Network  
THC  
   a   newtwork   of  opportunties    
WHOLE LOAN PRICING: Introducing CECL in ALCO Meeting  
THC Asset-Liability Management (ALM) Insight  
Issue 4  
Summary  
Yield measures are an approximate measure of profitability. Value measures have direct applications  
to transactions and provide a holistic methodology to assist in profitably, managing both long term  
interest rate risk and credit risk.  
Loan profitability varies when selling to the agencies because both the G-Spreads and the LLPA are  
only approximate measures of risk. Banks should select loans to sell to the agencies to maximize  
profits.  
CECL is an important component of loan value. For this reason, CECL can be used to facilitate  
transactions between buyers and sellers of whole loans. The analysis clearly shows that the  
combination of typical agency loan pricing and investors’ TBA (MBS) pricing create significantly better  
opportunities for banks to transact whole loans among themselves.  
Introduction  
FASB introduced Current Expected Credit Loss (CECL) on June 16, 2016. American Bankers’ Association (ABA)  
CECL Backgrounder June 2016 remarks: "the new standard represents the most sweeping change to bank  
accounting ever." By introducing the life-of-loan concept to financial accounting, CECL methodology has  
moved the banking industry toward fair value accounting. This is another major step away from book value  
accounting. The credit risk under CECL will be measured in economic value and not the current ALLL measure  
that is almost indifferent to the prevailing market conditions. For these reasons, CECL will be an important  
component of the regulatory reports such as Economic Value of Equity (EVE) report and Net Economic Value  
(NEV) report.  
Incorporating financial accounting to market risk reporting has tremendous implications to asset-liability  
management. The Asset-Liability Committee (ALCO) meetings will need to evaluate the bank’s or credit  
union’s market risk and credit risk exposure as an integral part of the projected earnings discussion; interest  
rate risk and credit risk will be inseparable in reporting earnings. I believe implementing CECL is not a cost,  
but a benefit to banks.  
The purpose of this paper is to explain the benefits of adapting CECL methodology in your ALCO decision  
making processes. In today’s dynamic, but uncertain banking environment, I believe this will clearly help you  
enhance your institution’s capital, Net Interest Income, and loan origination pricing.  
As an example of applying CECL to ALM, this paper will discuss a CECL-based ALM approach, which I refer to  
as the Value Attribution and Yield Attribution Model. I will then apply the model to analyze a sample loan  
portfolio of a bank.  
1
 Network  
THC  
   a   newtwork   of  opportunties    
The value attribution and yield attribution models are derived from a set of established and widely cited  
models including Ho Lee Interest Rate model, Dunsky-Ho Prepayment Default model (also described in  
FHFA Mortgage Analytical Platform) and Ho -Stoll Dealer Bid-Ask Pricing model. Extensive descriptions of  
these models are available in public websites for readers’ reference.  
Value Attribution and Yield Attribution  
Historically, earnings have primarily been evaluated on net interest income using $ value and the margin  
between the loans’ effective yield and the funding rate using %. In the ALCO process, the yield and $ value  
measures are used together to measure revenues. For example, the income of a 30- year fixed rate mortgage  
can be quoted as 3.5% rate with a one-point fee as opposed to a mortgage of 3.6% with no upfront fee.  
Government Sponsored Enterprises (GSE) charge for a G spread in rate (%) while adding the Loan Level Price  
Adjustment (LLPA) in value ($). The conversion between rate and value is expected to be done by institutions,  
even though the conversion is far from simple. Also, while banks use price convention to transact, much of  
ALCO management decisions are based on rates in percentages.  
How should ALCO evaluate the credit charge of loans? If all loans and funding sources are simple bullet  
payments, then the use of yields (or rates) is somewhat tractable. However, most loans are not bullets, and  
converting fees to rates is simple only if community banks and credit unions are willing to make  
approximations. Indeed, capital market participants exploit these approximations to profit in market  
transactions. Credit risk can illustrate the problem in mixing of rate and $ value in ALM. The current ALLL  
approach reports the expected annualized charge off as a rate, even though the balance sheet is $ value-  
based. In fact, CECL is introduced with a life-of-loan principle, in part, to remedy this problem.  
Also, the credit charge should relate directly to profitability. I have discussed many deficiencies using yield as  
profitability measure elsewhere. For example, there are significant problems with the yield measures for  
hybrid ARMS, such as 10-1, 7-1 and others as the project interest rates affect the repricing of these loans and  
hence the yield measure. Also, the prepayments of loans results in mismeasurements of yields. Furthermore,  
yields cannot capture the life-of-loan concept. Economic Value of Equity, a regulatory interest rate risk  
measure, was introduced to capture the long-term interest rate risk to augment the Earnings-at-Risk report  
that uses a short-term horizon. Likewise, CECL is introduced to make sure that the long-term credit risk is  
captured and incorporated in ALM.  
For these reasons, ALCO decisions should use rate and value interchangeably based on robust models. The  
ability to relate value and rate accurately and with one consistent methodology enables institutions to  
originate loans or execute any transactions more profitably. The tools used to present profitability in rates  
or value are yield attribution and value attribution.  
Yield Attribution  
Consider a whole loan based on the THC whole loan  
Yield attribution decomposes a reported yield to  
valuation model using a multi-factor prepayment-default  
model. The rate of a mortgage loan at origination can be  
decomposed as presented below.  
the rates of all the economic components of a  
balance sheet instrument.  
2
 Network  
THC  
   a   newtwork   of  opportunties    
Yield Attribution (%)  
YTM  
time  
value  
option  
spread  
credit  
spread  
clean  
OAS  
3.014  
1.422  
0.069  
0.638  
0.884  
YTM is the yield to maturity or the loan rate at par. The time value is the funding rate. Option spread is the  
additional rate to compensate for the extension risk or prepayment risk of a loan. The credit spread is the  
estimated net loss rate. The clean option-adjusted spread (Clean OAS) is the residual of YTM after netting the  
three cost components. Clean OAS represents the profit released annually to the bank.  
The use of yield attribution for asset-liability management is explained in THC Research.  
Value Attribution  
The value of a loan or portfolio of loans is decomposed into the values of all the economic components.  
Consider the following examples of fixed rate 1-4 family owner-occupied whole loans. The components of  
the attribution are explained below, using the first whole loan, 30-year FRM at 3.875% for illustration. All  
values are based on 100 par and the valuation is based on July 21, 2016.  
If the loan is valued as a bullet loan, like a bond based on a benchmark curve, say the Treasury curve,  
then the loan would be valued 112.348 or a premium of 12.348.  
The bank may sell the servicing fees. The value of the servicing fees is the Mortgage Servicing Right  
(MSR). The value is 1.429  
The loan CECL is 2.316 which has to be deducted from the premium to isolate credit risks from  
profitability.  
Based on the option-adjusted spread (OAS) estimated from the TBA MBS market, the investors’  
required profit of an equivalent TBA pool in the capital market is 4.858.  
The premium net of all the cost components is the loan excess profit. The loan excess profit is the profit additional  
to that required by the capital market, and in this case, the excess profit is 3.744 (= 12.348 1.429 2.316 4.858).  
Loan excess profits enable ALCO to understand a relative valuation or profitability of the loans.  
3
 Network  
THC  
   a   newtwork   of  opportunties    
Table A. Value Attribution of a Loan from a Bank’s Perspective in Loan Pricing  
Price  Attribution from a Bank's Perspective  
Profit when  
priced as  
Bank's  
excess profit  
to TBA  
total  
premium  
Loan  
Term  
MSR  
CECL  
Note Date Loan Rate  
TBA  
30  
30  
30  
15  
11/03/15  
11/05/15  
11/17/15  
11/23/15  
3.875% 1.429  
3.875% 1.429  
2.316  
4.858  
3.744  
3.744  
4.460  
3.348  
12.348  
2.316  
2.024  
0.894  
4.858  
4.123  
1.276  
12.348  
11.819  
6.406  
4.125% 1.213  
3.250% 0.888  
The loan excess profit can be used to determine the  
relative profitability of each loan, including loans to be  
sold to the agencies.  
The value attribution table enables ALCO to make  
an apple to apple comparison when evaluating  
the merits and demerits of each loan.  
Whole Loan Transaction: an Application on Value Attribution  
Value attribution is particularly important in any transaction, and valuation is often established through the  
arbitrage process of the capital market. Capital market participants seek to buy low and sell high almost  
instantaneously. I will first describe the GSE arbitrage pricing process.  
Government Sponsored Enterprise (Fannie and Freddie Agencies) Loan Pricing  
These agencies buy loans from loan originators and sell them as mortgage-backed securities. Therefore, the  
agencies’ loan pricing must take the capital market’s required return into consideration when determining  
the loan bid price to the loan seller. Further, the agencies must charge the loan seller for the following items:  
Guarantees for the credit risk of the loans. The agencies provide an implicit guarantee against any  
credit loss for the investors of mortgage-backed securities  
Capital charge. Even though the capital ratios are very low at the agencies, the capital charge is  
significant. Federal regulation restricts the US Government from infusing capital to the agencies.  
However, a line of credit is provided to support the operations of the agencies, resulting in an  
additional capital charge.  
Operating costs. Operating costs include transaction fees, interest rate risk and liquidity risk  
management and required profits to the shareholders.  
These charges are G-Spreads in % and LLPA fees in $ combined. Based on these costs, the agencies provide  
whole loan sellers non-negotiable bid prices. The Value Attribution Table B below illustrates the arbitrage  
between the bid price for sellers and the ask price for the capital market investors.  
4
 Network  
THC  
   a   newtwork   of  opportunties    
Table B: The value attribution on arbitrage pricing  
Price Attribution for Agencies' Price for Loan Sellers  
Investment  
Profit  
total  
premium  
G-spread  
loan profit  
Loan Term  
Note Date Loan Rate  
30  
30  
30  
15  
11/03/15  
11/05/15  
11/17/15  
11/23/15  
3.875%  
3.875%  
4.125%  
3.250%  
6.117  
6.293  
3.899  
3.138  
4.858  
4.858  
4.123  
1.276  
1.372  
1.197  
3.798  
1.992  
12.348  
12.348  
11.819  
6.406  
Arbitrage Pricing The loan profit ensures the loan originators supply the loans. The investment profit ensures  
that the securities can be sold to the capital market. The G-spread pays for the cost of risk transform and  
payments to the stakeholders.  
Application to Whole Loan Transactions  
Value Attribution can be used profitably for whole loan transactions. While whole loan transactions have  
gathered much attention among banks and credit unions in today’s market environment, the crucial  
impediment for buyers and sellers is to agree on a transaction price. Price discovery is always an important  
aspect in market formation. Extensive academic literature has been devoted to the subject called  
“Microstructure Theory” where I was one of the early contributors to the research in co-authoring the Ho-  
Stoll model that studies the bid ask quotes of dealers. By studying formation of bid and ask prices, instead of  
the transaction price, the model can provide a useful tool in submitting bid or ask prices to the market or to  
counter-parties.  
The bid and ask quotes must, in turn, be determined by the reservation bid and ask prices. The reservation  
bid price of the buyer is the highest price that the buyer is willing to pay. The reservation ask price of the  
seller is the lowest price that the seller is willing to sell. A necessary condition for any transaction to occur is  
that the seller’s reservation ask price has to be lower than the buyer’s reservation bid price.  
The Trade Margin is defined as the reservation bid price net of the reservation ask price. When the trade margin  
is positive, buyer and seller can find a transaction price that both sides can profit from. Otherwise, a transaction  
is not possible.  
Reservation Bid Price  
Trade Margin  
Buyer’s highest bid price  
Buyer and seller can find a  
transaction price in this region  
that both side can make a profit  
Reservation Ask Price  
Seller’s lowest ask price  
5
 Network  
THC  
   a   newtwork   of  opportunties    
Determining the Trade Margin and the Fair Transaction Price  
I will use the Value Attribution analysis of the sample fixed rate mortgage loans to illustrate the determination  
of the reservation bid and ask prices of the seller’s servicing retained loans. The figure below depicts the  
formation of these prices. The seller’s calculation and the buyer’s calculation are on the left and right  
respectively. Since the agency’s bid price determines the seller reservation price, I label the bar “Agency.”  
Consider the value attribution of an average of a sample of fixed rate mortgages from an institution’s loan  
portfolio from the buyer’s and seller’s perspectives in Table C below.  
Table C a. Value Attribution of the Loan from the Buyer Perspective  
Price Attribution from a Bank's Perspective  
Loan Term  
Note Date  
Loan Rate  
Investment Loan Excess  
Profit Profit  
Total  
Premium  
MSR  
1.429  
CECL  
2.316  
30  
11/03/15  
3.875%  
4.858  
3.744  
12.348  
Table C b. Value Attribution of the Loan from the Seller Perspective  
Price Attribution  for Agencies Price for Loan Seller  
Loan Term  
Note Date  
Loan Rate  
G-spread  
Investment Profit  
Loan Profit  
1.372  
Total Premium  
30  
11/03/15  
3.875%  
6.117  
4.858  
12.348  
Table C c. Comparing the Reservation Prices  
Reservation prices are the loan excess profit plus par.  
Total  
Premium MSR  
Investment  
Reservation  
Prices  
CECL G-spread  
Profit  
Loan Excess Profit  
Average  
(1)  
(2)  
-
(3)  
-
(4)  
6.117  
-
(5)  
(6)=(1)-(2)-(3)-(4)-(5)  
(7)=(6)+100  
101.372  
Bank (seller)  
12.348  
4.858  
4.858  
1.372  
3.744  
Bank (buyer) 12.348  
1.429  
2.316  
103.744  
As explained above and referring to the figure below, starting from the premium (a), after deducting the  
investor’s required profit and the G- Spread, the agency would be typically quoting a premium to buy from  
the bank. For this particular example, using all the information of this particular loan, THC model calculates  
the agency bid price to the loan originator to be 101.372 (= 100 + net premium). In this case, the Net premium  
= Total Premium Investment Profit- G-spread (12.348 -4.858 -6.117) = 1.372, which I also call the loan excess  
profit. Therefore, the loan seller would not sell the loan for less than 101.372 to another investor because  
the seller can always sell to the agencies. Hence, 101.372, in this case, is the seller’s reservation ask price.  
Now I consider the buyer’s calculation. Often a secondary market loan purchase is viewed conceptually as an  
“investment” in the ALCO meeting. A whole loan purchase is basically an investment, except that the  
6
 Network  
THC  
   a   newtwork   of  opportunties    
purchase goes through the bank’s loan origination process and the transaction is not regulated by FINRA as  
a securities transaction. For this reason, an alternative to buying whole loans for the buyer would be  
purchasing TBA (MBS) securities. Referring to the figure below, starting from the premium (a), the buyers  
would deduct the “investor’s profit” because the buyer would need to lower the price to receive the profits  
if he/she would buy a TBA with the same underlying mortgage loans. The buyer then marks down the bid  
price for CECL. Since the loan is seller servicing retained and will not receive the fee income, the loan again  
is marked down for the mortgage servicing right (MSR) value. The reservation bid price is therefore 103.744.  
Reservation Bid Price is = (Par +Total Premium – Investor’s Profit – CECL MSR) or (102.348 4.858- 2.316 -  
1.429) = 103.744  
These reservation prices, of course, are determined based on a set of  
assumptions. The buyer may need to adjust the MSR and CECL based  
on the ALCO’s estimates and not the THC model. Likewise, the seller  
may also adjust the ask price. However, since the THC mortgage loan  
The Trade Margin is established to be  
any transaction price between  
101.372 and 103.744.  
valuation model is based on objective industry data, both the buyer and seller can use the THC model  
calculated reservation prices as benchmarks to set their respective prices.  
Given the objectivity of the THC model, a fair transaction price can be the mean of the two reservation prices  
without being seller or buyer biased. In this example, therefore the fair transaction price is 102.558. The  
Trade Margin is 2.372 (= 103.744 101.372).  
I have calculated the indicative Trade Margin recognizing that there are other issues that have been ignored  
for clarity of exposition. For example, the mark up and mark down of the Agency pricing and other factors.  
On average, I believe, these effects do not affect the estimated Trade Margin in a consequential way.  
Conclusions  
This paper introduces the yield attribution and valuation attribution based on CECL as an analytical tool to  
convert the rate measure to and from the fair value measure. These tools can be used in many ALM decisions.  
This paper describes on application: whole loan transactions.  
Current Expected Credit Loss (CECL) will have significant impact on financial institutions’ asset-liability  
management. While the life-of-loan concept ensures economic valuation is used for financial reporting, I  
believe the impact of CECL will have broader impact on ALM, including loan origination, pricing, purchases,  
and sales. This subject I will discuss in the next THC ALM Insight.  
I welcome your comments.  
Regards,  
Tom Ho PhD  
President  
1-212-732-2878  
7
 Network  
THC  
   a   newtwork   of  opportunties    
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  
8