Feast, Then Famine

Jonah Coste is an economist at the FHFA and spear-headed the team’s research. Dr. Michael Seiler is a full-time professor at the College of William and Mary in Williamsburg, Va., and a visiting scholar at the FHFA. Will Doerner is a supervisory economist in the FHFA’s Division of Research and Statistics. Ross Batzer, a senior FHFA economist and the fourth member of the team, was unavailable at the time of the interview.

Q: What was the central question that you set out to answer with this research?

Jonah Coste: The first was just quantifying the exposure to lock-in. As of the end of last year, the average active mortgage was 3.2 percentage points lower than what they could get on the same mortgage at that time in the fourth quarter. You can also look at it in a dollar amount. It’d be about $500/month, and additional payments if they had to remortgage, which is about 40% higher principal and interest payments. Or, you can look at the total present value of those increased payments, which is about $60,000 for the average mortgage holder.

The second was asking, how sensitive is the average on average, and then various different groups to being locked in, in terms of their propensity to sell their home? About every percentage point of lock-in reduces your likelihood to sell by about 18.1%. People in more expensive homes are more sensitive than those in less expensive homes. After controlling for affluence across the expensiveness of your house and income and credit score, et cetera, non-white borrowers — Asian, Hispanic, and Black borrowers — are all more sensitive than white borrowers. For those three non-white groups, it’s about 22%, versus like 16% for white borrowers. 

The third point was, can we get an aggregate effect? That’s just applying those sensitivities to the path that mortgage rates have taken and the path of the stock of existing mortgages, and we find that 1.3 million sales have probably been lost due to this since rates started ticking up in 2022. That’s getting added to every quarter now, 200,000 to 250,000 each quarter. 

The final point is that the supply reduction due to this lock-in effect is dominating the direct effect of interest rates on home prices. Since we’re coming off such low rates, the increased rates are not only reducing demand, but they’re creating this lock-in which is decreasing supply. At least in the short run, we’ve shown that the supply effect can dominate. The best we can tell is it’s something like a 5% or 6% increase in prices due to the supply reduction compared to only about a 3% direct-effect decrease in prices due to the higher interest rates, that demand effect.

Q: Were there certain loan types or loan products that the lock-in effect becomes more pronounced with? You’re talking about the higher-priced homes, can you attach that to conventional or FHA borrowers being more or less sensitive to the lock-in effects?

Jonah Coste: We don’t see a lot of heterogeneity across things like loan attributes. Now, we do see very different baseline rates of sales across these things. A home with a 15-year mortgage has a higher baseline propensity to sell than someone who would buy it with a 30-year mortgage. But, both of them decreased by about 18% for each percent of lock-in. The heterogeneity we really see is more on borrower attributes, and like I said, affluence is a big one, and it seems that of the measures of affluence, it really seems that home value is the biggest driver there. You see it show up a little bit with credit scores and with income measures, but the effect is largest on home value. And then the second thing is race, where after accounting for those measures of affluence, the non-white groups seem to be much more sensitive than the white group. 

Will Doerner:

If you break it down by loan type, like what you’re asking, we’re still getting a really strong effect across all of those segments. It’s not something that disappears if you don’t focus on Fannie-Freddie loans if you focus on other loan types. It’s not something that disappears if you focus on borrower incomes and different kinds of affluence of borrowers or their ages or how much DTI or how much down payment they’re putting in — across all of those categories you’re going to get well over 10% effects. When Jonah was talking about that 18% effect, most of the cuts that we have are above 15%. That’s why we think this is such an important paper to bring up because we can get at this nationwide, not just with Fannie and Freddie loans, but with loans across the country and talk about how different segments are impacted differently.

Jonah Coste:

In theory, FHA and VA loans are assumable. That could mitigate some of this lock-in effect because you can capitalize some of the value of your below-rate mortgage and into your selling price. There’s two problems with this. One is research that predates ours that shows that you can actually only recoup about a third of the present value of your below-market-rate mortgage. So, even if you were able to find a borrower that could qualify to take over your mortgage and you successfully navigated that process, you’re only going to get about a third of the value, so you’re only really decreasing your exposure to lock-in by a third. The second problem is that servicers are putting up all kinds of roadblocks to actually getting that done, to the point where in all of 2023, a total of 6,400 assumptions were processed by the FHA and VA combined. Over that same period of time, we’re looking at well over 200,000 lost sales in that group.

Q: Yes, if servicers could figure out what that requires on their end, then there’s opportunity. I don’t know if that question has really been answered by the servicers yet. I’ve talked to other people who’ve said that assumptions are not as complicated as people think.

Michael Seiler

The other thing no one mentions though, is if assumable loans were more likely to be assumed than you imagine, then they would have to increase the interest rate in the first place. If a bank is willing to loan money at a fixed rate for 30 years, that really puts them on the hook. Part of the reason why that is not a much higher interest rate is because people don’t hold them to that task for 30 years. But, if loans that are assumable actually got assumed over and over and over and over again so that they had a 30-year maturity, then the duration would be higher, the risk would be higher, and ergo you’d have to charge a higher contract interest rate on the mortgage.

Will Doerner

We modeled a couple of different ways of thinking about what is the lock-in effect, and then asked, how long do we think it’s going to be around? These kinds of estimates haven’t really been done in the past in terms of trying to describe to what extent we think this is going to continue. There have been folks who have speculated about if rates were to decline 50 basis points or 0.5%, or a percentage point, then we’re out of this challenge. Our takeaway is that the lock-in effect is going to be here for a while. If rates begin to decline pretty substantially even, we’re still going to see it around for another five to 10 years at minimum, just because it’s become such a prominent problem that hasn’t been really talked about or explored a whole lot.

Michael Seiler

It’s not like this rate delta, the difference between the rate on your mortgage and current rates, occurred 20 years ago. This occurred very recently. You can’t just say, ‘Oh, we can just be patient, time will pass, loans will fully amortize, and then that will be gone.’ I mean, that’s a super long time to wait. So, that’s not the solution out of the lock-in problem at all.

Q: What are some solutions? No one’s got a crystal ball, but what could be a path forward, or at least various paths?

Jonah Coste

We are careful. We’re neutral on potential policy solutions, but we do go into some ways that other markets are structured internationally or at different periods of time in the U.S to avoid this going forward. Some things could help, like tax credits to sell your home is one thing that’s been proposed. We’re not taking a stance on this, but we do think our research provides a lot of tools for someone who is trying to scope out whether that would be a good policy, to understand if it is sized appropriately, what the costs would be, and how to effectively target it. Going forward, thinking about different ways we can structure the mortgage market — assumability is one, but does suffer from the problems that I outlined. Portability is something that could potentially have a higher uptake and also would avoid the problem of only being able to recoup a third of the value on an assumption. But, much like the assumable mortgage, you have to think about what the general equilibrium effect is on rates. The effect would be a higher interest rate being offered, particularly in times of low rates because those would be the mortgages most likely to keep getting ported, ported, and ported, against the desires of whoever owns the mortgage. Both assumability and portability exist in other countries, and more adjustable-rate mortgages (ARMs). We show empirically that adjustable-rate mortgages are not subject to this effect. 

Michael Seiler

If we’re talking about servicers not wanting to play ball on assumables, I’d be curious what people think about loan originators. If you’re going to port mortgages, it would seem that people wouldn’t need nearly as many originations. How is that going to fly in the industry?

Q: That’s a great question. I don’t think anyone’s asking that on account of portable mortgages not getting much uptake. Is my understanding correct that the number of lost home sales compounds over time, simply becoming more and more lost home sales?

Jonah Coste

There’s fewer sales every quarter, which adds to that cumulative number. The question is, does the effect fade over time or does it create sort of a pent up demand that when rates do come down will manifest in a big surge in home sales? This is pretty unprecedented. It’s very unprecedented in that timeframe. The last time you could maybe see this was back in the early 1980s, rates rising this much in a short period of time. Of course, they were rising from an already higher level then. A three- or four-percentage-point rise from like 10% to 14% is probably not as big of an effect as the rise from 3-7%. So, this might be truly unprecedented. What we can see from lower levels of lock-in is very little evidence of either a fading effect or a creation of pent-up demand. To the extent that those will eat away at the cumulative effect, they’ll do so even slower than the natural process of loans maturing and new loans getting made at the higher rate that we explicitly model out. This does seem to be every quarter we’re seeing more and more lost sales that are probably just lost. These are moves that didn’t happen, a family that didn’t downsize or whatever it may be.

Michael Seiler

Jonah’s talking primarily about price effects, but imagine that you are a young family and you are expanding. You feel like you’re locked into your home, so maybe you’re getting squeezed, wish you had greater square footage, but you don’t really want to change that. On the other side, imagine your children are grown up, going to college, or you’re an empty nester. Maybe you want to downsize your home, but again, don’t want to sell it. We call this not being right-sized in your home. From a pricing perspective, everything we talked about, that’s great. But from an overall utility perspective, people are not necessarily living where they want to live and maybe even working where they want to work. It has some very difficult to quantify, but absolutely important, life impacts as well.

Q: At the same time that those sales are lost, would you still expect home price appreciation to continue on account of still-restricted supply and still-decreased demand?

Jonah Coste

You can’t necessarily translate the effects that we have into a projection because this is sort of an ‘all else being equal’ scenario. There are all sorts of other things that are going to drive national and certainly local home prices. What we can say is that the demand effect in the earliest quarters, at least, is being swamped out by this decreased supply, and the decreased supply doesn’t look like it’s ending anytime soon.

Q: You’ve been using the word this phenomenon, which implies a certain amount of unpredictability around this issue, that we couldn’t have seen it coming. Could some kind of lock-in effect not have been anticipated as a consequence of everybody purchasing or refinancing into rock-bottom rates, and then being unable to afford to move when rates go up seemingly overnight by three or four percentage points?

Jonah Coste

I definitely don’t want to speak for everyone that no one saw this coming. I also want to caution that even if it was well understood that this was going to happen, this still could very well have been the right thing to do. Putting interest rates towards zero and other monetary actions taken at the time were not done with the idea of what’s going to lead to the healthiest housing market in the short and long terms. They were done for other macroeconomic reasons. Monetary policy is a blunt tool and will affect different sectors differently. Ultimately it’s the aggregate economy that policymakers are looking at. So, I don’t know that even with perfect foresight, that anything should have been done differently. That’s sort of outside of my department. 

Michael Seiler

You also have to kind of pick your poison because if it had not been done, then you’d ask, ‘Why wasn’t that done?’ We would be asking if it’s possible to defend the inactivity, the inaction of that. It’s not like there was a course of action that would have solved all problems and had no unintended consequences. Like Joanh said, even with perfect foresight, it's not like you can take a particular action and have nothing negative come out of it.

Will Doerner

When we’re looking at the monetary actions that happened, particularly during Covid, had the Fed not acted very fast and on the fiscal side, with just the regular federal government combating what we were going through with the pandemic, we almost undoubtedly would’ve been in a much worse situation. That refinancing that actually got people at this rate and started to create the problem actually might have been a stimulus of sorts because when people have this lower mortgage rate that they’ve been able to lock into, they all of a sudden have more money in their pocket, they can spend it and create aggregate demand that stimulates macroeconomic activity. The mass refi wave that went through that period actually probably helped stabilize the economy, and housing, by and large in the early part of the pandemic, was a savior for a lot of the economic challenges we were going through.

Jonah Coste

To the extent that our paper speaks to policy discussions and decisions that could be made differently, I think it has less to do with the decisions made in 2020 and 2021 and more to do with very slow, long-term decisions we’ve made about how we structure our mortgage markets dating back one hundred years in some cases, how we’ve ended up with this non-assumable, non-portable, fixed-rate, 30-year mortgage dominating our market. Now that we understand this is a phenomenon that could happen with that, do we want to make any changes to it? Could we add some policy solutions so the next time that we need a dramatic decrease of interest rates to combat a crisis, there’s not a big hangover from that, to combat inflation without hurting the dynamism of the economy and people’s ability to move for work and live in the types of places they want.

Q: All these homeowners we’re discussing are in the best mortgage financing of their lives. It’s many of those folks who are finding it increasingly difficult to afford those mortgages, though, because of the rising flexible costs of homeownership, like property tax increases, insurance costs, and utilities. By trapping homeowners in their properties, lock-in effects also trap homeowners in these rising flexible costs. Have you studied this consequence?

Jonah Coste

In economics we have a name for this, at least the property tax piece of it. The “Tiebout mechanism” is where people can vote with their feet and select the municipality that is offering them their optimal bundle of services, like a tax rate. That’s hampered by these effects. You’re trapped into the services you have now because yes, maybe you can save a few hundred dollars a month on your property taxes somewhere else, but the only way to do that is to pay $700 more every month in interest. That’s no longer a trade-off you’re going to make.

Q: For mortgage bankers, decision makers, operations teams thinking about this on a daily basis, how does your paper impact their planning and how they should be thinking about the market? Is there anything that you think is important to drill down on?

Will Doerner

I was talking with some academics recently and they were saying that they can’t even attract high-level professors to come to the university because they’re not going to give up their mortgages. One of the things they did in the late 1980s and early 1990s was universities would actually offer to pay down part of the mortgage rate for whoever’s coming in to kind of attract them to come there. We don’t have all the solutions and that’s why we wanted to write this paper, to begin a policy discussion and to begin a public discussion. That’s where we could use the feedback from the industry thinking through this. There are folks who probably have plenty of ideas in terms of what they would like to see or what they would think would be possible.

Jonah Coste

Let’s say we wanted to adopt things like portability, what would be needed to do that? If we enter a period of low rates relatively soon again, could the market provide that as a feature? Would people be willing to pay a slightly higher rate in a period of low rates for it? Does it need to be something that can be sold to the government sponsored enterprises? Do they need that policy to exist, that Fannie and Freddie will buy a portable loan and have procedures in place? We can talk about the economics of it. We can model out how rates would need to be higher or the appropriate fee to charge for reporting a mortgage. But, in terms of the institutional details, like with the assumables, you can see with the FHA and VA assumptions right now the incentives just aren’t there. I don’t really buy that it’s an issue of complication. It’s really an issue of the incentives not being there for the servicers. They want these loans actually to prepay, or the holder at least wants the loans to prepay. Making it easier to hand it over to someone else is really not in their interest, and certainly the small fees that are associated with it aren’t enough to overcome that. That’s where industry folks have a whole lot more insight.

Michael Seiler

The last point that I’m reminded of is we’ve heard, ‘Why don’t you just build your way out of it, create more supply that’ll push down prices?’ Think about the timing of doing that. Then they’re like, ‘Well, just take old commercial real estate and rezone it.’ You’re talking about a massive number of years and money. So, building out of it is not a short-term solution. People are being creative, but it really is an issue that’s here to stay.

And the cost of building right now, from materials to labor to new inspection codes for environmental resilience, means the cost of a new home doesn’t make that an easily affordable option for many buyers.

Will Doerner

One of the challenges in the U.S. is that a lot of the places that are most expensive and have the highest amount of lock-in effect, if you’re looking at places like California and Florida in particular, they also have property taxes that are capped. That creates a lock-in effect there, as well. So, people haven’t moved in Florida in some locations because they’ve accumulated this really large gap. When you have mortgage lock-in effects on top of property tax lock-in effects, it just really compounds in these markets that are having high house price appreciation.

Jonah Coste

Obviously your readership is very mortgage focused, but even people who are outside of the mortgage space or outside the finance space, policymakers and decision makers of all stripes need to be thinking about these knock-on effects. If you’re someone who’s doing a nationwide recruiting search, this may be adding to your costs, relocating someone. Being in Wasington, D.C., there’s a very steady migration of people that once they have kids at school age, a good percentage of them move out of D.C., they move to the suburbs. Is that not going to happen to the same extent? What does that do to enrollments, demographics, and school districts? I wouldn’t be surprised to see that change and have people with nothing to do with mortgages or even housing grappling with this effect that is going to touch a lot of different parts of policy.