Housing Policy

PolicyCast: In the Room Where CRT Happened

Episode 2 – 9/14/2020

In the room where CRT happened

Today’s guest is former Freddie Mac CEO Don Layton. Mr. Layton and Kirk Willison discuss the GSEs transferring mortgage credit risk.

Episode 2 Transcription

Kirk Willison: I’m Kirk Willison with a second episode of the Arch Mortgage Insurance PolicyCast,  a podcast about key issues shaping housing in America. One of those key issues is the future of credit risk transfer or CRT. It’s a valuable tool for Fannie Mae and Freddie Mac to move risk off their books and off the taxpayers’ backs by sending it to private investors instead. They’ve used CRT to move 70% of the credit risk on new loans they’ve purchased the past seven years. But now credit risk transfer is in jeopardy, because of a newly proposed rule by Fannie and Freddie’s regulator. That could mean higher mortgage costs for borrowers and more risk exposure for American taxpayers. So what is CRT? And, why is it so important to housing in America? To find out, I’ll be talking with former Freddie Mac CEO, Don Layton. Don was in the room where the first CRT deal happened. He calls CRT the biggest GSE reform of the last decade. Don spent seven years as Freddy’s Chief Executive before retiring in June of 2019. Well, retiring is too strong of a word. Today, Don is known as a prolific writer of policy papers as a Senior Industry Fellow for Harvard Universities Joint Center for Housing Studies. Well, welcome to Policy Cast Don, before we start our conversation about credit risk transfer, I thought I’d just ask you how did you end up at Freddie Mac?


Don Layton: Well, I was a retired banking finance executive at the time. Well, semi-retired I was on a few boards and had a long standing interest in public service from when I was young. Then I was approached by the board of Freddie acting on behalf of the FHFA as a conservator to recruit a new CEO. And it was a public service position working basically for the government to run Freddie and help make the finance system better. So that’s what attracted me. So I came out of retirement at a relatively advanced age and started commuting to Virginia, and there went seven years.


Kirk Willison: So when you arrived there Don the GSEs practice what’s known as a buy and hold model, but it failed with awful consequences. What happened and why?


Don Layton: Okay. So bear with me there are a few layers on this. First, obviously, in 2008 we discovered that having four and a half trillion dollars of mortgage credit risk, one asset class divided up between just two companies, was not a good idea from a systemic risk point of view. I mean, it’s just obviously a bad concentration of risk. There’s this old phrase, “The only free lunch in finance is diversification”. This is like the anti-free lunch. The more colloquial version of that is that there were a lot of eggs and just two baskets. So it was an accident waiting to happen and the bad things of the housing bubble happened and the chickens came home to roost. The companies ended up failing and going to conservatorship. Second, I want to clarify and really explain that it wasn’t technically just a buy and hold model. It was a highly subsidized buy and hold model that allowed the core business to be run quite honestly not all that well. It was more about the subsidies. There were two big subsidies. One was of course the giant subsidized investment portfolios they built up and when I say subsidized the funding to carry it was subsidized by the government indirectly. It became the largest share profits of the GSEs, not the guaranteed business, but the subsidized investment portfolio. That was scheduled to be taken out and it has been taken out and it’s gone. So it leaves the core business more standing on its own. The second subsidy was the ultra low capital requirement, even against relatively low capital requirements of the era, that also is scheduled to go away with whenever the FHFA finalizes a capital rule. But even in conservatorship, we’ve been running with hypothetical capital of a full level. That means the stripped down core business is now naked. It has to perform well. It has to stand on its own two legs and be economically viable. That’s actually something that’s never been seen in the GSE world before.


Kirk Willison: One of the ways that then you tried to work that concentration of risk is that you created a credit risk transfer model at Freddie Mac. So just so everyone understands what is credit risk transfer? What is CRT?


Don Layton: I want to back up and talk about the basic securitization model of the GSEs. They buy mortgages in and they issue the agency MBS out. So now the investors would be on the hook for all the risks. But, then the GSEs guaranteed with government support the credit risk to the investors. Functionally, that means the interest rate risk of the mortgages pools are passed through to these global investors who can handle it. And the credit risk unfortunately then piles up on the GSE books, it all comes back with these guarantees. CRT is designed to mimic the pass through mechanism of agency MBS that passes through this interest rate risk to be a mechanism to pass through the credit risk to investors separately. Two separate streams, leaving the MBS market undisturbed. So CRT is technically a financial contract. There are several types, one is most dominant and they pass through the credit risk on specific pools of mortgages to investors who buy it. It’s actually patterned after what is colloquially called cat bonds or catastrophe bonds and are more officially known as insurance link notes. They have been used for hurricanes by insurance companies for years, it’s a tried and true structure. So, we weren’t inventing anything new in that sense. The first one was done in 2013 by Freddie, our brand name is called Stacker and it’s a bond execution, that’s the most common form.


Kirk Willison: And, how did that actually work?


Don Layton: The way this works again patterned after cat bonds is first we put out giant database releases to the public now. So these investors can see all the history of credit and what losses have been and the characteristics of the pool. And then the characteristics of the specific pool we’re showing. So they have the data to make a good judgment in their view of how much risk this is and what they think is a reasonable price for it. We then sell a bond, which is called a reference pool bond in which we go, on this pool of assets which we’ve defined for you if cumulative credit losses are greater than X basis points but less than Y basis points, you reimburse us for any losses there. And to make sure that we do get reimbursed, they put the money upfront with the trustee for the maximum loss. At the end, if we don’t use some, they get it back. That’s the basic structure. And they then get an interest rate that is well above a normal like low risk rate to take the risk of this. It’s done in tranches, if it’s a riskier tranche you get paid more, if it’s a less risky tranche you get paid less, and that’s your cap bond structure apply to.


Kirk Willison: So under that structure, the GSEs don’t face any risk of nonpayment whatsoever because all the cash has been delivered by the investors?


Don Layton: That’s correct.


Kirk Willison: Now you’ve created a couple of other structures that are a bit different than that. What kind of things have you done beyond the capital market store?


Don Layton: The second biggest one is using insurance companies, we usually say re-insurance technically it doesn’t have to be. But, insurance companies whereby we have instead of a bond execution is contracts with insurance companies, insurance style contracts. But again, the same structure of more than X basis points loss less than Y basis points loss you reimburse us. We call that ASIS again a brand name, and it very clearly mimics the Stacker bond, but it’s insurance execution. In that case, the nature of insurance companies, we don’t get a hundred percent cash collateral upfront. We get a negotiated, partial cash collateral, depending upon the rating of the insurer. If the rating is very high, they’d give us less the ratings lower, they have to give us more. So we think the risk of reimbursement losses are pretty low.


Kirk Willison: You’ve called CRT the biggest single reform in the GSE model in the past decade. How have the capital markets, insurance markets reacted? How much of the risk are they now taking off the books of the GSEs?


Don Layton: There are two answers to that, the stock and the flow. Up until the virus so let’s say 2019 the credit risk on new mortgage transactions as measured by FHFA formula. The CRT transactions of both types between both Freddie and Fannie are taking north 70% or north of 70% of the risk off our hands.


Kirk Willison: Don, what are some of the other benefits of CRT?


Don Layton: The biggest one is market discipline. So here are these two GSEs with this four and a half at the time, now over 5 trillion of single family mortgage credit risk. And it’s just their internal judgment whether it’s good risk. There’s no way to second guess it, there’s no way for them to double check, now we do. When you have a package of mortgages, the quarters production of mortgages, and you do a credit risk transfer bond on, stack of bond on it, those investors look at it, they run their models, they analyze it and they give you their feedback well, informally, directly, and formerly, and what price they’re willing to buy it at if any price. So that means that it’s much less likely the GSEs could sort of get out of line and be weird, and the credit could be too tight or too loose. It keeps them on the straight and narrow. That is at those kinds of numbers of trillions of dollars. That’s a really important function.


Kirk Willison: Well, one of the things that have come up in recent conversations is that there are some perhaps not fans of CRT, who suggest that in times of economic turmoil it can’t be trusted. That investors are going to flee. But what have we learned from the most recent episode of the pandemic and seeing how investors are reacting to it?


Don Layton: So let me explain. I think most of the accusation really is just political stuff. Putting aside people who just don’t like hearing about any change whatsoever, which there are a lot in housing finance. You said, “investors flee”, that somehow implies that when we have a loss, and it’s time to be reimbursed, the investors are gone. They’re not there to give you the money, but as we said, they give you the money upfront, it’s there. We only give it back to them later with whatever’s leftover. So if you have trillions of dollars outstanding of this, which we do, that all gets paid, it works just as expected. The issue shrinks down to, can you do a new deal this month, this quarter versus all the old deals, the old deals are fine. So the reality is markets go into distress, and we saw that in the pandemic. It was not practical to do some new deals starting in the last deals. I think we’re done about mid March and then the markets fundamentally closed. There was always secondary trading, but it was a weird market obviously. The market’s closed for about six to eight weeks. And then Freddie started issuing again, there was a delay with FHFA having to approve things. So I think with that two and a half months.


Kirk Willison: And then that really isn’t out of the ordinary for?


Don Layton: No. Well, so what did we do during that time, we bought and we did buy and hold. Like we had done for half a century before that, except in this case, we don’t buy to hold forever. We just buy and hold until the markets open up again. And then you go back and do old deals. I want lastly to go after this false logic. When people say that you go, so this is what you’re proposing, If we can’t promise to do CRT 100.00% of the time instead it’s maybe 98 or 90% of the time. Getting cheaper capital which we do via CRT and reducing systemic risk. Then, the argument is, if you can’t do a hundred percent of the time, but only 98, then you should do it 0% of the time. That makes no sense whatsoever, it is 99% politics and the other 1% is people who don’t like change. When this was invented from 2013 to about 2018, the republicans and the democrats in Congress and the administration all loved this. It put private capital in front of the taxpayer and only later did the politics end true. And so we just need to go back to those days where it was considered a professional nonpartisan matter. And, this program could be one of the great reforms of the GSE system.


Kirk Willison: It brings us to today. The federal housing finance agency has issued a new capital rule. And at its best.


Don Layton: They’ve issued a draft of a new capital rule.


Kirk Willison: A draft of new capital rule well put. And, at its best, it cuts the capital credit for CRT in half. And in some cases, it could go all the way down to zero. What’s that going to mean for borrowers, for taxpayers, for investors, for the GSEs themselves?


Don Layton: The problem is if you make CRT uneconomic via the capital rule, then you’re back to the old world, all those mortgage eggs and those two baskets, which means you’re asking for a problem if the markets get bad. Now the FHFA’s answer to that is, well, then we just need because of this concentration risk to have really premium levels of capital on it. That does solve the problem, but it solves it so expensively it’s ridiculous. You’re talking about another $50 billion in capital or something.


Kirk Willison: Is it going to be an incredible hurdle anyway for the GSEs to raise the amount of private common equity that has been called for in this rule?


Don Layton: The rule calls for about 240 billion of equity between the two companies, some could possibly be preferred. Even if you take the original 2018 capital rule, put out earlier which this replaced as a proposal, the number then would have been about 130 or 140 billion. Now you can use over many years retained earnings to finance some of a lot of this, but I’ll remind people that the largest IPO in history was Alibaba for $25 billion. And this is not a high growth internet success like Alibaba. I believe the largest US one was about 20 billion. When General Motors was refloated, 20 billion versus a hundred or 200 billion. You’re talking about not just attracting equity, you’re talking about tracking it in amounts never seen before. So this is not a one and done IPO. This is going to take more issues and more tranches and timing than anything ever seen. They’re just, the companies are just that huge. If the negative approach to CRT stands, it will make raising common equity very hard, because you’re back to a concentrated risk business model and equity investors will demand a premium return for that risk. Which means the price is lower, many investors may regard as too risky. You know you have to make the company attractive to the investors, or they will just vote with their feet and stay away. And given the amounts we’re talking about, you can’t let many vote with their feet and stay away.


Kirk Willison: But, if CRT was given greater credit, it might ease that transition out of conservatorship?


Don Layton: Yes, the reality is CRT absorbs loss, they reimburse the GSEs and capital is designed to absorb loss. So CRT is an unusual form of capital. Therefore, if you have a lot of CRT, you need less for the rest of the company. So the answer is yes, a robust CRT program means it’s easier and quicker to raise capital in effectivel and no CRT program means it’s longer, more expensive, and more difficult.


Kirk Willison: Let’s talk about another form of CRT, mortgage insurance is a form of CRT, and I’d like to specifically focus in on a product that both Freddie and Arch Mortage participated in. We call it MRT for mortgage risk transfer. Freddy called it Imagine. How is that different than typical mortgage insurance? And why was Freddie attracted to that product?


Don Layton: Traditional mortgage insurance is just a form of CRT. It’s a specific form that goes back a long way in terms of its structure, but it’s the same thing. We get reimbursed for losses. There were two weaknesses with it, number one, it’s very expensive. The other of course is this issue of payment, the reimbursement to us was insecure. So our specialists, after a while on CRT went, let’s address these problems and do a new structure of MI that solves the two problems, expensive and that the reimbursement is secure like we talked about for CRT. And, that became Imagine, and that meets all the charter requirements, but it’s a wholesale business to business product. So it’s much less expensive. Estimates are 20, 30% less expensive. And in a similar way, I’d mentioned between credit ratings and partial collateralization. We’re in a much better place of getting reimbursed than traditional MI, which has weaker credit ratings and no collateralization. It’s proven itself. We figure it’s about $600 to $800 a year less for the users, which is a significant amount of money. And if you apply to the entire MI book, that’s billions of dollars.


Kirk Willison: In keeping with the MI theme here and being with Arch Mortgage I guess that makes sense. The MI companies themselves are using CRT for their own portfolios. I guess, following up on the success of the GSEs and Arch Mortgage again was first and that with the Bellamead transactions. What’s your reading of how those are going?


Don Layton: They’re going great because CRT at the GSEs is going great because it’s the exact same thing. GSE model line company mortgage credit, MI model line mortgage credit, is not a good business model. And as I said to one then CEO of an MI, your expected tenure is lower if you have no credit risk transfer, and all your eggs are in the one basket, the light bulb went on there too. And so people have started to they did it and what did they get? They found the capital was cheaper and they didn’t have all the concentration of risk. So if the housing credit markets went down, they only went down partially not wholly. And so they love it and they’re doing it and exactly imitation of what the GSEs were doing.


Kirk Willison: So do you think the situation now where CRT is at risk is that due to politics?


Don Layton: The only current issue about CRT moving forward is the capital rule. There are two parts of the capital rule, which makes CRT problematic, ie literally would render it, what I think of, is falsely uneconomic. The GSEs with reprivatization are supposed to be privately owned stock holder companies, which means they work on the economics. They are not command and control government agencies who do what they’re told regardless of the economics. And if the economics are rendered falsely uneconomic by the treatment of CRT or on a different topic, a leverage ratio being combined constraint capital versus risk based capital, then you could literally have a stoppage of CRT being done. We’ve got the canary in the coal mine for that because when the markets reopened I guess it was late May, early June, Freddie Mac started doing CRT transactions, Fannie Mae did not. And they announced they were not until it was clear whether the capital rule would be accommodating CRT or make it falsely uneconomic. And so literally today, as we speak risk is being reconcentrated in Fannie. So we’re literally going backward in terms of addressing the systemic risk concentration in the two GSEs right now. So I consider that to be a very bad result, and that’s a warning of what might happen.


Kirk Willison: If you could narrow it down Don to let’s say three or five recommendations that FHFA should modify as it reconsiders the proposal after it looks at all the comments, what would those be?


Don Layton: I think they need to, uh, they currently have a hybrid of risk based and not risk based. Even the risk based calculation has many non-risk based items in it. They need to get more consistent on a risk based calculation being risk based and a non-risk based is not. They now have them intermixed, that’s one. Number two, the reality is their leverage ratio is more appropriate for a bank and not a GSE, which while it has many risks the same as a bank not packaged the same way. GSEs are not banks. And they’d be better off going back to the leverage ratio mechanism in the 2018 proposal, which was consistent with how a GSE should run as opposed to a bank. I know one commenter did their calculations and said if you did that kind of thing, the 240 billion of capital would drop to about 180 they thought, they didn’t address other issues. Just those few and you wouldn’t have false economics about CRT stopping the usage.


Kirk Willison: You spent a long time at Freddie Mac. You spent a lot of time likely in meetings with FHFA, are you an optimist? Do you think that they will reconsider their position on CRT?


Don Layton: What I feel comfortable predicting is that the current proposal is enough out of the mainstream of financial regulation, that it will not have broad credibility. That means the first time a Democrat administration appoints an FHFA director, they’ll throw this one out and put a different one.


Kirk Willison: Let’s hope they get it right.


Kirk Willison: Hey, thank you Don for joining me on Policy Cast.


Don Layton: You’re quite welcome, Kirk. I am happy to do it. As you know, my only objective now as a retiree is to move the system in a good direction. I have no other mode of going on.