Audit guidance for mortgage company auditors

The servicer advance detail (subledger) houses the information that will drive the reserve and includes data such as advance balance, type of advance (P&I, T&I and foreclosure related) bifurcated between agency and non-agency buckets, delinquency status, foreclosure status, and whether it is recoverable or non-recoverable. Engagement teams should test the completeness and accuracy of the advance detail.

As it relates to the completeness of the advance detail, a reconciliation of the detail to the servicing system is a common completeness test. However, in practice, due to system limitations or the significant volume of transactions, it could be challenging to maintain a precise advance subledger. Given this, engagement teams should also consider performing additional completeness procedures such as the following:

• Select advances in other audit areas such as loan detail testing and trace the advances through to the advance ledger
• Select from sources such as bank statements that contain advance activity and trace the cash activity to the advance ledger
• Perform a servicer advance roll forward and test cash and payment activity
• Review borrower statements and trace advance details back to the advance population
• Send investor confirmations and reconcile confirmation responses to servicer detail
• In many situations the investors are relying on the detail provided by the servicer, so investor reporting controls are also important to test
• Review P&I and T&I reconciliations as required by regulators
• Test servicing accounts where advance expense activity is recorded to ensure expenses shouldn’t have been capitalized
• Test IT general controls over the systems that produce the reports (servicing system and general ledger)
• Test controls such as servicing-specific controls (payment controls, borrower collection controls, investor reporting controls, which address the completeness of investor remittances) that support the completeness of the information within the servicing system

The completeness of the subledger is important from a reporting standpoint, but also from a valuation perspective, as an understatement of the detail could lead to an understatement of the reserve (i.e., less advances to reserve for).

For the accuracy of the advance detail, the engagement team should determine what detail is ultimately driving the advance loss reserve and subject them to substantive testing. The advance balance is an obvious driver of the reserve and engagement teams should perform procedures over the existence and accuracy of the balance. Engagement teams should also look at agency type, as each agency has different guidelines for recoverability. The Government National Mortgage Association or GNMA (Ginnie Mae) guidelines generally result in smaller recoverable balances, and as such, a GNMA advance reserve could reflect a larger amount of non-recoverable advances whereas a Federal National Mortgage Association or “FNMA” and Federal Home Loan Mortgage Corporation or FHLMC reserve may reflect larger recoveries.

Also, engagement teams should test the type of advance (P&I, T&I, foreclosure/other) as certain advances are typically non-recoverable or have limits on what is recoverable and therefore, such advances would have a larger amount of reserve or would be directly expensed. For agency type, advance type, or other pertinent fields that drive the advance reserve, engagement teams should leverage their existence and accuracy procedures performed over the advance balance. Alternatively, teams can perform a test of details, commonly known as an attribute test, to ensure accuracy. The engagement team should also test controls around how advances and related vendor expenses are coded in the servicing system and the controls around the accuracy of vendor remittances and investor reporting.

The general ledger (GL) typically is a separate system or record for the advance balance. Therefore, the completeness of the advance GL balances is supported through a reconciliation to the advance subledger. If there are discrepancies between the GL and the advance detail, engagement teams should understand the reconciling items and its potential impact on the advance reserve. Through routine audit procedures such as GL completeness, journal entry testing and various other account reconciliations, engagement teams are able to gain comfort over the completeness and accuracy of the general ledger. However, it is important to note that the GL itself (or reconciliation of the subledger to the GL), is not always a sufficient audit procedure to ensure the completeness of the servicer advance detail, as there could be a “garbage in, garbage out” argument if the GL is being populated from faulty information.

Historical claim losses by type are often a key input to the reserve. Engagement teams should perform completeness and accuracy procedures over the claim losses.

• For completeness, engagement teams should reconcile the claim loss data to the GL or perform a roll-forward to the GL. Additionally, tracing known claim losses from other audit procedures directly into the claims detail helps ensure the completeness of the detail.
• For accuracy, engagement teams should perform a detail test over the claim loss and include fields such as balance, agency and type – similar to the procedures performed over the advance subledger detail.

When leveraging subservicers, the engagement team should assess the relevant Statement on Standards for Attestation Engagements “SSAE” 18 System and Organizational Controls “SOC” 1 reports to verify that the controls highlighted above are in place and operating effectively.

The recoverability of advances is often dictated by the agency. Engagement teams must evaluate the criteria used to properly classify or segment advances and determine recoverability. Advances may be 100% recoverable, or there may be a percentage that is not recoverable. The team should challenge management’s assumption on recoverability by examining past performance via servicer data.

The utilization of historical losses to predict losses that currently exist in the advance portfolio is a common input utilized in advance reserve models. Additionally, management typically has an associated lookback period for the losses and may place more weight on losses in current periods as they may be more indicative of the current state, in terms of recoverability guidelines, servicing operations, and borrower performance. Audit teams should examine the lookback period for any potential bias and consider if the data was included or excluded and whether that would change the reserve significantly. For instance, if the period utilized for historical losses is from a period with minimal losses, but the upcoming period is expected to yield higher losses, the auditor should challenge whether the lookback period is appropriate. In addition, a lookback test provides additional support to identify if losses recognized were recoverable or non-recoverable under investor and agency guidelines or whether it is related to operational errors. This lookback test can provide further evidence that management has historically reserved for and bucketed their advances appropriately.

The type of advance and its corresponding recovery rate can vary. Principal is often considered largely recoverable whereas interest is generally not fully recoverable at the current note rate. Also, each agency dictates which advances are recoverable, such as their guidelines for various corporate advances, and as such, agency type can play an important role in the reserve estimate. Accordingly, it is not uncommon for losses on GNMA advances to be higher than losses on other agencies due to the nature and volume of the advance transactions. According to the Congressional Budget Office (CBO), “Ginnie Mae would be exposed to the risk of large losses in a period of severe economic stress that included high default rates on mortgages.” Due to the nature of the GNMA program, GNMA loans are typically on houses with lower valuations to a broader range of homeowners with qualifying credit criteria; therefore, liquidation costs could be higher related to property preservation.

One additional nuance to advance type, relates to whether advances were recognized through the normal course of business or acquired through a business combination. In a business combination, the purchase accounting adjustments related to the fair value of the advances would factor into the required reserve calculation. When advances are acquired, the fair value mark typically represents the expected losses plus the time value of money component related to the fair value discount. As such, segmentation by agency type and the recoverability of advances by type, are important assumptions that engagement teams should subject to their testing procedures.

Status of an advance can relate to whether there is an active MSR, if the loan is paid in full or if the servicing was sold or transferred. The active versus inactive status of an MSR drives a few considerations. If active, the chance of recovering the advance will be higher as the company being audited is still servicing the loan; however, if the company does not have an active MSR, it becomes increasingly more difficult to collect advances since the company’s servicing department is no longer in charge of collection or is actively working the customer. This is also likely the case for loans that have been paid in full. Accordingly, there may be assumptions related to collectability based on the active or inactive status of an MSR and whether the loan has been paid in full.

Another consideration for active MSRs is when a Company elects to fair value their MSRs, where they often build in an advance loss/recoverability rate in the valuation. As such, when reviewing the advance reserve, it is important to test that advances associated with MSRs carried at fair value are not grossing up the advance reserve estimate.