 ⨉ $8.69 ≈ $41.55LCL
             Universe projection:                               $3,927,839 – 1.658 ⨉ $609,276 ≈
                                                                $2,910,000LCL
             Monthly projection of overpayments:                $2.55 – 1.658 ⨉ $0.319 ≈ $2.02LCL
             Forward 1 year of overpayments:                    70,196 ⨉	12 ⨉ ($2.55 – 1.658 ⨉
                                                                $0.319) ≈ $1,700,000LCL33


We updated our audit universe to include loans that had closed as of December 31, 2014.
Therefore, the audit universe contained 79,012 active loans totaling more than $12.1 billion. We
found that borrowers for 62 of the 150 rehabilitation loans were billed and paid an incorrect

32
     The percentages were rounded to the nearest hundredths.
33
     This amount was rounded.



                                                           19
monthly amount for their mortgage insurance premium (either too much or too little). This
amounts to a weighted average of 41.7 percent of our sample. Projecting the results of our
sample to the universe of 79,012 loans, we can say, with a one-sided confidence interval of 95
percent, that at least 27,800 borrowers were billed and paid the wrong monthly amount for their
premium, thus yielding more than $3.2 million in overpaid premiums. Looking forward 1 year,
if this behavior continues, the annualized projected amount of overpaid premiums for
rehabilitation loans will be more than $1.9 million in funds to be put to better use.

         Count of loans:                            79,012 * (41.7% – 1.658 ⨉ 3.87%) ≈ 27,800LCL
         Universe projection:                       $4,421,512 – 1.658 ⨉ $686,614 ≈ $3,280,000LCL
         Forward 1 year of overpayments             79,012 ⨉ 12 ⨉ ($2.55 – 1.658 ⨉ $0.319)
                                                    ≈$1,910,000LCL

Computation of Borrower’s Monthly Mortgage Insurance Premium – Explained

According to HUD’s Home Mortgage Insurance Division’s working group, the after-improved
value shown in the appraisal report should be the denominator used to calculate the loan-to-value
ratio for determining a borrower’s monthly mortgage insurance premium.34 Once the loan-to-
value ratio is calculated, the annual mortgage premium rate is determined for the mortgage
insurance premium computation, depending on the following:

        The loan case assignment date,
        Whether the calculated loan-to-value ratio is (1) 95 percent or less or (2) greater than 95
         percent, and
        The type of loan transaction (purchase, full qualifying refinance, streamline refinance).

For instance, for a purchase loan case assigned on or after April 9, 2012,35 the annual mortgage
premium rate would be 1.20 percent if the loan-to-value ratio equaled 95 percent or less;
however, the annual mortgage premium rate would be 1.25 percent if the loan-to-value ratio was
calculated at greater than 95 percent.

Using the Single Family Servicing mortgage calculator in FHA Connection, we recalculated a
borrower’s monthly mortgage insurance premium by entering applicable loan data, including the
loan amount, after-improved value, interest rate, loan term, beginning amortization month and
year, and case assignment date. We also adjusted the annual mortgage premium rate
accordingly, based on our calculation of the loan-to-value ratio. Examples of how we calculated
the overpayment and underpayment of the mortgage insurance premium are below.

Examples for Overcalculated Loan-to-Value Ratios36
For FHA case number 277-1777708, HUD’s System and the Office of Inspector General (OIG)
calculated the loan-to-value ratio at 96.49 percent and 94.43 percent, respectively. Using the

34
   See footnote 21.
35
   For loans with mortgage terms greater than 15 years. See Mortgagee Letter 2012-04.
36
   See appendix G.



                                                        20
prescribed annual mortgage premium rate in Mortgagee Letter 2012-04, OIG calculated the
borrower’s monthly mortgage insurance premium as $331.50 instead of $345.31, using HUD’s
loan-to-value ratio. The borrower’s monthly premium calculated by HUD’s System was $13.81
($345.31 – $331.50) more than OIG’s amount for the first 12 months37 and $13.56 more over the
next 12 months, resulting in the borrower’s overpaying for mortgage insurance from April 2013
through September 201438 by $247 ($13.81 x 12 months + $13.56 x 6 months). Additionally,
with a 96.49 percent loan-to-value ratio, the borrower would be required to pay monthly
premiums for 108 months totaling $34,214. However, using a 94.43 percent loan-to-value ratio,
as calculated by OIG, the borrower would pay the monthly premium for only 99 months totaling
$30,377. Therefore, using the loan-to-value ratio in HUD’s System, the borrower would pay
insurance premiums for an additional 9 months and overpay for mortgage insurance by $3,837.39

For FHA case number 156-1664066, HUD’s System and OIG calculated the loan-to-value ratio
at 96.49 percent and 95.93 percent, respectively. Therefore, with both loan-to-value ratios above
95 percent, the same annual mortgage insurance premium rate of 1.25 percent applies, resulting
in the same monthly mortgage insurance premium of $84.32. However, based on HUD’s
incorrectly calculated loan-to-value ratio of 96.49 percent, the borrower would pay premiums for
2 (118 – 116) additional months because it would take longer for the ratio to decrease to the 78
percent threshold.

Undercalculated Loan-to-Value Ratios40
For FHA case number 156-1541251, HUD’s System and OIG calculated the loan-to-value ratio
at 93.64 percent and 103.01 percent, respectively. Therefore, using the prescribed annual
mortgage premium rate in Mortgagee Letter 2012-04, OIG calculated the borrower’s monthly
mortgage insurance premium as $81.93 instead of $78.66, using HUD’s loan-to-value ratio. The
borrower’s monthly premium was $3.27 ($81.93 – $78.66) less than OIG’s amount for the first
12 months and $3.22 less for the next 12 months, resulting in the borrower’s underpaying for
mortgage insurance from January 2013 through September 2014 by $68 ($3.27 x 12 months +
$3.22 x 9 months). Further, with a 93.64 percent loan-to-value ratio, the borrower would be
required to pay the monthly premiums for 96 months totaling $7,012. However, using the
103.01 percent loan-to-value ratio, as calculated by OIG, the borrower would pay the monthly
premium for 131 months totaling $ 9,605. It would also take longer for the ratio to decrease to
the 78 percent threshold. Therefore, using the loan-to-value ratio in HUD’s System, the
borrower would pay the insurance premiums for fewer months (35 months) and underpay for
mortgage insurance by $2,593.41

For FHA case number 093-6893539, HUD’s System and OIG calculated the loan-to-value ratio
at 96.49 and 100.34 percent, respectively. Therefore, with both loan-to-value ratios above 95

37
   The monthly premium amount was reduced after each 12 months.
38
   Mortgage insurance premium calculations were as of September 2014.
39
   Borrowers will continue to overpay their premiums for at least 5 years and until their loan-to-value ratio reaches
78 percent or is no longer insured by HUD.
40
   See appendix H.
41
   Borrowers will continue to underpay their premiums for at least 5 years and until their loan-to-value ratio reaches
78 percent or is no longer insured by HUD.



                                                           21
percent, the same annual mortgage insurance premium rate of 0.55 percent applies,42 resulting in
the same monthly mortgage insurance premium of $36.57. However, based on HUD’s
incorrectly calculated loan-to-value ratio of 96.49 percent, the borrower would pay premiums for
16 (145 – 129) fewer months because it would take less time for the ratio to decrease to the 78
percent threshold.

We relied in part on information maintained in HUD’s Neighborhood Watch and Single Family
Data Warehouse systems for informational and sampling purposes only. We also relied on data
maintained in the lenders’ systems, such as electronic loan files. Although we did not perform a
detailed assessment of the reliability of the data, we performed a minimal level of testing and
found the data to be adequately reliable for our purposes. The HUD System data for sampled
items were validated by reviewing documents maintained by the lenders. The audit results were
based on our review of electronic and supporting hardcopy documentation maintained by the
lenders and local government authorities.

We conducted the audit in accordance with generally accepted government auditing standards.
Those standards require that we plan and perform the audit to obtain sufficient, appropriate
evidence to provide a reasonable basis for our findings and conclusions based on our audit
objective(s). We believe that the evidence obtained provides a reasonable basis for our findings
and conclusions based on our audit objective.




42
     See Mortgagee Letter 2008-22.



                                                22
Internal Controls
Internal control is a process adopted by those charged with governance and management,
designed to provide reasonable assurance about the achievement of the organization’s mission,
goals, and objectives with regard to

   Effectiveness and efficiency of operations,
   Reliability of financial reporting, and
   Compliance with applicable laws and regulations.

Internal controls comprise the plans, policies, methods, and procedures used to meet the
organization’s mission, goals, and objectives. Internal controls include the processes and
procedures for planning, organizing, directing, and controlling program operations as well as the
systems for measuring, reporting, and monitoring program performance.

Relevant Internal Controls
We determined that the following internal controls were relevant to our audit objective:

   Effectiveness and efficiency of operations – Policies and procedures that management has
    implemented to reasonably ensure that a program meets its objectives.

   Reliability of financial reporting – Policies and procedures that management has
    implemented to reasonably ensure that valid and reliable data are obtained, maintained, and
    fairly disclosed in reports.

   Compliance with applicable laws and regulations – Policies and procedures that management
    has implemented to reasonably ensure that resource use is consistent with laws and
    regulations.

We assessed the relevant controls identified above.

A deficiency in internal control exists when the design or operation of a control does not allow
management or employees, in the normal course of performing their assigned functions, the
reasonable opportunity to prevent, detect, or correct (1) impairments to effectiveness or
efficiency of operations, (2) misstatements in financial or performance information, or (3)
violations of laws and regulations on a timely basis.

Significant Deficiencies
Based on our review, we believe that the following items are significant deficiencies:

   HUD lacked adequate procedures and controls to ensure that lenders required borrowers or
    contractors to obtain building permits to rehabilitate properties. It also did not ensure that



                                                  23
    lenders required that (1) contractors be licensed or certified to perform rehabilitation work or
    (2) contractors’ cost estimates contain detailed descriptions of the proposed repairs to
    determine eligibility for the Streamlined (k) program (see finding 1).

   HUD lacked adequate procedures and controls to ensure that (1) it correctly calculated loan-
    to-value ratios when determining borrowers’ monthly mortgage insurance premiums and (2)
    lenders properly entered borrowers’ loan information into FHA Connection (see finding 2).




                                                  24
Appendixes

Appendix A
           Schedule of Questioned Costs and Funds To Be Put to Better Use
        Recommendation                                             Funds to be put
            number            Ineligible 1/    Unsupported 2/      to better use 3/
              1A                                $792,837
                1B                                 305,395
                1C                                 83,715
                1D                                 39,367
                1E                                 $97,355
                2A             $12,576
                2E                                                  $1,910,000

              Totals           $12,576          $1,318,669          $1,910,000


1/   Ineligible costs are costs charged to a HUD-financed or HUD-insured program or activity
     that the auditor believes are not allowable by law; contract; or Federal, State, or local
     policies or regulations.
2/   Unsupported costs are those costs charged to a HUD-financed or HUD-insured program
     or activity when we cannot determine eligibility at the time of the audit. Unsupported
     costs require a decision by HUD program officials. This decision, in addition to
     obtaining supporting documentation, might involve a legal interpretation or clarification
     of departmental policies and procedures.
3/   Recommendations that funds be put to better use are estimates of amounts that could be
     used more efficiently if an OIG recommendation is implemented. These amounts include
     reductions in outlays, deobligation of funds, withdrawal of interest, costs not incurred by
     implementing recommended improvements, avoidance of unnecessary expenditures
     noted in preaward reviews, and any other savings that are specifically identified. In this
     case, if HUD implements our recommendation 2E, it will reduce the risk of borrowers
     overpaying monthly mortgage insurance premiums under the 203(k) program. Our
     estimate reflects only the initial year of this benefit.




                                              25
Appendix B
             Auditee Comments and OIG’s Evaluation



Ref to OIG    Auditee Comments
Evaluation




Comment 1



Comment 2




Comment 3




                               26
Auditee Comments and OIG’s Evaluation




Ref to OIG             Auditee Comments
Evaluation




Comment 4




Comment 5




Comment 6




                                        27
Ref to OIG   Auditee Comments
Evaluation




Comment 7




Comment 8




Comment 9




                            28
                         OIG Evaluation of Auditee Comments


Comment 1   HUD acknowledged that some procedural and system updates are needed to better
            control risks associated with the 203(k) program. Further, it stated that it agrees
            that clarification of 203(k) related policy was needed, and this was accomplished
            with its recent publication of the new FHA Handbook 4000.1, which becomes
            effective on September 14, 2015.

            We commend HUD for including clarification on 203(k) related policy in its new
            FHA Handbook 4000.1. We reviewed the origination section of the handbook,
            and determined that HUD had addressed recommendations 1K and 2D. However,
            since the new handbook would not be effective until September 14, 2015, these
            recommendations will remain in the audit report and addressed during the
            management decision process.

            Further, HUD Handbook 4000.1 does not fully address all the issues cited in the
            audit report. For example, the handbook does not address how (1) lenders are to
            ensure that building permits are obtained for repairs when required by local
            government authorities and (2) contractors and sub-contractors are licensed.
            Therefore, additional procedural and system updates are still needed.

Comment 2   HUD noted that the language in the draft audit report with respect to FHA’s
            efforts at managing its program risk related to the 203(k) product. Further,
            Housing stated that the very nature and design of the FHA program, with
            underwriting authority delegated to FHA-approved lenders, assumes some
            measure of risk-taking. With finite resources with which to manage program,
            operational, counterparty, and credit risks, Housing faces practical limits in its
            capacity to ensure that lenders comply with 203(k) program requirements in every
            instance.

            We acknowledge (1) that HUD has finite resources and (2) the risks involved in
            administering the 203(k) program. However, as detailed in the audit report,
            HUD’s current controls for monitoring lenders for compliance could be improved
            or enhanced to reduce the risk to the Mutual Mortgage Insurance Fund. We
            reconsidered and revised the language for finding 1 to state that “HUD needs to
            improve its monitoring and oversight of lenders’ compliance with the Section
            203(k) program requirements because lenders did not ensure that (1) borrowers or
            contractors obtained building permits to rehabilitate properties and (2) contractors
            were licensed or certified to perform rehabilitation work.”

Comment 3   HUD stated that it believes that it has controls in place to monitor lender
            compliance with 203(k) program requirements, which had been evidenced to be
            appropriate especially as related to the scope of the Section 203(k) program. It
            also stated that during the period covered by the draft audit report, Housing staff



                                              29
            reviewed approximately 430,000 loans for compliance; almost 2 percent of this
            sample were 203(k) loans.

            We acknowledge the Section 203(k) loan program portfolio is small in
            comparison to FHA’s forward mortgages. However, assessing risks associated
            with the 203(k) program should not be based solely on the size of its portfolio.
            The 203(k) program loan is a two-tier loan designed to finance both the
            acquisition costs and the costs of property improvements into one mortgage loan;
            thus, making it a risker loan than the Section 203(b) mortgage loan. Essentially,
            FHA insures a 203(k) loan before the condition and value of the property offers
            adequate security. According to U.S. General Accounting Office (GAO) report
            GAO/RCED-99-124 issued in June 1999, the Section 203(k) program was more
            risky than the 203(b) single family loan program because the Section 203(k)
            program possesses both the risk of a traditional mortgage and the risk of a
            construction loan. Further, in its May 2013 Insights report, the Office of the
            Comptroller of the Currency (OCC) stated that the construction phase of the
            203(k) program presents the highest risk of loan default.

            Underwriting and approving 203(k) loans involve (1) reviewing borrowers’
            income and credit risks and (2) determining whether repairs are eligible and in
            compliance with the 203(k) rehabilitation requirements. As a result, besides
            ensuring borrowers meet income and credit risk underwriting requirements, HUD
            needs to improve its controls to ensure (1) the repairs to properties are in
            compliance with building codes, (2) borrowers live in safe and sanitary housing;
            and (3) HUD’s interests are properly protected. The review of borrowers’ income
            and credit underwriting analysis was not our audit objective.

Comment 4   HUD stated that the report cites roughly 40 loans with alleged violations of FHA
            policy and notes that almost 85 percent of the violations were related to permits
            not being obtained before the start of construction. However, FHA does not
            require copies of building permits to be included in the endorsement case binder,
            and the absence of a building permit is not sufficient evidence that the borrower
            or the borrower’s contractor did not obtain such documentation.

            The report cited that 55 of the 192 loans reviewed (more than 28 percent) were
            not in compliance with FHA’s policy. We acknowledge that more than 80
            percent of the cited deficiencies were related to licensing and permits and that
            FHA does not require lenders to include the building permits in the endorsement
            case binder. However, as a part of our review procedures, we contacted both the
            lenders and appropriate local government authorities to verify whether building
            permits were obtained for repairs when required. Therefore, we did not solely
            rely on the documents included in the loan files. HUD Handbook 4240, REV-2,
            section 4-9, requires that all licenses and permits be obtained as required by local
            government authorities, and the lender (direct endorsement underwriter) should be
            assured that these requirements have been satisfied and fees had been paid before



                                              30
            draw releases can be given. However, the contractor licenses and building
            permits were not obtained as required for the cited loans.

Comment 5   HUD contended that the requirement to obtain a building permit when necessary
            was that of the borrower or contractor, not the FHA-approved lender. Thus,
            Housing believes that the loans were compliant with FHA requirements.

            Although we agree that HUD specified in the rehabilitation loan agreement form
            that all licenses and permits required by local governmental authorities to
            rehabilitate the property should be obtained by the borrower or contractor, it is the
            lender’s responsibility to ensure that the requirements have been satisfied and the
            fees have been paid before it releases the draw to contractors, as specified in the
            HUD handbook 4240, REV-2, section 4-9.

Comment 6   HUD suggested we revise the recommendation 1I regarding requiring the lender
            to certify to the determination that the work required building permits and that the
            contractors were required to be licensed or certified to perform repairs. HUD
            stated that the lenders’ staffs do not have the knowledge to render a certification
            pertaining to building permit requirements and contractor licensing requirements
            for multitude of jurisdictions throughout the United States. Such requirement
            would likely undermine a lender’s willingness to participate in the 203(k)
            program.

            We acknowledge that there are requirements for the multitude of jurisdictions
            throughout the United States. However, it is the lender’s responsibility to ensure
            that these requirements have been satisfied and the fees have been paid before
            funds are paid to contractors, as specified in the HUD Handbook 4240, REV-2.
            Currently, HUD does not have a requirement concerning how lenders are to
            ensure contractors or borrowers determine whether building permits and licensing
            are required and obtained. Thus, we considered the suggested wording and
            revised recommendation 1I accordingly, recommending that HUD’s policy
            require lenders to maintain documentation supporting their review and
            determination concerning required contractor licensing and building permits.

Comment 7   HUD suggested we revise recommendation 1J regarding the training of borrowers
            that participate in the 203(k) program. As suggested, we revised recommendation
            1J accordingly, removing the reference to training borrowers that was included in
            the draft report.

Comment 8   HUD expressed its confusion about the indication of loan eligibility related to
            “funds put to better use” in appendix A of the draft report. Further, it stated that
            the report and findings do not evidence that any 203(k) loans studied were
            discovered to be ineligible under the program guidelines.




                                               31
            We reconsidered our definition of funds to be put to better use and revised the
            definition in appendix A accordingly, to reflect the future savings in mortgage
            insurance premium dollars to the borrowers who participate in the Section 203(k)
            program.

            Apart from reviewing the selected loans for computation of the borrowers’
            monthly mortgage insurance premiums, our review was generally limited to
            verifying lenders’ compliance with HUD’s 203(k) requirements; specifically
            regarding contractor licensing, building permits, and the eligibility of work items
            under the Standard (k) and Streamlined (k) programs. We did not focus our
            review on the borrower’s income, credit, or funds to close, to determine a loan’s
            eligibility for FHA insurance under the 203(k) program. Additionally, we do not
            provide assurance that no other issues exist with the loans reviewed during the
            audit.

Comment 9   HUD suggested that we revise recommendations 2A and 2C regarding reference
            to adjustments to loan amortization schedules. We agree and revised
            recommendations 2A and 2C accordingly to specifically address borrowers’
            monthly mortgage insurance premiums. Further, for the logical organization of
            the recommendations, we have switched the order of recommendations 2A and
            2B cited in the discussion draft audit report, which are now recommendations 2B
            and 2A, respectively for the final audit report.




                                              32
Appendix C
                                             Criteria

Finding 1
HUD Handbook 4240.4, REV-2, paragraph 3-2(C), states that the improvements must comply
with HUD’s minimum property standards (24 CFR (Code of Federal Regulations) 200.926(d) or
HUD Handbook 4905.1) and all local codes and ordinances. Further, cost estimates must
include labor and materials sufficient to complete the work. Home buyers doing their own work
cannot eliminate the cost estimate for labor because if they cannot complete the work, there must
be sufficient money in the escrow account to hire a subcontractor to do the work.

HUD Handbook 4240.4, REV-2, section 4-9, states that at loan closing, the mortgage proceeds
disbursed by the lender and the cash from the borrower must equal the total cost of acquisition or
refinance. The lender must establish the rehabilitation escrow account and place into the account
the total amount to finance the construction plus the contingency reserve, inspection fees, and
any mortgage payments when applicable. Additionally, the borrower must obtain all licenses
and permits that are required by local governmental authorities. Draw releases cannot be given
until the field office or direct endorsement underwriter is assured that these requirements have
been satisfied and the fees have been paid.

HUD Handbook 4240.4, REV-2, appendix 2, Rehabilitation Loan Agreement, item 8, states that
the borrower should cause all improvements to be made in a workmanlike manner and in
accordance with all applicable statutes and regulations. All licenses, permits, and privileges
required by local governmental authorities to rehabilitate the property should be obtained by the
borrower(s) or contractor.

HUD Handbook 4155.2, paragraph 8.B.2.b, states that the documents in the case binder for the
Section 203(k) program must include but are not limited to the rehabilitation agreement, work
writeups, cost estimates, draw request, 203(k) borrower’s acknowledgment, borrower’s identity-
of-interest certification, homeowner or contractor agreement(s), and 203(k) consultant identity-
of-interest statement.

Mortgagee Letter 2005-50 states that while lenders are not contractors, participation in the
Streamlined (k) program requires that they examine the contractor’s bid(s) and determine that
they fall within the usual and customary range for similar work. Lenders must also ensure that
the selected contractor(s) meet all jurisdictional licensing and bonding requirements. Further, if
“self-help” arrangements are used, the borrower must provide written estimates from the
suppliers of the materials. Those repairs and improvements must comply with any local codes
and ordinances, and the borrower or contractor must obtain all required permits before starting
the work. The cost estimate(s) must clearly state the nature and type of repair and the cost for
completion of the work item. Further, major rehabilitation or major remodeling, such as the
relocation of a load-bearing wall and repair of structural damage, are not eligible for financing
under the Streamlined (k) program.



                                                 33
HUD’s Valuation Protocol Frequently Asked Questions – Environment Protection Agency’s
New Lead-Based Paint Rule states that on April 22, 2010, the Environmental Protection Agency
changed its requirements regarding renovation, repair, and painting for houses built before 1978
as follows:

        Homeowners performing renovation, repair, or painting work on their own home are
         exempt from the rule but are encouraged to learn to perform lead-safe work practices.
        Property owners or landlords who renovate, repair, or prepare surfaces for painting in
         pre-1978 rental housing must be certified and follow lead-safe work practices required by
         the rule.
        Contractors who perform the repair must be certified and must follow specific work
         practices to prevent lead contamination.

HUD’s Post Endorsement Technical Review Desk Guide, dated October 2010, chapter 1, states
that the postendorsement technical review process is one of several FHA processes used to help
monitor and mitigate risk to the FHA insurance fund by conducting technical reviews on a
selection of postendorsement loans to ensure lender compliance with FHA credit and valuation
policies and procedures. These reviews help to identify areas of lender origination
noncompliance, permitting FHA to require corrective actions to mitigate risk, including
indemnification or referral to the Mortgagee Review Board. In addition, chapter 4 states that
reviewing the case file in a thorough and analytical manner is crucial to protecting the FHA
insurance fund. FHA depends on the experience and expertise of underwriters and appraisers to
use their training, experience, and analytical skills when reviewing case files to determine
whether a lender complies with FHA policies and guidelines when underwriting loans to be
insured by FHA. A successful reviewer must have the ability to comprehensively evaluate the
data contained within the entire file and then determine whether the file presents an acceptable
risk to the FHA insurance fund.

Finding 2

Requirements of 12 U.S.C. (United States Code) 1709(c)(2)state that the HUD Secretary must
establish and collect, at the time of insurance, a single premium payment in an amount not
exceeding 343 percent of the amount of the original insured principal obligation of the mortgage.
In the case of a mortgage for which the borrower is a first-time home buyer who completes a
program of counseling with respect to the responsibilities and financial management involved in
home ownership that is approved by the Secretary, the premium payment under this
subparagraph must not exceed 2.75 percent of the amount of the original insured principal
obligation of the mortgage. Upon payment in full of the principal obligation of a mortgage
before the maturity date of the mortgage, the Secretary must refund all of the unearned premium
charges paid on the mortgage under this subparagraph, provided that the borrower refinances the
unpaid principal obligation under this subchapter. In addition to the premium under
subparagraph (A), the Secretary must establish and collect annual premium payments in an

43
  The upfront premium changes throughout the years. Refer to applicable HUD mortgagee letters in this appendix
for applicable changes.



                                                       34
amount not exceeding 0.50 percent of the remaining insured principal balance (excluding the
portion of the remaining balance attributable to the premium collected under subparagraph (A)
and without taking into account delinquent payments or prepayments) for the following periods:

“(i) For any mortgage involving an original principal obligation (excluding any premium
collected under subparagraph (A)) that is less than 90 percent of the appraised value of the
property (as of the date the mortgage is accepted for insurance), for the first 11 years of the
mortgage term. (ii) For any mortgage involving an original principal obligation (excluding any
premium collected under subparagraph (A)) that is greater than or equal to 90 percent of such
value, for the first 30 years of the mortgage term; except that notwithstanding the matter
preceding clause (i), for any mortgage involving an original principal obligation (excluding any
premium collected under subparagraph (A)) that is greater than 95 percent of such value, the
annual premium collected during the 30-year period under this clause should be in an amount not
exceeding 0.55 percent44 of the remaining insured principal balance (excluding the portion of the
remaining balance attributable to the premium collected under subparagraph (A) and without
taking into account delinquent payments or prepayments).”

HUD Handbook 4155.2, paragraph 7.3.c, states that for loans closed on or after January 1, 2001,
with terms of more than 15 years, FHA’s annual mortgage insurance premium is automatically
canceled when the loan-to-value ratio reaches 78 percent, provided the borrower paid the annual
mortgage insurance premium for at least 5 years. For loans with terms of 15 years and less and
loan-to-value ratios 90 percent and greater, FHA’s annual mortgage insurance premium is
automatically canceled when the loan-to-value ratio reaches 78 percent, regardless of the length
of time the borrower has paid the annual mortgage insurance premium. For the mortgages with
terms of 15 years and less and loan-to-value ratios of 89.99 percent and less, FHA does not
charge borrowers annual mortgage insurance premiums.

HUD Handbook 4240.4, REV-2, section 1-10, states that the maximum mortgage calculation is
based on the lesser of (1) the estimate of as-is value or the purchase price of the property before
rehabilitation, whichever is less, plus the estimated cost of rehabilitation and allowable closing
costs or (2) 110 percent of the expected market value of the property upon completion of the
work plus allowable closing costs.

Mortgagee Letter 2008-22 states that effective with new FHA case number assignments on or
after October 1, 2008, FHA will no longer base its mortgage insurance premiums on a
combination of the credit bureau score and loan-to-value ratio. FHA will charge an upfront
premium in an amount equal to 1.75 percent of the mortgage for purchase money mortgages and
full-credit qualifying refinance. In addition to the upfront premium, FHA will charge an annual
premium based on the initial loan-to-value ratio and length of the mortgage. For loans with
mortgage terms greater than 15 years and a loan-to-value ratio less than or equal to 95 percent,
the annual rate is .50 percent; with a loan-to-value ratio greater than 95 percent, the annual rate is
.55 percent. Further, for insurance premium purposes and eligibility for FHA mortgage

44
  See Mortgagee Letters 2010-02, 2010-28, 2011-10, and 2012-04 for applicable upfront mortgage insurance
premium and annual mortgage insurance premium rates.



                                                       35
insurance, the loan-to-value ratio,45 computed to two decimals, is calculated by dividing the
mortgage amount, before adding on an upfront mortgage insurance premium, by the sale price or
appraised value, whichever is less. For refinance transactions, which often include closing costs
in the loan amount, the loan-to-value ratio is determined by dividing the loan amount, before
adding on an upfront mortgage insurance premium, by the appraiser’s estimate of value.

Mortgagee Letter 2010-02 states that effective for FHA loans for which the case number is
assigned on or after April 5, 2010, FHA will collect an upfront mortgage insurance premium of
2.25 percent. The annual premium will not change at this time.

Mortgagee Letter 2010-28 states that effective for FHA loans for which the case number is
assigned on or after October 4, 2010, FHA will lower the upfront premium to 1 percent (from
2.25 percent). In addition, for mortgages involving an original principal obligation of less than
or equal to 95 percent of the appraised value of the property, the amount of the authorized annual
premium is increased to 0.85 percent (from 0.50 percent) of the remaining insured principal
balance. For mortgages involving an original principal obligation that is greater than 95 percent
of the appraised value of the property, the amount of the authorized annual premium is increased
to 0.90 percent (from 0.55 percent) of the remaining insured principal balance.

Mortgagee Letter 2011-10 states that effective for FHA loans for which the case number is
assigned on or after April 18, 2011, with a mortgage term greater than 15 years, for mortgages
involving an original principal obligation of less than or equal to 95 percent of the appraised
value of the property, the amount of the authorized annual premium is increased to 1.10 percent
(from .85 percent) of the remaining insured principal balance. For mortgages involving an
original principal obligation that is greater than 95 percent of the appraised value of the property,
the amount of the authorized annual premium is increased to 1.15 percent (from 0.90 percent) of
the remaining insured principal balance. The upfront premium remains the same (1 percent).

Mortgagee Letter 2012-04 states that effective for FHA loans for which the case number is
assigned on or after April 9, 2012, with mortgage terms greater than 15 years, for mortgages
involving an original principal obligation of less than or equal to 95 percent of the appraised
value of the property, the amount of the authorized annual premium is increased to 1.2 percent
(from 1.10 percent) of the remaining insured principal balance. For mortgages involving an
original principal obligation that is greater than 95 percent of the appraised value of the property,
the amount of the authorized annual premium is increased to 1.25 percent (from 1.15 percent) of
the remaining insured principal balance. In addition, FHA increased the upfront premium from 1
to 1.75 percent of the base loan amount.

Mortgagee Letter 2013-04 states that effective for FHA loans for which the case number is
assigned on or after June 3, 2013, for all mortgages, regardless of their amortization terms, any
mortgage involving an original principal obligation (excluding financed upfront mortgage
insurance premium) less than or equal to 90 percent loan-to-value ratio, the annual mortgage
insurance premium will be assessed until the end of the mortgage term or for the first 11 years of
45
 The instruction for calculating the loan-to-value ratio in HUD’s Mortgagee Letter 2008-22 does not mention
whether it was also applicable to the 203(k) loans.



                                                        36
the mortgage term, whichever occurs first. For any mortgage involving an original principal
obligation (excluding financed upfront mortgage insurance premium) with a loan-to-value ratio
greater than 90 percent, FHA will assess the annual mortgage insurance premium until the end of
the mortgage term or for the first 30 years of the term, whichever occurs first. Additionally, for
case numbers assigned on or after April 1, 2013, it increased the annual mortgage insurance
premium from 1.20 to 1.30 percent if the loan-to-value ratio is less than or equal to 95 percent
and 1.25 to 1.35 percent if loan-to-value ratio is greater than 95 percent.

Computerized Homes Underwriting Management System manual, dated November 2011,
paragraph 4.5.2c, states that for 203(k) loans endorsed on or after November 12, 2008, the ratio
of the mortgage (without upfront mortgage insurance premium) to the lesser of the appraised
value or the sum of the sales price plus repair amount is used for the loan-to-value ratio.

HUD’s Office of General Counsel’s legal opinion clarifying the calculation of the mortgage
insurance premium for 203(k) loans, dated July 29, 2013, states that for purposes of determining
the obligation to pay the mortgage insurance premium, the mortgage insurance premium is
calculated based on the original principal balance. The period for payment of the insurance
premium is established based on the appraised value at the time of endorsement, which may
differ from the value used to determine the maximum insured value for a 203(k) loan. There are
no further calculations necessary for closing costs or repair costs because the calculation of the
mortgage insurance premium is based solely on the appraised value46 of the property as of the
date the mortgage is accepted for insurance.

OIG’s Office of Legal Counsel’s legal opinion regarding the definition of appraised value for
determining borrowers’ monthly mortgage insurance premium, stated that HUD’s Computerized
Homes Underwriting Management System contradicts the statute and the statute controls.
Counsel opined that the appraised value as it pertains to Section 203(k) Rehabilitation Mortgage
Insurance Program is the after-improved value of the property and should be applied consistently
when calculating mortgage premiums.




46
  HUD’s Home Mortgage Insurance Division’s working group further clarified that the after-improved value (the
value according to the appraisal report as determined by the appraiser) should be the only appraised value used for
calculating the loan-to-value ratio for determining borrowers’ monthly mortgage insurance premiums.




                                                          37
Appendix D
                              Schedule of Lender Compliance Deficiencies
                             Permits not
                               obtained           No evidence
                                before          contractor had a         No lead-
           FHA case          construction      specialty license to    based paint       Unclear cost
           number47             started          perform repair        certification      estimate          Other issue
1         137-5861827                                                                         X
2         264-0943855              X
3         061-4391510              X
4         446-0244535              X                                         X
5         221-4986227              X                     X
6         023-3940503              X
7         412-7073745              X                     X
8         251-5012404              X                     X
9         566-0363087              X
10        263-4726983              X                     X
11        461-4837349              X
12        137-6191569              X
13        011-6367805              X
14        277-1529457              X                     X
15        544-0339793              X                     X
16        241-9717127              X
17        341-1255632              X
18        093-6893539              X
19        052-6270908              X
20        197-5774289              X
21        251-4327095              X                     X
22        548-5367429              X                     X
23        566-0021633              X
24        137-7354506              X                     X
25        091-5004081              X                                                           X
26        412-7541464              X48
27        412-7029735              X
28       264-130103949                                   X




47
   FHA case numbers 061-4391510 and 251-4327095 were paid in full as of January 2015.
48
   As a result of local code violations, the required building permits were obtained after the repairs and renovations
were completed. The repairs and renovation associated with issued permits were inspected by the local city
authority. Therefore, we did not request the unsupported cost.
49
   There was no evidence that the contractor had a specialty trade license for the repairs; however, permits were
obtained by the borrowers, and repairs passed the local city inspection. Thus, we did not question the cost.



                                                             38
                     Schedule of Lender Compliance Deficiencies (Concluded)
                            Permits not
                              obtained           No evidence
                               before          contractor had a        No lead-
                            construction      specialty license to   based paint      Unclear cost
      FHA case number          started          perform repair       certification     estimate          Other issue
29     251-466994250                                   X
30      264-0785150                                                        X
31      277-1310097                                                        X
32      044-5206203               X
33     412-701157651              X                    X
34      413-5745825               X                    X
35      352-7227989               X                    X
36      241-9513470               X                                                                          X52
37      263-5007325               X
           Totals                 32                   14                  3                2                 1




50
   There was no evidence to support that the contractor had a specialty license to perform the repairs; however, the
borrower obtained all required permits, and the work passed the local city inspection. Thus, we did not question the
cost.
51
   FHA case number 412-7011576 was paid in full as of January 2015.
52
   The contractor did not sufficiently address all the repairs that were identified on the appraisal.



                                                            39
Appendix E
                                  Schedule of HUD Review Deficiencies
                                                        No
                                  Permits not       evidence
                                    obtained       contractor
                                     before           had a         No lead-      Unclear
                 FHA case         construction      specialty     based paint       cost
                 number53            started         license      certification   estimate     Other issues
        1       023-4597598             X                               X
        2       501-8198149                             X                            X              X54
        3       501-8290589             X                                            X
        4       501-8474260             X                                            X
        5       249-5834018             X               X
        6       137-6358439             X               X
        7       061-4106346             X
        8       251-4486882             X
        9       451-1165810                                                                         X55
                                         56
        10      137-6895748            X                X
        11      042-9383114            X
        12      052-6519334            X
        13      374-6193319            X
        14      277-1438986                                                                         X57
        15      374-6097723                                            X
        16      197-4944463             X
        17      052-4308836                                                          X
        18      043-8239100                                                          X
                   Totals              12               4               2            5               3




53
   FHA case numbers 249-5834018, 042-9383114, 197-4944463, and 374-6097723 were paid in full as of January
2015.
54
   There was no evidence that the repairs indicated in the direct endorsement underwriter form, form HUD-54114,
were adequately addressed.
55
   The cost estimate prepared by the HUD-approved consultant did not separate the costs for labor and materials.
Therefore, we could not determine whether the borrower was reimbursed only for the cost of materials.
56
   Required building permits were obtained after the repairs and renovation were completed as a result of local code
violations. The repairs and renovations associated with the issued permits were inspected by the local city authority.
Therefore, we did not question the unsupported cost.
57
   Appraiser conditioned on the appraisal report that peeling paint in the stairwell needed to be repaired to meet
minimum property standard but was not addressed by the contractor.



                                                            40
     Appendix F
                                 Estimated Losses to HUD From Deficiencies

                    Unpaid                                           Estimated losses
     FHA case      principal     Recommendation    Recommendation    Recommendation     Recommendation    Recommendation
      number       balance58          1A                1B                 1C                1D                1E
1    011-6367805      $120,647            $5,270
2    023-3940503       235,625            31,901
3    023-4597598        84,254            31,780
4    044-5206203       300,689            30,308
5    052-6270908       116,756             7,590
6    052-6519334       223516             10,088
7    061-4106346       102,651            23,344
8    093-6893539        75,519            14,836
9    137-6191569       130,291            26,560
10   137-6358439       120,818            27,435
11   137-7354506       152,256           121,736
12   197-5774289       197,765            14,263
13   221-4986227       255,879            58,108
14   241-9717127       141,091            13,651
15   251-4486882       197,452            21,938
16   251-5012404       276,950            67,085
17   263-4726983        60,409            10,174
18   263-5007325       123,831            32,365
19   264-0943855        45,420             5,346
20   277-1529457       170,491            15,443
21   341-1255632       104,557            10,382
22   352-7227989       146,211            17,890
23   374-6193319       256,068            10,075
24   412-7029735        67,277            11,878
25   412-7073745        44,874            21,485
26   413-5745825        71,799            22,912
27   446-0244535       140,019             9,962
28   461-4837349        67,474            20,695
29   544-0339793        76,638            24,804
30   566-0021633       117,681            33,483
31   566-0363087       115,664             8,400
32   548-5367429      155,100             31,650
33   501-8290589      114,749                              $57,375
34   501-8474260       86,520                               43,260
35   137-5861827      142,280                               71,140
36   091-5004081      100,596                               50,298
37   052-4308836            0                               61,363
38   043-8239100            0                               21,959
39   451-1165810       167,429                                                $83,715
40   501-8198149        78,733                                                                  $39,367
41   241-9513470       105,322                                                                                    $52,661
42   277-1438986       $89,387                                                                                     44,694
      Totals       $5,380,688           $792,837          $305,395            $83,715           $39,367           $97,355



     58
      Unpaid principal balance for active FHA insurance loans downloaded from HUD’s Single Family Data
     Warehouse as of January 2015. HUD paid claim on two loans, FHA case numbers 052-4308836 and 043-8239100.



                                                          41
     Appendix G
                         Schedule of Overcalculated Loan-to-Value Ratio Deficiencies
                                                                       Total
                                                                      amount
                                                                        of        Borrowers    Incorrect data entered into FHA
                          Loan-to-                                   overpaid     would pay              Connection
                         value ratio    Loan-to-     Borrower        premium      premiums
                         calculated    value ratio   overpaid        through         for                             Repair
            FHA case     by HUD’s      calculated    monthly           Sept.      additional               Sale      escrow
             number        System       by OIG       premium           2014        months      Estimate   price59    amount
1          277-1777708     96.49%        94.43%         X                $247         X            X
2          156-1664066     96.49%       95.93%                                0       X
3          352-7829014     96.49%       84.05%          X                  106        X           X
4          105-7187318    100.22%       97.38%                                0       X                                 X
5          137-5665261     94.07%       78.07%                                0       X                     X
6          352-7148959     96.49%       75.07%          X                  255        X
7          501-8500108     97.52%       67.24%          X                    59       X                                 X
8         251-4791488*     97.05%       77.15%          X                  232        X                                 X
9          387-1395159     96.49%       93.73%          X                    57       X           X
10         052-7035272     97.70%       85.20%          X                  167        X
11         332-5379521     96.49%       86.02%          X                  459        X
12         412-7531161     96.49%       88.12%          X                    40       X           X         X
13         264-0943855     84.96%       73.94%                                0       X                     X           X
14         263-4726983     96.49%       91.57%          X                  147        X
15         372-4184540     96.49%       84.08%          X                    95       X           X         X           X
16        422-3176468*     96.49%       96.17%                                0       X
17         292-5995474     96.50%       86.34%          X                  118        X
18        446-0719569*     97.49%       72.75%          X                  108        X                                 X
19         501-8409794    100.27%       75.92%          X                    93       X                                 X
20         446-0654308     95.40%       89.14%          X                  143        X                                 X
21         263-4702841     96.49%       76.46%          X                  189        X
22         105-4680866     96.49%       74.84%          X                  309        X
23         441-8836210     95.84%       87.37%          X                  111        X
24         093-7523939     90.02%       83.55%                                0       X           X         X
25         411-5077139     96.50%       73.93%          X                    72       X           X
26         544-0339793     96.50%       64.09%          X                    64       X           X
     Legend - *FHA case numbers with an asterisk (*) presented loans paid in full as of January 1, 2015.




     59
       For a refinanced loan, lenders would enter a borrower’s unpaid principal balance into FHA Connection rather than
     the property’s sale price.



                                                                42
             Schedule of Overcalculated Loan-to-Value Ratio Deficiencies (Continued)

                                                               Total      Borrowers    Incorrect data entered into FHA
                     Loan-to-                               amount of     would pay              Connection
                    value ratio    Loan-to-     Borrower     overpaid     premiums
                    calculated    value ratio   overpaid    premium          for                               Repair
       FHA case     by HUD’s      calculated    monthly      through      additional                           escrow
        number        System       by OIG       premium     Sept. 2014     months      Estimate   Sale price   amount
27    264-1301039     96.49%       92.93%          X               $70        X            X
28    048-7080310     96.49%        95.92%                            0       X            X          X
29    105-7078213     96.49%       94.37%          X               109        X                                  X
30    201-5269369     97.79%       81.39%          X                 70       X                                  X
31    412-7541464     96.49%       86.27%          X                 83       X           X
32    581-4664948     96.50%       80.49%          X                 81       X           X                      X
33   221-4909890*     96.49%       62.50%          X                 72       X
34    181-2700596     96.49%       87.39%          X               168        X
35    413-5640987     96.49%        95.75%                            0       X
36    387-1318640     96.49%       91.07%          X               107        X
37    492-9113318     97.00%        95.70%                            0       X                                  X
38    566-0363087     96.48%       93.85%          X               193        X
39    221-4557177     88.53%        87.99%                            0       X                       X
40   277-0408863*     99.12%       98.85%                             0       X                                  X
41    374-5556132     97.02%       65.28%          X               279        X                                  X
42    446-0250899     96.49%       89.69%          X               211        X
43    566-0021633     96.47%       82.73%          X               250        X
44    105-4643472     99.90%       88.90%          X               275        X           X
45    501-7905081     96.76%       94.81%          X               315        X           X                      X
46    277-1436224     96.49%        95.70%                            0       X
47    061-4045269     96.86%       88.31%          X               231        X                                  X
48    387-0931290     94.65%        92.55%                            0       X
49    481-3335469     97.13%       93.27%          X               224        X                                  X
50    137-6191569     96.49%       86.29%          X               239        X
51    566-0177026     96.48%       94.22%          X               279        X
52    182-1193066     96.49%       61.93%          X                 53       X           X
53    446-0244535     96.72%       81.04%          X               304        X           X                      X
54    156-0289929     89.91%        84.67%                            0       X                       X          X
55   061-4391510*     96.49%       84.60%          X               117        X           X
56    277-1529457     96.49%       83.46%          X               172        X           X
57    156-0777800     96.86%       71.10%          X               141        X                                  X
58    011-6367805     92.65%       83.00%          X               249        X                       X
59    137-5987962     93.28%        78.82%                            0       X                                  X
60    221-4986227     96.18%       75.47%          X               181        X           X
61    341-1375139     96.49%       82.19%          X               135        X
62    231-1212449    101.15%       94.80%          X               171        X                                  X




                                                           43
                Schedule of Overcalculated Loan-to-Value Ratio Deficiencies (Concluded)
                                                                    Total      Borrowers    Incorrect data entered into FHA
                          Loan-to-                               amount of     would pay              Connection
                         value ratio    Loan-to-     Borrower     overpaid     premiums
                         calculated    value ratio   overpaid    premiums         for                               Repair
         FHA case        by HUD’s      calculated    monthly      through      additional                           escrow
          number           System       by OIG       premium     Sept. 2014     months      Estimate   Sale price   amount
63     251-5052715         96.43%       86.46%          X              $167        X            X
64     352-7504706         96.49%       89.08%          X               285        X            X          X
65     412-7484656         96.49%       92.80%          X               160        X
66     501-8492625         96.49%       72.34%          X               207        X
67     581-4452607*        96.49%       96.07%                            $0       X           X
68     352-7317194        105.94%       94.84%          X               264        X                                   X
69     541-9603091         96.68%       83.38%          X               261        X                                   X
70     043-8233137*        96.49%       94.24%          X               297        X
71     221-4622406*        88.32%       82.86%                            0        X                       X
72     371-4256486         96.67%       90.88%          X               410        X                                   X
73     451-1035298         96.49%       88.05%          X               478        X           X
74     412-6267818        103.81%       97.38%                            0        X                                   X
75     251-5012404         96.49%       94.92%          X               279        X
76     081-1008837*        96.96%       94.65%          X               575        X                                   X
77     411-4915136         96.50%       95.35%                            0        X
78     374-5946100        100.09%       94.45%          X               598        X           X
79     352-6987789*        96.76%       90.99%          X               623        X                                   X
80     277-1492783         96.49%       83.60%          X               122        X           X
           Totals                                       61           $12,576      80          25           11          28
     Legend - *FHA case numbers with an asterisk (*) presented loans paid in full as of January 1, 2015.




                                                                44
 Appendix H
                    Schedule of Undercalculated Loan-to-Value Ratio Deficiencies


                                                                                           Incorrect data entered into FHA
                      Loan-to-                                      Total                            Connection
                     value ratio                                  amount of    Borrower
                         as         Loan-to-     Borrower         underpaid    would pay
                     calculated    value ratio   underpaid        premium      premiums                           Repair
       FHA case      by HUD’s      calculated     monthly          through     for fewer               Sale       escrow
        number         System       by OIG       premium          Sept. 2014    months     Estimate   price60     amount
1     093-6893539      96.49%       100.34%                               $0       X           X        X
2     156-1541251      93.64%       103.01%         X                     68       X           X        X
3     094-6448960      93.63%       103.00%         X                    90       X           X
4     446-1464771      96.49%        97.42%                               0       X           X
5     023-4444095      96.49%       102.45%                               0       X           X
6     544-0538777      96.50%        99.99%                               0       X           X
7     251-4669942     106.15%       108.18%                               0       X           X                     X
8     332-4872169      98.83%       102.13%                               0       X           X
9     372-4544873      96.49%        99.56%                               0       X           X
10    562-2337382      97.14%        97.57%                               0       X           X                     X
11    442-3644053      96.46%       106.12%                               0       X           X          X          X
12    044-4545929      96.49%        97.68%                               0       X           X          X

         Totals                                     2                  $158       12         12          4           3




 60
   For a refinanced loan, lenders would enter a borrower’s unpaid principal balance into FHA Connection rather than
 the property’s sale price.



                                                             45