The conduit market does not absorb a lot of bandwidth in my day-to-day practice; I’m more of a CRE/CLO/warehouse/SASB/new products/innovation sort of guy. But it’s painful to watch this marquee capital markets product wither away, a product that transformed $200 billion of mortgage loans into securities in a single year. That biz might limp over the finish line with a meager $25 billion this year. What happened? The demand for CRE leverage certainly hasn’t changed. The CRE market has gotten significantly bigger since 2008 and consequently, the need for leverage has grown concomitantly. The nature of the underlying real estate assets hasn’t changed all that much, nor the nature of the ownership structure, albeit it is probably a bit more institutional today than it was in 2008. The product is no different, in large measure, today than it was back then and indeed in some minor, twiddling respects might even be better from the perspective of the borrower.
Fixing the conduit is hardly a new idea. We’ve surely tried. Over the past decade, we have made some modest improvements to the servicing experience for the borrower, particularly in connection with the control of unanticipated and what, in the borrower’s view, would be unwarranted fees, but we’ve gotten little to show for it.
At the outset, let me acknowledge that I obviously don’t know if the conduit is dying or whether it’s just dormant, hibernating like a bear waiting for spring. Is this a lacuna or an extinction event? If you are convinced of the former, please skip the balance of this note and wait for spring and my next commentary. For those who are worried, perhaps read on.
The problem seems to be simple enough. It’s like we’re trying to sell a car with a governor on the engine that caps the speed at 40 mph. It works. It gets you from point A to B. Solid, safe. But if no one wants to buy it, we either have to make a new kind of car or get out of the car biz entirely. That may be where we are today. We keep telling the borrower to like our damn product, but seemingly they don’t.
The calculus of the average commercial real estate borrower hasn’t really changed in the years since the heyday of the conduit. How much money do I get; and, how expensive is that money?
The somewhat dismissive joke was oft told while the business was thriving that all the stuff that borrowers disliked about conduits was worth 4 bps.
Demonstrably, that turns out not to be true. It seems the transactional friction, inefficiencies and the simple functional failures of the conduit product are worth a great deal more than 4 bps.
Let’s begin by acknowledging that the problems of the conduit are not all self-inflicted. First, since its day in the sun, many other sources of liquidity have entered the market. There are lots of alternatives for borrowers. Banks have lent more, lifecos have lent more, debt funds breed like tribbles , the GSEs are bigger, bolder and brassier, etc. Some of these, perhaps many, are offering products which, if not precisely competitive with the conduit, are quite close. During the years of the Great Moderation 2.0, we had super-benign interest rates and tons of liquidity. In that environment, the value proposition of the ten-year fixed rate loan was seriously diminished.
But does that explain it all? At some level, a ten-year fixed rate, non-recourse loan should be an extraordinarily attractive product for the commercial real estate space. It particularly ought to be super attractive right now with short term rates on fire, but the data suggest that the conduit business is not reflating.
The problem may be that there are numerous structural flaws in our conduit product which have never been effectively addressed.
It can be boiled down to just two problems. First, while the borrowers like ten-year money, they are not fond of the straightjacket of yield maintenance. For many years, yield maintenance was simply economically unfeasible. (In an every-cloud-has-a-silver-lining sort of thing, the current run up in treasuries has actually made defeasance less painful, but that’s a moment in time, one suspects.)
Second, borrowers hate the fact that their asset management judgments are second-guessed by a seeming army of bureaucratic naysayers. They hate that it takes too long to get a decision. They hate that it costs too much money to process cases and make formalistic arguments for things which appear blindingly obvious to the property owner. They hate the fees they have to pay to get the servicer and directing certificateholder approvals. Frankly, they hate the fact that they have to ask at all. They see the servicing construct of the conduit loan as inconsistent with solid asset management.
No one has seriously suggested that a change to the lockout on ten-year money is a good idea. It’s lucrative and baked into the cake of the calculus of servicers and investors. The yield maintenance mechanic is, like the weather (but, according to the best among us, not this climate), to be tolerated not fixed.
On the other hand, our industry has tried to address the servicing limb of these problems for years without success. We’ve always been aware that the current model is not terrific, but for many years the consensus was that as long as the borrowers just grumble but vote with their dollars by continuing to come to the window, there’s not a compelling reason to change, recognizing that all change would probably negatively impact the lender and investor value proposition.
Perhaps now fixing it is compelling.
Maybe, if all the constituencies in the conduit process stick to their discrete and unique pecuniary interests, stay in their lanes, stick to their conviction that the way the structure works is essentially God-given the borrowers will come back…but what if they don’t come back?
This is the age-old Tragedy of the Commons. This observation about human nature originated with the British economist, William Forster Lloyd, in the early 19th century. The estimable Mr. Lloyd used as an exemplar of the problem, the structure of a medieval village, surrounded by common fields on which the folks were entitled to graze their cattle and sheep. Everyone knew that overgrazing would endanger the viability of the village and, let’s face it, in that time and place, cause them all to die of starvation. But individually, it was certainly advantageous to graze all your cattle and let some other poor bastard’s cattle starve. Perfect solution, except it doesn’t work, if everyone had the same idea and pursued the instincts of their common reptilian brain, starvation and death would ensue.
Something like that may be going on here (okay, no one will actually starve). Everyone knows there’s a problem, but like the boiling frog, no one is inclined to change position, to get out front, to take one for the team.
Let’s do a whip around the principal constituencies.
The investment banking community wants to sell bonds and hopes that everyone else would just get this right. From the banker’s point of view, the easier a product is to sell, the better it is. Doing anything that makes selling the bonds more difficult is not in their pecuniary best interests. Not much help there.
The servicing community is in a pickle. We don’t pay them very well, and we get what we pay for. Putting aside for the moment the servicer’s driving need for additional fee income from the borrowers, which annoys the borrowers no end, the servicers are deeply concerned with liability, particularly when measured against the modest returns the servicing business delivers. But that is problematic for borrowers looking for increasingly efficient asset management. (Sidebar memo: One of the false conceits about our conduit business is that while the CRE CLO leverages unstable assets that require active management and hence has enhanced focus on servicing flexibility, the conduit business finances stabilized commercial real estate where active management is less critical. That’s sort of silly, isn’t it? If you think about it, there’s no such thing as a commercial real estate asset that’s stable over a ten-year time horizon.)
Everyone in the servicing community is, with reason, afraid they’ll be sued by the borrowers; or they’ll be sued by the IG investors (even though they really can’t be) and the B-buyer. They’ll be sued by somebody. Remember, the standard of liability throughout the servicing structure is simple negligence. In a country where everyone has a God-given right to sue everyone over every possible bad outcome (the servicing equivalent of pouring a hot McDonald’s coffee on your lap and blaming the franchise), those concerns about liability are real.
In our current servicing paradigm, it’s easier to get to “No” than to get to “Yes.” It’s easier to demand more and more information before making a decision. Numerous constituencies need to be consulted. Numerous constituencies might need different data, data which is expensive and time consuming to develop. The consequence is that we are stuck with a creaky system that absorbs a monstrous amount of time and energy as servicing decisions bounce from the borrower through the servicing structure, from master to special, from special to B-piece buyer and back again.
Moreover, the servicing standard is itself problematic. As we all know, that standard is to protect the investors as a collective whole with a view to the timely recovery of all payments of principal and interest. That standard elevates certainty of return over actions that might improve, on a long-term basis, the prospect of the mortgaged property. Getting cash back is elevated over long-term value creation. That means that a lot of decisions a competent asset manager of the underlying commercial real estate makes would be inconsistent with the servicing standard. This tension has never been resolved.
In the CRE CLO space, which is purpose-built to leverage transitional assets, the tension between the servicing standard and good asset management became obvious and the consequence was that certain decisions have been removed from the servicing standard. The so-called Administrative Modifications and Criteria-Based Modifications are made by the collateral manager without consent of the special servicer (and, of course, there is no B-buyer). This is deemed to better align the securitization experience with the compelling needs of an asset manager who simply wants to create value.
This hasn’t migrated to the conduit business for a number of reasons. Conduit loans are allegedly stable, the tax regime is REMIC which limits flexibility and the B-buyers are independent players exposed to the risk of first loss exposure. The B-buyer is paid to absorb first loss exposure on a pool of loans and so for very obvious reasons, is highly focused on whether the servicer is protecting the interest of the certificateholders as a collective whole of which they are part – and the first party to get obliterated when losses ensue. Early dollars in are clearly better than dollars deferred. This creates a bias in terms of the rapid realization at par of a pool of loans and does not encourage flexibility in performing asset management (where the annoyed borrower’s option might be prepayment). This essentially substitutes the judgment of the B-buyer for the asset manager. Entirely understandable, but also problematic.
IG investors? Happy with the current structure and, as most are multi-strat investors, there’s not much incentive for them to get behind any actions to grow or indeed preserve this business… they will just take their custom elsewhere.
How about the role of the ratings agencies? Let’s save a little bit of love for them. For better or worse, the ratings agencies’ model rewards unwavering fealty to the existing model and its policies and procedures. Variations are credit negative in almost all cases. Not much incentive there to address changing the conduit paradigm.
So, largely we just wish this would all just go away and the borrowers would come back. Servicers want to hold on to the meager compensation they receive without getting sued. B-buyers want to run the deals and get to par ASAP. Ratings agencies with their hardwired, fully baked-in models, have little incentive to change. IG investors don’t really care for the business contracts. Banks just wish it was easier to sell bonds.
Everyone’s individual pecuniary interest is tied, with considerable fidelity, to the existing conduit paradigm. If everyone sticks to their principles and individualistic pecuniary interests, this product may indeed be ready for a nice glass case at the museum.
Okay, so there you go. What could be done if the key constituencies could come together and think hard about building a structure ab initio — a conduit 2.0?
This may not be possible, or even advisable. It’s indeed possible that the current structure of the conduit will come back into favor. It’s indeed possible that in a period of significant escalating floating rate coupons, the borrower community will once again fall in love with the existing conduit model. Maybe the past several years is just an aberration. Maybe it’s simply a long, cycle where change in demand will be reversed and once again bring us a $100 billion plus conduit market. Maybe.
But, it’s arguably worth a moment to at least wonder whether that’s true.
Here are some ideas.
We could change the standard of care in the servicing model. We could specify that the amount of information provided under the applicable loan documentation is, except in extraordinary circumstances, sufficient for prudent decision making. We could specify that the periods of time available under most loan documentation for making decisions about things like leasing, property releases, etc., are adequate for servicing decisions to be made and servicing decisions should be judged (and liability assessed) within the structure of the available information and timeframe.
Maybe the standard of care is gross negligence, not negligence. (Who’s going to make a lot of money suing servicers anyway?)
Maybe we can import into the conduit things such as the Criteria-Based Modification and Administration Modifications found in current CRE CLOs. Some of this is inconsistent with REMIC, although not all. If these tenets are inconsistent with REMIC, maybe we should consider a non-REMIC shelf! Maybe virtually all leasing decisions are within a very wide box would not subject to servicer or B-buyer consent.
Maybe it’s possible to bake into the servicing standard a notion that the duration of the receipt of interest from an underlying loan and not just the rapidity of recovery of principal is a valuable goal to be weighed in the servicing decision process. One might argue that it is not always good to economically compel a borrower to seek an exit from the loan in order to do what it, as a competent asset manager, thinks it should do about the property. Do investors always want their money back early?
Maybe we should think about paying servicers more and demanding a higher level of service and touch. This could eliminate the need for servicers to chase borrowers for additional payments for off the run servicing decisions, reducing an area of tension with the product.
Let’s have a third rail moment. Maybe we need to do something about yield maintenance. Maybe we need to cap yield maintenance and hence defeasance costs at some sort of dollar amount with a stable relationship to the value of the underlying property. Or, how about five years of full metal jacket yield maintenance and then a sliding scale for the final five years of a ten-year term?
Velocity is critical for the conduit market to thrive. I know the markets are thinking about five-year product, but production volumes and hedging concerns have kept that idea on the sidelines. Can we look into alternatives for the current dearth of effective hedging strategies? Maybe an exchange, like the Merc, could be created where folks could bid on providing hedging for not only interest rate spread but economic spread in an active marketplace. This could be built on a co-op model.
The rating agencies need to consider seriously new criteria for new models. Everything discussed above is probably credit negative to the existing model, but to what extent? This needs to be quantified before an alternate model could even be considered.
All changes to the existing model will make debt more expensive. More flexibility to the borrowers, less certainty in the servicing function, more risk on the B-buyer, will all require risk to be repriced. Maybe borrowers confronted with a more expensive ten-year fixed rate product will once again develop a yearning desire for conduit 1.0. Maybe these 1.0 and 2.0 versions of the conduit are viable alternatives that can exist in parallel, like homo sapiens and Neanderthals for 100,000 years or so (didn’t work out for the latter, but oh well).
I’m not smart enough to capture all the things that might be considered to improve the current dismal straights of our conduit product nor courageous enough for everyone to get at it, but it does seem to me to be something worth thinking about. And, as we think about it, remember the comments of the estimable Holmes who once said, “Once you eliminate the impossible, whatever remains, no matter how improbable, must be the truth.” (And once again, for those of you who think this is a really bad idea, please pretend I didn’t write this.)