…First, credit losses on pools (securities) of mortgage loans come from four sources: 1) The program parameters of the loan (also called product or loan/lending guidelines). For example, a mortgage with little or no down payment/equity. (The lower the down payment/equity, the higher the loan-to-home value ratio, and generally the higher the credit risk/risk of borrower default.) Or a borrower with a low credit score. (A borrower with lower or poor credit, is generally a higher credit risk.) 2) An error by the mortgage underwriter in determining whether the borrower and the property/collateral meet the program parameters. 3) Fraud on the part of the buyer/borrower, seller/builder, lender, appraiser, Realtor, or other party to the transaction. and 4) Macroeconomic events beyond the control of the borrower, mortgage lender, or investor. Examples would include significant changes in home prices and/or employment.

In preparing this blog posting, I sent a note out to several industry veterans asking the following question: “How does a mortgage underwriter know that a mortgage application is prudent and responsible and therefore they should approve it? Seems like a pretty simple question, but is it?” Everyone in the industry agrees that a loan with 20% down payment, where the borrower has good credit, a good job that pays enough to easily afford the mortgage, and solid cash reserves (for a “rainy day”) in the bank is a prudent and responsible mortgage. Everyone knows intuitively that’s a good loan, but the reality is they don’t really know. And it’s even harder to determine what’s a “prudent and responsible” mortgage, when you deal with the reality that most mortgages today (especially first-time homebuyers) involve very little down payment, to borrowers who are stretching to afford the home and mortgage. Why don’t mortgage underwriters and other mortgage industry vets know what’s a “prudent and responsible” mortgage? Because, believe it or not, nearly everyone in the mortgage banking industry is not involved in assessing mortgage credit risk. They may think they are (even seasoned veterans), but they aren’t. Why is that? Any mortgage loan can go bad, even one that appears very safe upfront, because a percentage of mortgage borrowers, suffer a personal tragedy…death, divorce, job loss, serious health issue, etc….each year (I have heard it’s 5% a year) that causes them to struggle with their mortgage. So think about it, what would an underwriter do knowing that one mortgage they approved defaulted and the lender/investor suffered a loss, as result? Chances are they would draw the wrong conclusions and make the wrong underwriting decisions going forward. To properly assess mortgage credit risk, you have to look at mortgage credit losses in large, statistically-valid pools (by product-type, by geography, and by many other characteristics) and by vintage (year) over many years and over various economic cycles. Almost no one in the mortgage industry does that. The ones who did that, were the rating agencies, Fannie and Freddie, FHA, and a handful of other large institutions and they set the lending guidelines, for the entire industry, from what’s called today “Big Data.” To restate, the bulk of the mortgage industry is mainly just data collectors and fact/data verifiers, with a hunch (that could just as easily be wrong, as right) about mortgage credit risk.

I know I am repeating myself, but this is important to understand. Underwriters in the mortgage lending industry, even ones with years of experience, have little-to-no ability to know whether a particular mortgage loan is a “prudent and responsible” mortgage/acceptable credit risk or not. Why is this? Mortgage underwriters are only responsible for making sure that a borrower and the home (value, as collateral) meet the loan program’s parameters/lending guidelines. For various reasons to extensive for me to discuss here, while they might hear about a loan they underwrote and approved that defaulted or had to be repurchased from an investor or indemnified against loss, they never get to see the overall results of their underwriting decisions. They never see the actual performance of all the loans they underwrite and in most instances it wouldn’t be helpful anyway, because most mortgage underwriters and even most mortgage lenders, don’t underwrite a large enough population of loans, over diverse economic cycles, to be able to draw any statistically-valid conclusions. In other words, mortgage underwriters are mainly fact-checkers. Their role is to ensure that the mortgage loan file (some now over 1,000 pages of numerous documents) meets the mortgage loan program guidelines and if it does not (and cannot), to generally decline or reject the borrower’s mortgage application. Mortgage loan programs have guidelines, rather than absolute rules, so borrowers with a minor weakness in one area, may still appropriately be approved for a mortgage, if they have strengths in one or more other areas. In industry jargon, these strengths are called “compensating factors”, that compensate for a weakness in another area. For example, let’s say a loan program’s guidelines requires a borrower to have a minimum FICO score of 680 and the borrower has an actual score of say 670 (a minor difference), but has cash in the bank that far exceeds the minimum and a better job and income than normally required. The underwriter, in their judgment, may approve the loan based on these positive “compensating factors.” My belief is that some, and possibly most, of the mortgage defects cited in government and other claims against the industry, are minor guideline discrepancies, where other compensating factors, resulted in the mortgage’s approval. In future blog postings on mortgage credit risk, I will show how initial mortgage defect rates cited by the government and others, fall dramatically (are reduced by 80% to 90%) as mortgage lenders and investors like FHA engage in a proper review (back-and-forth) of these initial mortgage defect claims. Much like the back and forth you have when you are audited by the I.R.S.. I believe the government deliberately stopped this “back-and-forth” process for crisis era (bubble/bust period) mortgages (2006-2011), because they knew it would dramatically reduce their Red Herring allegations of high mortgage defect rates and therefore their claims and moral authority.

Bottom line, the activity of mortgage underwriting plays a relatively small role in credit risk management and any errors that they might make would generally be fairly small and not systemic. As result, despite all the hype to the contrary, mortgage underwriting had little-or-nothing to do with the U.S. mortgage and housing market collapse in 2008 and resulting financial crisis. And while mortgage fraud might have been elevated during the bubble, again it’s a smaller, non-systemic issue. (No one, as far as I am aware, has provided any statistically-valid facts to support mortgage fraud being a material cause of the mortgage crisis.) So what caused the U.S. mortgage crisis? I believe its #1 above, expanded/riskier loan program guidelines over time, and #4 above, macroeconomic events. I have discussed/documented the macroeconomic issues at length on this blog and so I will focus my next blog posting on #1, expanded/riskier loan program guidelines.”, Mike Perry, former Chairman and CEO, IndyMac Bank

This May 17, 2016 email from me to Mark Calabria of The Cato Institute (responding to his article advocating portfolio lending for mortgages over securitization), further explains my points above:

Dear Mark (Calabria of The Cato Institute),

My name is Mike Perry. I am the 53-year old former Chairman and CEO of IndyMac Bank, which failed during the 2008 financial crisis (and was a major mortgage lender and issuer of private MBS, pre-crisis). I have a blog at www.nottoobigtofail.org that is all about IndyMac, mortgage lending and the financial crisis, that you might be interested in. I am also a small donor and follower of The Cato Institute.

I read your April 21 Commentary this morning and I found much to like and agree with you on.

In regards to securitization vs. the originate-and-hold model of mortgage finance, you miss two HUGE benefits of securitization.

Let me step back for a minute. I recently asked a number of senior mortgage industry folks, “how do you know when you have underwritten a prudent/responsible mortgage loan?” Pretty simple question, right? No one could answer this simple question, in my view.

Why? First, because mortgage underwriting is mis-named/mis-understood. Unlike underwriters in say insurance, mortgage underwriters do not have any skills or experience in determine the credit risk of a particular mortgage (borrower and collateral). Mortgage underwriters are just fact checkers. Making sure the borrower and the collateral (property), meet the program guidelines (whether those are established by the government mortgage firms like Fannie, FHA, whether those are established by private securitization or secondary whole loan markets, or whether they are established by the underwriters own bank, in an originate-and-hold environment). In fact, I checked recently, and mortgage underwriters still rarely see the actual performance of the mortgage they underwrite and even if they did, how would they evaluate that, when its not a big enough or representative enough sample to determine with any statistical relevance, whether the loan should have been made or not.

Clearly, I think we all can guess that loans with 20% or more in down payment/equity, where the borrower has excellent credit, and a solid and stable job where the income clearly allows them to afford the mortgage is a prudent risk. But, we live in the real world and the real world is that most purchase mortgage loans (and most of these are provided through a government guarantee) these days, involve very little in the form of down payment and often to borrower with less-than-excellent credit, who are stretching to pay the mortgage, and have little in cash reserves or other assets.

In that situation, the only way to know that a mortgage is prudent/responsible….is if you have lot’s and lot’s of loans and can evaluate their performance over time via statistical models. That, to me, was the biggest benefit of securitization.

Also, I see a lot of originate and hold mortgage lenders putting 30-year fixed rate mortgages in portfolio these days, at these historically low rates. In fact, because a portfolio lender does not have to mark to market their loans held for investment, they can hide a loss on their balance sheet and bleed it out (through lower spread income) over the life of a mortgage. In the originate-to-sell mortgage finance model, if you make an underwriting or pricing error, you immediately are penalized for it, when you try to sell it into the secondary market.

You are right about investors needing to take on risk and understand risk. You are right about government mortgage and deposit insurance distorting the marketplace. You are right about Basel 3 and risk-based capital resulting in too much leverage.

Those all need to be addressed. Take the National Statistical Rating Agencies out of the process and force investors to evaluate the underlying collateral (mortgages) in the bonds…..they would then determine how much subordination the originator needed to retain, before they would buy a senior tranche.

Something like that. Happy to discuss anytime. Best, mike