Network  
THC  
   a   newtwork   of  opportunties    
 Where Asset Liability Management and Transactions Meet  
WHOLE LOAN SALE TO AGENCIES: A Strategy  
key words: risk capacity, G-spread, LLPA, yield attribution, fixed rate 1-4 family mortgage, whole loan pricing  
THC Asset-Liability Management (ALM) Insight  
Issue 5  
1-4 family residential mortgage originations have picked up in 2018. To facilitate the increase in demand, financial  
institutions can (1) keep more loans on the balance sheet, (2) sell conforming loans to the Agencies or (3) sell loans as  
servicing retained or released to other community banks or credit unions. How do you determine your best option? Our  
Insight #5 explains.  
In previous issues, I have discussed the use of duration and the liquidity coverage ratio to measure your Risk Capacity,  
which is then used to enhance profitability by applying the Fund Transfer Pricing (FTP) approach. As FTP shows, the  
profitability is driven primarily by your loan volume and pricing. My last Insight issue outlines a loan pricing strategy  
using CECL. This issue continues this discussion by using both on and off-balance sheet strategies.  
Overview  
Must you or should you sell loans to Fannie and Freddie?  
Our bank clients often sell whole loans to Fannie Mae (Federal National Mortgage Association) and Freddie Mac  
(Federal Home Loan Mortgage Corp). These agencies have fulfilled their role in “disintermediatingthe residential loan  
market, bringing individual borrowers to the lenders in capital markets via securitization in structuring MBS and CMOs,  
bypassing the traditional rules of banking. Disintermediation was welcomed by academics in the 90’s believing that  
allowing borrowers transact directly with the investors in the capital market would provide liquidity and efficiency to  
all. In the subsequent two decades, disintermediation has become a significant disruptor to the residential loan market.  
The chart below depicts the significance of government initiatives in disintermediation, the trends of First Lien Purchase  
finance by alternative channels, and illustrates that the Government Sponsored Enterprise (GSE) and FHA/VA  
securitization dominates the volume of originations. The private label volume is negligible and the volume retained on  
balance sheet is only 21% in 2017 Q1.  
1
 Network  
THC  
   a   newtwork   of  opportunties    
 Where Asset Liability Management and Transactions Meet  
In the Post 2009 financial crisis era, regulators are rethinking the implications of disintermediation. They  
often consider regulating the agencies or adjusting the G-spread, a haircut on the whole loan interest rate  
paid to the agencies to cover the credit guarantees and servicing cost. But disintermediation has a broader  
impact on banking. Our bank clients also know the importance of community banking in the lending market;  
borrowers simply cannot be pigeon holed into conforming loans. Further, disintermediation can also change  
the culture of banking. With disintermediation, many banks are acting as brokers: completing loan  
applications subject to agency underwriting guidelines and selling to the Agencies electronically. Brokering  
loans was favorable prior to the financial crisis when new loan volumes were strong. Today, however,  
origination volume is low and many banks want to keep loans on the balance sheet. What loans should banks  
sell to the agencies? Are there better alternatives? Banks need to rethink the benefits of selling conforming  
loans to the Agencies. With the growing economy, the Whole Loan Sale to Agencies Strategy may offer a  
solution in the current environment.  
Whole Loan Sale to Agencies Strategy  a Description  
The agencies in disintermediating the residential mortgage market provide price transparency to investors  
and borrowers. The basic pricing method is described below. Whole loan pools are sold as TBAs to the capital  
market to determine the market price. Whole loan prices are based on the market price of investments with  
a particular coupon rate, taking the G-spread into consideration. Refer to THC Research article CECL ALM  
Approach (2016) for details. In addition, Fannie Mae charges for the credit component based on a Loan Level  
Price Adjustment (LLPA) matrix that takes LTV and FICO into consideration. This procedure determines the  
Fannie Mae (FN) purchase price. The pricing model also considers mark up, mark down, and other factors  
when determining value. For clarity, this document will not cover those issues other than FICO and LTV  
considerations.  
The LLPA table is quite simplistic. For example, the same credit charge on the price, not on the yield, is  
constant for any maturity over 15 years. Further, the LLPA charge for low FICO does not take other pricing  
factors into consideration, such as lower FICO score borrowers tend to have a lower prepayment speed.  
Given this agency pricing process, THC can determine the whole loan rate for a TBA coupon rate and  
determine the loan values. Our research then compares the Fannie Mae (FN) offer prices to the  
corresponding THC model prices. THC default-prepayment model has over 30 parameters to capture the  
default risk, the borrowers’ prepayment behavior, and their combined effect. Below are the results of the  
comparison.  
2
 Network  
THC  
   a   newtwork   of  opportunties    
 Where Asset Liability Management and Transactions Meet  
Comparing Agency Offer Rates and THC Offer Rates for 1-4 Family FRM Conforming Loans  
6.500  
6.000  
5.500  
5.000  
4.500  
4.000  
3.500  
3.000  
THC Rates  
FN rates  
Conforming FRM types  
Implications of the Comparison Results  
The chart above compares the origination loan rates based on the THC model and the loan rate based on the  
FN pricing with LLPA. The orange line represents the FN pricing of whole loans over a range of FICO Scores  
and LTVs for conforming 10, 15 and 30- year loans. The results show that the LLPA credit charge for LTV is  
significant for the 10-year FRM and decreases gradually with the term of the loan. This result comes from the  
deficiency of using price adjustment to charge for credit risk exposure instead of using basis points on  
spreads. Also, the LLPA risk charge remains the same for all maturities and that deficiency comes from the  
LLPA matrix that assumes a constant credit charge on price for any term over 15 years.  
By way of comparison, the blue line depicts THC loan pricing following capital market convention. The results  
show that the THC rates are lower than FN prices for shorter term loans and are higher for longer term loans  
to compensate for credit risk and higher LTVs.  
Note that the higher rate results in a lower offer price. In comparing THC prices with Agency prices, the bank  
should prefer keeping the loans on their balance sheet when the THC price is lower. The results can also assist  
3
 Network  
THC  
   a   newtwork   of  opportunties    
 Where Asset Liability Management and Transactions Meet  
ALCO in determining which conforming loans should be sold to the Agencies based on the borrowers’ FICO  
scores and LTV ratios. I know some banks are already using this optimization strategy. For example, banks  
tend to sell conforming 30-year FRM with relatively high LTV’s and low FICO scores.  
The Whole Loan Sale to Agencies Strategy suggests that banks can determine the relative pricing of loans  
between the capital market and the agency. Furthermore, using the bifurcating loan strategy, conforming  
loans can be constructed such that the 1st mortgage can attain the highest profits in selling to the agency.  
THC Research CECL ALM Approach (2016) and Insight #4 provide a more detailed analysis on the transaction  
pricing analysis.  
Why a Strategy for Whole Loan Sale?  
Access Loan Market Liquidity  
Banks can gain access to the loan market. Loan markets offer multiple options for banks to optimize whole  
loan balance sheet strategizes:  
Selling loans to the agencies  
Keeping loans on the balance sheet  
Selling loans in the loan market  
Originating two loans to a single borrower using any of the above options to maximize earnings.  
ALM Analysis and Reporting  
The above optimization depends on multiple factors: the sale price, the impact on the balance sheet risk  
profile, and member needs. You can use a systematic approach to making these decisions using pre-  
transaction simulations.  
Impose minimal operational changes  
Many bank CFOs, CLOs, and legal counsels have worked together in establishing guidelines for selling whole  
loans to Agencies. To implement, the bank would just have to review the loan analytics when deciding the  
optimal portfolio of loans to sell to the Agencies or utilize an alternative option, as listed above. This strategy  
is consistent with the processes many banks have in place already.  
Flexibility of the Strategy to adapt to banks’ preferences.  
Banks can adjust the Strategy in Loan Sale to meet the bank’s risk culture. This is discussed in further detail  
in Insight #1 by simply considering the Agencies as a potential investor, rather than the sole source.  
Increase Loan Product offerings  
The 30-year conforming Fixed Rate Mortgage type dominates the residential market in part because of the  
Agencies’ pricing of the long dated high credit quality loans. By monitoring the Agencies’ residential mortgage  
loan pricing, banks can offer a broader range of loan products where the Agencies prices are less competitive.  
4
 Network  
THC  
   a   newtwork   of  opportunties    
 Where Asset Liability Management and Transactions Meet  
Conclusions  
Much of banking has changed since the financial crisis. With the increase liquidity in the loan market,  
enhanced transparency and standardization of risk measures of residential whole loans, banks and credit  
unions can use multiple channels to off load loans from their balance sheet to increase profitability and  
manage the balance sheet risk. The Agencies need not be the sole investors of your loans. And you can be  
flexible to meeting customers’ borrowing needs.  
I welcome your comments.  
Regards,  
Tom Ho PhD  
President  
1-212-732-2878  
About THC  
THC is a financial technology company founded by Dr. Thomas Ho, a former professor at New York  
University, who introduced the first balance sheet valuation (Ho-Lee model 1986) called "option model" by  
regulators and key rate durations (1992), one of the most popular interest rate risk measures.  
THC was selected as the sole provider of the risk reporting to all regulated institutions under a federal bank  
regulator. THC continues to dedicate its research and resources to supporting community banks.  
THE THC CONTENT IS PROVIDED AS IS, WITHOUT REPRESENTATIONS OR WARRANTIES OF ANY KIND. TO THE MAXIMUM EXTENT PERMISSIBLE  
UNDER APPLICABLE LAW THC HEREBY DISCLAIMS ANY AND ALL WARRANTIES, EXPRESS AND IMPLIED, RELATING TO THE THC CONTENT, AND  
NEITHER THC NOR ANY OF ITS AFFILIATES SHALL IN ANY EVENT BE LIABLE FOR ANY DAMAGES OF ANY NATURE WHATSOEVER, INCLUDING, BUT  
NOT LIMITED TO, DIRECT, INDIRECT, CONSEQUENTIAL, SPECIAL AND PUNITIVE DAMAGES, LOSS OF PROFITS AND TRADING LOSSES, RESULTING  
FROM ANY PERSON’S USE OR RELIANCE UPON, OR INABILITY TO USE, ANY THC CONTENT, EVEN IF THC IS ADVISED OF THE POSSIBILITY OF SUCH  
DAMAGES OR IF SUCH DAMAGES WERE FORESEEABLE  
5