These are my notes on:
- The concepts behind underwriting loans.
- Loans that will likely end in disaster.
- Finding excesses in lending.
- Areas which might lead to unusually high shareholder returns.
The key skill in underwriting loans is predicting the difference between the revenues and expenses on a loan when the loan is made. Prudent lenders will take into account how (in)accurate their predictions are. Prudent lenders will give themselves a suitable margin of safety when determining how much leverage to use.
There is always some uncertainty as to how well a loan will perform. Underwriting is a challenging and open-ended problem, especially for the riskiest loans.
Loan performance is affected by macroeconomic factors: unemployment, fluctuations in the market value of the collateral, etc. The difficult in forecasting future macroeconomic trends is one reason why loan underwriting is difficult.
In some cases underwriting can be unintuitive. For mortgages, rising home prices will lower loan losses. Loans originated in the past will be more profitable than they otherwise would be. However, after home prices have risen, new loans at the same loan-to-value ratios will be riskier. Home prices have more room to fall and therefore loan losses are potentially greater. Lenders may wish to tighten lending standards to maintain similar loan quality. In practice, the positive reinforcement from past profits can be difficult to fight. Past profits tend to attract new entrants which leads to increased competition. In response to this competition, some lenders will price more aggressively to maintain or increase loan volumes. While some lenders may be tightening their standards, it is the loosest lenders that will have the most economic impact. Their aggressive lending increases the number of buyers who can finance their real estate purchases. This leads to increased demand for real estate which leads to higher real estate prices which leads to better loan performance (*ignoring refinancings). Historically, these self-reinforcing behaviours tend to get out of control. The end result is a boom-bust cycle. Auto lending also tends to exhibit boom-bust cycles due to changes in the availability of capital to lenders.
Many loans are made where the originator is not the same as the lender. The originator will be motivated to find the lender with the lowest effective price. They will shop around to find the lender which provides the cheapest financing. Lenders which have made underwriting errors are more likely to have the lowest price and are more likely to be chosen. Underwriting mistakes can get magnified by brokers looking for the lowest pricing.
Similarly, changes in the competitive environment can cause adverse selection to be more problematic. A lender that prices more aggressively on underpriced risks (but not overpriced risks) will reduce the quality of competitors’ loans. Conversely, each lender is affected by what their competitors are doing.
Typical loan fraud involves a borrower with overstated income. A borrower may decide to lie about his/her income. A broker or loan officer may coach and/or encourage the borrower to overstate their income. One way lenders protect themselves against this fraud is to verify income. An underwriter may compare the job description to typical incomes for that job (e.g. by searching the Internet) to see if the income is plausible. The lender might also call the employer to verify that the borrower is actually employed. Most lenders will ask for proof of income such as pay stubs, though some people may try to forge these documents.
Sometimes brokers or loan originators will overstate income without the borrower’s knowledge. Some lenders will call the borrower as a safeguard against this practice.
Intricacies in underwriting
Most lenders look at many different variables to try to predict a borrower’s creditworthiness:
- Credit score
- Credit history- previous bankruptcies, repossessions, lack of a credit history, etc.
- Stability of employment – time at current job, self employed or not, etc.
- (Auto loans) Stability of residence – time at current residence
- (Auto loans, homes) Cost of servicing the debt relative to income, including the all-in costs of maintaining and using the asset.
- Debt relative to disposable income. An auto lender may look at a borrower’s disposable income after rent/mortgage is paid.
- etc. etc.
Lenders also typically consider the collateral:
- If the collateral backing the loan is seized, what will the proceeds be? A home lender may appraise the property and/or perform an inspection to get a more accurate reading on the current value of the house, which will affect foreclosure proceeds and the borrower’s willingness to keep paying the mortgage.
- For auto loans, a lender may (or may not) look at the age, make and model of the car to determine reliability and the cost of repairs. Generally all auto lenders look at the odometer reading.
- For pawn loans, pawn shop employees will use their experience to appraise an item. Pawn shop employees also typically check eBay for items that are difficult to appraise (in a back room where customers cannot see them checking eBay). Some pawn shop employees love their jobs because they enjoy constantly learning about new things.
There are many, many little areas in which a lender may look at to try to price risk more accurately or to avoid risks that are difficult (for them) to quantify. For example, Home Capital Group’s operating subsidiary has an esoteric policy against homes with urea formaldehyde foam insulation (UFFI), even if UFFI was previously removed from the home.
At the origination stage, lenders may ask for information that will help them collect on defaulted loans:
- Many subprime auto lenders will ask for references. Some lenders require different addresses. This helps the auto loan servicer to find the car when the borrower has moved and stopped paying (skip tracing).
- Bank account information. Lenders can go to court and freeze bank accounts to collect on a defaulted loan.
The underwriting of a loan is tied to the servicing of a loan as the servicing affects loan losses. Ultimately, underwriting is a very open-ended problem. It’s not as simple as looking at a credit score.
Skin in the game
Lenders whose incentives are affected by loan quality tend to originate higher quality loans.
However, the US housing bubble demonstrated that this does not guarantee prudent lending. Some of the companies that made terrible loans had significant insider ownership (e.g. Golden West Financial). My theory is that some people are simply crazy and that these crazy people have an unusually large financial impact.
In rare cases, a lender may lend in a way that will likely end in disaster. Some loan structures will show deceptively low delinquencies in the beginning and show massive losses later on. With homes, an option-ARM mortgage can have very low payments initially during the initial “teaser” period before resetting to a much higher level. Worse still, many lenders made these loans to borrowers without verifying income. They basically invited mortgage fraud in the quest to increase loan volumes.
After a loan is made, it is possible for the lender to try to hide loan delinquencies. Golden West Financial, a portfolio lender, would modify its loans so that the company would not have to show a delinquency. Nowadays, the excesses of American lenders have gone down:
- Minimum downpayments have largely increased from 0% to 3.5%/5%. 0% downpayments still exist for VA and USDA loans. However, most borrowers/homes do not qualify for a VA or USDA mortgage.
- Portfolio lenders (lenders who hold onto the loans they originate) no longer make loans that allow for the principal balance to grow over time (negative amortization). Nowadays, portfolio lenders offer “interest-only” mortgages where the principal balance stays flat for the first 3/5/7 years of the loan. I think that interest-only loans are still potentially toxic because they attract irresponsible buyers.
- The loosest lenders will perform some income verification work rather than none.
- Fannie Mae no longer offers 40-year mortgages. Fannie has gone back to offering fixed-rate mortgages that are “only” 30 years without any prepayment penalties.
Personally, I think that the US is a bizarre place. Government tax breaks on mortgage payments and mortgage subsidies encourage homedebtorship rather than homeownership. Encouraging Americans to use their homes as ATMs does not strike me as socially beneficial (e.g. Freddie Mac’s cash-out refinance mortgage product). Secondly, the US government gave Fannie and Freddie a special advantage (low cost of capital) via the implicit guarantee of their debt. Some portfolio lenders, MBS investors, and private insurers are crazy enough to try to compete against their special advantage.
In the auto lending space, one Canadian subprime lender had provided loan extensions/deferrals to 10.4% of its portfolio in 2014. This may (or may not) be a tool used to hide loan delinquencies. It is unclear to me as to what portion of its portfolio currently has missed payments due to loan extensions. This lender does not seem to report this information.
Unfortunately, this lender is difficult to short because it trades on the TSX Venture exchange. It is not worth researching.
Finding bad lending
One trick is to listen to mortgage brokers and other intermediaries. As part of their job, they help risky borrowers get financed. The brokers have figured out which lenders are willing to make the riskiest loans, have weak income verification standards, how to game the lenders’ underwriting standards, etc. etc. Many of them will post good information on the Internet (to attract clients).
The Mortgage Reports by Dan Green is a “good” blog on the US mortgage market. The blog is written with the bias of a mortgage broker; it continually advocates mortgages and refinancings over other alternatives (e.g. renting, paying off a mortgage, etc.). Some of the advice on the blog is very interesting in hindsight. The blog started around the height of the US housing bubble with a post in 2005 about “The Self-Fulfilling Housing Bubble Prophecy Of 2005“.
EDIT (8/25/2014): For advice from a more responsible mortgage insider, check out The Truth About Mortgage. A good post from the bubble era is the one titled “Pick-a-Default-or-a-Foreclosure-Pay Mortgage“.
Another approach to finding bad lending is to look at lenders with the most leverage compared to its peers. Sometimes high leverage coincides with bad management.
Making unusual returns on lending
I see lending as mainly a commodity business. There is not too much room for value creation. Without value creation, a business will not be able to generate superior risk/reward over long periods of time. Here are the areas where I see potential for unusual shareholder returns:
- Businesses with some type of special advantage or moat (e.g. Moody’s). However, Fannie/Freddie shareholders did extremely poorly in 2008/09 despite these companies’ low cost of capital.
- Loan servicers that create value and find efficiencies in servicing bad loans (e.g. OCN/ASPS, CACC). Some debt collection agencies that focus mainly on credit card debt have also generated strong returns for shareholders (e.g. PRAA).
- Gouging/upselling retail consumers. The proceeds of this gouging generally goes to the party originating the loan. Some lenders are integrated with retail operations that originate loans. The retail side is interesting as it can potentially generate unusual returns, though this requires excellent management. Examples would be Carmart (CRMT), pawn and payday lender First Cash Financial (FCFS), WRLD, etc. However, I am uncomfortable with the ethics of payday lending.
Some lenders seem to generate good returns through skilled underwriting of subprime loans (e.g. NICK, HCG.TO). I like these companies a little less. Nicholas Financial is a fine company but (A) I don’t like that they are changing CEOs and (B) I like CACC better. With Home Capital Group, I’m not sure how much of their historical returns should be attributed to luck and how much to skill. Home Capital has benefited from the tailwind of rising Canadian home prices for a very long time, so there is an element of luck. Their investor presentations do not inspire confidence about their risk management. One presentation claims that the company is not risky because there is no Canadian housing bubble. This claim is just silly. If there is a greater than 0% chance that Canada is in a housing bubble that would destroy Home Capital… then there is a risk management problem. Nowadays I stay away from management teams that aren’t straight and honest with investors.
*Disclosure: I own OCN, ASPS and CACC. No position in MCO, PRAA, FCFS, WRLD, NICK, HCG.TO, Fannie/Freddie debt or equity, or WFC.