(Bloomberg Markets) -- Deepak Narula says wide spreads in the agency mortgage-backed security market offer opportunities approaching those that surfaced in the wake of the global financial crisis.

The founder and chief investment officer of Metacapital Management LP should know. Born in New Delhi, Narula earned a Ph.D. in management science at Columbia Business School before joining Lehman Brothers Holdings Inc. as a research analyst in 1989. There, he rose to become head of mortgage strategy, then to lead Lehman’s agency MBS and derivatives trading desks.

Narula founded Metacapital in 2001 and started its first fund the following year. In 2007, amid the subprime bubble, risks looked so lopsided, he says, that the firm liquidated the fund and returned capital to investors. “The amount of froth in the system was incredible,” Narula says.

Metacapital started a new vehicle in 2008 that went on a tear in the wake of the financial crisis. It was the best-performing large hedge fund in Bloomberg Markets’ 2012 ranking, with a 41% return for the year. After that, with the Federal Reserve holding rates low, the fund had some tough years, Narula says. He was optimistic going into 2020. Then Covid-19 hit. “There was a week in 2020 that was a brutal week for us,” he says. The firm’s assets under management are now about $125 million.

That lower supply and other factors are now creating interesting opportunities in the market, Narula says. On a sunny Wednesday in May, he spoke with Bloomberg Markets at Metacapital’s offices, next to New York’s Carnegie Hall. BI’s Adelberg joined the conversation, which has been edited for length and clarity.

JON ASMUNDSSON: Could you tell us about where you see the opportunities now?

DEEPAK NARULA: Absolutely. What makes agency mortgages interesting is that you’re never worried about getting your money back. As long as the US agencies [Fannie Mae, Freddie Mac, Ginnie Mae] that have the backing of the US government are there to provide the guarantee on the principal and interest payments, that’s never a concern. Yet you will find that these ­government-guaranteed Ginnie Maes and quasi-government-guaranteed Fannie and Freddie securities will trade at significant spreads to Treasuries—even net of what is the embedded option in these, which is the prepayment option. That is really what differentiates these securities from a lot of other government debt and also a lot of corporate securities.

The interesting thing about agency MBS is not that you’re worried about getting your money back, you just don’t know when you’ll get it back. It’s the uncertainty of the timing of cash flows that makes it particularly interesting. And the reason for that is homeowners in the US—very unique mortgage market—can take out a 30-year loan and, without penalty, pay that loan anytime. That prepayment is what makes the timing of cash flow uncertain, which produces basically a risk premium that is embedded in these securities. One is explicit in the form of the fact that when rates go down, homeowners refinance.

We’ve lived through a big refinancing boom post-Covid. You had basically 2.5% mortgage rates available to the vast majority of the population. That’s just very compelling. As rates go down, borrowers refinance their debt. That makes a lot of sense, you can save a lot of money. But that makes these bonds callable. So it’s a government-guaranteed security that has embedded in it a call option that is linked to rates. But it’s not linked only to rates. It’s also linked to the state of the housing markets, which are evolving and uncertain and hard to predict. Trying to forecast housing over any long period of time is like trying to forecast GDP. It’s generally a fool’s game. What you find is there’s a risk premium embedded in these securities in the spread that they offer over Treasuries.

In our world, we basically call it the option-adjusted spread—adjusted for the prepayment option to basically come up with what is the net spread that you’re getting over Treasuries. And so—long answer to your question—what makes the sector interesting is when these OASs are really wide.

Fundamentally, when you’re talking Ginnie Mae, you’re talking US Treasury credit, and so for it to trade for significant periods of time at wide spreads to Treasuries is unusual. It’s very unusual. Currently spreads look extremely wide, and there’s a confluence of factors that’s made it happen.

One of the bigger contributors is the Fed’s QT [quantitative tightening] program and the distress that we’ve had with small and midsize banks recently. That is something we think makes for a very compelling investment.

JA: Can you talk a little bit about the banks?

DN: Yeah, post-­financial crisis, the big banks, many of them looked like they weren’t going to make it. That, we all know, gave rise to a lot of the regulation focused on the too-big-to-fail banks.

Part of the regulations was to basically make sure the banks have enough liquid assets to be able to meet any run on deposits. You’ve got to own either Treasuries or Ginnie Maes. Fannies and Freddies have a place there also. What is missing in that piece of regulation—after the fact, it’s pretty obvious—is that these securities have enormous interest-rate risk. If you don’t put in place regulations, or if the banks don’t manage that interest-rate risk, then you can have the kind of problems we’ve had recently.

JA: Why do you feel like that was ignored in the last round of regulations?

DN: If you look at rate moves that we’ve seen in most of our lifetimes, they haven’t been that large, and they’ve been orderly. I remember ’94, when the Fed hiked pretty rapidly and raised rates 300 basis points and sent the mortgage market and lots of other markets into a tailspin, right?

But subsequent to that, we haven’t had that. If you look at the hikes that got us to the financial crisis, rates were hiked at 17 consecutive meetings [June 2004 through June 2006] by 25 basis points. That took a long time, and it was really orderly, and in the end it tipped us into the financial crisis with the lag because it finally took housing down. It finally took the subprime market down, and the excesses all came to the surface.

Compare that to what we’ve got this time. You’ve got short rates at zero for a long period of time, long rates at very low levels thanks to QE [quantitative easing], and the central bank tells you this inflation is transitory. It can lead to some managers saying the risk is such that we can afford to own these securities. They produce positive returns to where our deposits are, and they’re accretive even though the returns are not that high.

This is clearly a place where the regulators messed up. I mean, you push the banks to buy liquid securities. You talk about inflation being transitory, and then you raise rates by 525 basis points in under a year. Something’s going to break, right?

ERICA ADELBERG: Do you think the Fed panicked? Do you think they were just trying to maintain credibility or do you think that’s what they had to do?

DN: The Fed was a year late, and they had a year. And so that year of excess is something they had to make up for. QE should have ended, and rates should have started to get normalized at the start of ’21, not at the start of ’22. Housing was fine, equities were fine, unemployment stabilized. The economy handled it really well thanks to the Fed and what came out of DC, the Fed just overstayed its welcome by a good year.

JA: Could you talk a bit about interest-only and inverse ­interest-only strips? What do you see happening there?

DN: Absolutely. I mentioned that government-guaranteed mortgages are really not a credit play. It’s really the timing of cash flow that is uncertain. But there are a set of securities that actually take that prepayment uncertainty, and that risk is concentrated in a much larger way in these securities. The reason that happens is because a lot of the end buyers are like, “We like the credit, we like the fact that we get a spread over Treasuries, but we really don’t want this prepayment risk that we don’t understand too well. It’s hard to model. There’s a lot of work that’s been done to model it. Yet every time we go through a new prepayment wave, there’s always new learning, and we are like, ‘Oh my God, I never thought about that.’ ” Right?

So a lot of institutions will say, “Just give us par price bonds, so we know what our yield is going to be. We’re going to earn the coupon, and maybe we’ll own it for longer, maybe we’ll own it for shorter, but we know that’s what we’re going to earn on it.” The collateralized mortgage obligation market will chop up agency MBS into what are prepayment-protected par bonds, par price bonds, where the principal will come back at par and the balance of the risk is put in securities where you basically own the prepayment risk. That’s one way to tranche them. You make what is called a floating-rate security at par, and you make an inverse IO against it. That sounds complicated, but you just take a fixed-rate bond, and you sell a floater. And what you’re left with is a security with no principal, just the fixed rate minus the coupon on the floater. So it’s inverse floating rate, but it’s interest only, since all the principal went to the floater. The other way to tranche the risk up is just take your agency MBS and break it up into the two cash-flow streams that correspond to monthly payments that are either interest or principal. Every month when you pay your mortgage, part of that payment is used to retire some principal and part of it is an interest payment against the debt outstanding. If you can separate those two cash-flow streams, you create securities that are called for the interest portion, interest-only securities. The IO. And for the principal only, the PO.

The IO is a very unique security. All it does is pay you interest, never pays you principal. The timing of how much interest you get is uncertain. There’s no natural buyer for IO, and folks like us will buy IOs if they are priced at levels where we get compensated for the prepayment risk and more. So currently, for example, our portfolio of IO—and I’ll use the word “IO” loosely to refer to IO and inverse IO—it’s got a prepayment option-adjusted spread that’s close to 600 basis points over Treasuries. You’ve got to ask the question: Government-guaranteed risk paying you 600 over Treasuries, what gives? How is that possible? That’s because it has in it the concentrated prepayment risk that is hard to model over the next 30 years. For all the work that we’ve done in modeling it, the models continue to evolve. And you need to be compensated for that risk, and you get paid excess spread. So IOs have a place in portfolio for a couple of reasons.

One of the things that’s very unique about IOs is when rates go up, most bonds go down in price. The IO security, on the other hand, when rates go up, prepayments slow down. You earn the interest stream for a longer period than you were planning to when you bought the bonds. The IO security actually goes up in price. That’s called having negative duration. A lot of folks will buy IOs to manage portfolio duration. Mutual funds and money managers will do that. The second thing that’s interesting about IO is that it just offers wide spreads. Right now they’re yielding double digits, north of 10% yields. So folks will buy it for yield enhancement.

The way we look at it is neither of those two. The way we look at the IO market is: What are the option-adjusted spreads that the security can offer, and what is the likely path of this OAS going forward? Is it too wide? Are we going to get prepayment surprises to the negative side that will cause investors to be leery of this market and look to sell it and prices will go down, causing spreads to widen further? Or are we at a point in time when there’s good prepayment news and, on top of that, you’re getting paid 600 over Treasuries.

That scenario is really when IOs become very interesting. And that’s where we find ourselves today. That’s why probably our top pick in the agency marketplace right now is a portfolio of IO and inverse IO, hedged for rates. We think expected returns there are fairly high. They are well into the teens, if not more. This is without using any leverage.

The reason I say we really like where we are in the marketplace today is, look, Covid caused the Fed to cut rates to zero. And with a lag, housing markets reacted very favorably to what came out of the Fed and Washington in terms of the stimulus. The big surprise was how well housing did. Who would have thought that the US is going to be hit by this pandemic and home prices are going to go up at 20% per annum?

The prepayment models are still calibrated to the Covid experience. To those models we are buying IO and inverse IO, at option-adjusted spreads of 600 over Treasuries. What is the likely outcome for how these models are going to be changed going forward? They’re going to have to be dialed back. For most folks that have built mortgage models, they have not seen a move like this in their lifetime. A 500-basis-point move in rates, where if you look at the average interest rate that US homeowners are paying on their mortgage, it’s about 350 basis points below current market rates. We’ve never seen this. How can you build a model to data that you don’t have? You can’t fit any data. It will take a while for these models to catch up, and they will all get slower. The market’s expectations of future prepayments will move faster than the models.

There are sectors where that might be a little faster, a little slower, but in aggregate you’re seeing prepayments that continue to leave most analysts scratching their heads and saying, “Wow, a lot slower than expected.” And so, yeah, the opportunity in the IO and inverse IO market—to us—is right now, right here. 

Asmundsson is FFM editor of Bloomberg Markets.

--With assistance from Anibal Arrascue.

©2023 Bloomberg L.P.