Build-to-Rent and Housing Supply & Demand

Brian Potter wrote a post at “Construction Physics” that is thorough and informative, as is typical of his writing. Here I will discuss his post a bit, and note how some details about investor activity and the build-to-rent market fit well with my general position about the housing market: that (1) the most important factor in housing market trends, still today, is the sudden exclusion of 10 to 15 million households from mortgage access in 2008, and (2) the primary remaining consequence of that is that we permanently lowered the American capacity to build new homes by more than 1 million units. A drop that we are still slowly growing back out of.

Related to those points, and underlying them, is an observation that is so obvious it shouldn’t need to be stated, and yet is at odds with practically all the writing about this issue - both lay and academic: Funded homeowners systematically pay more for single family homes than investors do. Always and everywhere. This is why there was never a build-to-rent market outside of small niches until the last few years. It’s why there were never any large firms in the single-family rental business before 2012. It’s why the scattered mom & pop owners of single-family rentals mostly owned older depreciated homes in low-tier neighborhoods where tenants didn’t have the wherewithal to buy their homes.

If your research question is “Are large investors pricing families out of housing?” Your strong presumption before filling a single cell of your spreadsheet should be “No.” And, the reason for your answer should be “The entire history of modern housing economics. Duh.”

It’s sort of an odd situation because you might collect data on the historical trends of large investor activity and their correlation with home prices, and if you did, you would find that investors overwhelmingly are outbid by families for housing. Except, you can’t find that because the case is so overwhelming that there is no historical data of large-scale investments into single-family housing. Now that we have some data, it might seem like an opportunity to freshly ask the question. But, the lack of historical data already answered the question.

You might say, “But any source of demand will raise the price.” If I pick up a box and hold it over my head, and then a 5 year-old comes and stands next to me to try to help push it higher, does the extra “demand” from the 5 year-old to push the box higher, definitionally, mean that there is more “demand”? No. The 5 year-old is a rental investor in the history of American housing markets. The 5 year-old might help keep the box from falling all the way to the floor if I happen to break my arm. They don’t under normal circumstances.

This is like asking the question, as Federal Reserve researchers did in 2012, “How much excess housing production occurred during the boom phase of the cycle and how far along is the correction?” There is no answer to that question. The answer lies outside the set of answers that question will accept. There was never an excess of homes, and there certainly wasn’t by 2008 or by 2012. Questions can presume their set of answers, and when the real answer lies outside that set, only confusion can ensue.

Anyway, on to Potter, who tends to get things right.

The modern BTR industry, where developers build entire communities consisting of dozens or hundreds of single-family homes for rent, is a product of the 2008 Global Financial Crisis. Prior to the financial crisis, single-family home rental wasn’t uncommon — in 2005, there were over 8 million detached single-family homes being rented — but the business was mostly the purview of small “mom and pop” operators that owned a relatively small number of scattered rental properties. As late as 2011, no single company owned more than 1,000 rental homes in the US.

But the financial crisis shifted the housing landscape. Huge numbers of people lost their homes to foreclosure: foreclosure rates in 2009 and 2010 were four times rates from 2005, and between 2007 and 2010, there were four million foreclosures. The homeownership rate in the US fell from a high of 69% in 2005 to 63% in 2016. At the same time, to rein in the subprime lending that had precipitated the crisis, banks tightened their lending standards, and average mortgage credit scores rose by more than 50 points. In 2003 buyers with a credit score of less than 620 made up 7% of all mortgages. By 2011 that had fallen to essentially zero.

The raft of foreclosures and the tightening of lending standards had two simultaneous effects on the housing market.

First, they pushed millions of Americans into renting. Between 2010 and 2015 the number of renter households in the US rose by roughly six million, while the number of homeowner households declined by roughly 800,000.

Second, this shift created a huge pool of homes available for purchase at very low prices. Between 2006 and 2010 the value of US homes dropped by 26%, greater than the average decline during the Great Depression.

Potter goes into a lot of interesting details about the different companies that started to dip their toes into the build-to-rent market as existing home prices re-inflated. He estimates that today there are about 350,000 single-family homes that were constructed to be rentals - about 1.5% of the single-family rental market.

He also goes into details about the various physical forms that build-to-rent neighborhoods can take.

Potter concludes with:

It’s clear that many folks strongly believe that large-scale corporate ownership of rental housing (which would include BTR communities) is something that can have negative effects on the housing market. But I think it’s more useful to think of the popularity of rental housing as something that’s a product of the housing market: it’s a natural consequence of housing getting increasingly unaffordable thanks to high interest rates and skyrocketing housing prices. Shutting down BTR is a poor way to address that problem; what we need to do is build more housing and develop construction methods that let us construct buildings more cheaply.

Absolutely. This is correct, and better stated than 90% of what you will read on the topic.

He notes that the literature on the effect of investors on rents or prices is mixed, but that the literature really has only looked at investors in existing homes. Building new homes is likely to have more of a direct effect on supply, and so it would surely help to lower both rents and prices, unless build-to-rent is crowding out owned homes that would have been built, anyway.

I think we are currently supply constrained, nationally, and so, the crowding out story is actually plausible under current conditions. But, this goes back to my comments above. If there is crowding out, it is from families. Your intuition should be that where build-to-rent is increasing, it is because families are crowding out investors less than they were before.

Just turn the conventional story upside-down and you’ll have a big advantage over practically anyone who is trying to understand housing, and you’ll even have a bit of an advantage over someone as smart and accurate as Brian Potter.

I have a few points to make on that, and some pedantic nitpicks of Potter’s post.

First, a nitpick. Potter writes:

But while affordability issues seem to be the primary driver of BTR’s popularity, there also seems to be some fraction of residents that simply prefer renting over owning, due to a desire for less maintenance or simply because they don’t perceive owning a home as a major life goal. CBRE, NAHB, and NexMetro all mention various demographics of “renters by choice” (such as retirees), an analysis echoed by several BTR developers I talked to. As construction of BTR communities continues, this growth might create a sort of reinforcing cycle: more people move into rental housing, which makes it more accepted, which draws even more people in, and so on.

There is some truth to this. There are new forms of rental housing available now that weren’t before, and forms of productivity and innovation in the space that would possibly lead to some permanent increase in rental housing even if mortgage access fully returned to pre-2008 norms. But, I think it is broadly overstated, and it’s overstated because observation is biased.

There is a similar issue with homelessness. There are many more homeless people today than there used to be, because of the housing shortage. But, even many homeless advocates will push back. “No. I work with these people. They have problems. Those problems are why they are on the street.”

Here’s a picture I use to make this point. Observationally, the population of unhoused people in both of these scenarios will look similar and will have similar problems. In both cases, the typical homeless person will have personal issues that aggravate their condition. It is very hard to observationally see why there are more homeless people.

There is a similar thing going on with renter households. There would have been 40 million renter households, and they would have had any number of reasons for wanting or having to be renters, including some number that were on the margin, and could go either way. Then, after the mortgage crackdown there were 50 million renter households, and they had any number of reasons for wanting or having to be renters, etc.

A property manager will tell you sincerely - even the renters will tell you sincerely - that observationally and personally, those renters generally have a preference for renting, and if there are 10 million more of them, then, observing the average of the 50 million will suggest that there are just more households now that have reasons for being renters.

It’s just not a claim that you can make through observation. There may be marginal changes in preferences that have moved in this direction, but we wouldn’t be having a debate about the sudden rise of this new asset class if it was just about the scale of change that preferences would have caused.

Potter notes that large buyers started buying up properties around 2010-2012. A few firms built new rentals, but the main activity was in existing homes. That’s because the mortgage crackdown made them too cheap, because families - who always are the price drivers - disappeared from the market.

You can see a common confusion in a paper Potter cites from the Philadelphia Fed, which notes “the increasing presence of institutions in the housing market explains over half of the increase in real house price appreciation rates between 2006 and 2014.”

That’s a weird way to put it. Figure 1 is a chart with the Case-Shiller home price index and an estimate of the change in the number of homes that are owner-occupied. [Each year, I subtract the number of new homes built and add the number of new owner-occupier households. The result is cumulative estimate of the number of existing homes that switched to owner-occupied (up) or to investor-owned rentals (down).]

It’s weird to talk about price appreciation during that period. At most, more investors after 2011 might have been associated with the partial reversal from a deeply underpriced market. Before that, at most, investor activity might have been associated with a collapse that wasn’t as sharp as it would have been.

In any case, this was clearly a market with unprecedented price depreciation that suddenly lost a ton of purchaser demand. Surely we can agree that a straight red line from 2006 to 2020 would have been a better outcome than the line we did follow. But, more importantly, this language betrays the confused approach that is common on this topic. Somebody has to own each home, and if one category of owners declines, the other categories will have to rise. Associating the rise of the other categories with price appreciation when their rise wasn’t nearly responsive enough to keep prices from cratering and defining the economy for a generation is deeply confused.

Figure 1

That’s not Potter’s fault. There isn’t an unconfused literature to draw upon. That paper is actually not bad. They find mixed results. Some papers from geography or sociology departments just begin with an implicit presumption that large investors are bad, and whatever they find is described in a negative light. If they buy homes at a discount, they are stealing equity from families. If they buy homes when prices are rising, they are pricing families out. The Philadelphia Fed paper finds some trends that it characterizes as positive and some that it characterizes as negative.

But, still, the presumption that investors are a source of power, moving housing markets, creates confusion, even when the research isn’t stacked to be negative.

Maybe a useful analogy is the sun and the moon. A full moon is brighter than a crescent moon. So, you could say that the moon provides light. But, the topic only has meaning when the sun isn’t in the sky. A full moon is really helpful to have when the sun isn’t in the sky. Would it make sense to ask if the moon is crowding out the sun? Would it make sense to note that, since the full moon makes the night brighter, that a new moon might be making the days too bright?

This is the conversation we’re having. It’s not coherent. And, yet, if you model the moon as a provider of light, which it certainly is, and conceive of it implicitly as a source of power, a competitor, and an outsider, it would be hard to believe that there was any other conversation worth having.

Potter writes:

In response to these market conditions — millions of homes available to buy cheaply, and millions of Americans who couldn’t afford to buy them — various real estate ventures were formed to take advantage of the situation.

“Couldn’t afford to buy them” is technically true, I guess, as a household who has lost access to mortgage capital won’t be able to afford a house, but this is a tic I find in the literature, in general. Acknowledging the credit crackdown, but then proceeding as if the loss of homeowner demand was some sort of economic state of nature, as if the country wasn’t loaded with existing homes for the next decade where families were paying $1,500 rent for homes they could have purchased with a $1,000 mortgage payment.

In more recent years, affordability has been more of a constraint, since home prices corrected back up until existing homes were selling at prices that builders could compete with. Rents had to rise for that to happen, so the affordability problem is a rent problem as well as a price problem. In some markets, where land rents became inflated because the American homebuilder industry hasn’t been able to regrow capacity quickly enough to build homes in inflated locations, rent inflation leads to more price inflation. (For each 1% increase in rents, prices rise by more than 1%.)

You can see that at work in Figure 2. Median income (black) has risen faster than general inflation (blue), but it hasn’t risen faster than rent inflation (orange). That’s bad. For 30 years, as incomes have grown rents have grown along with them. This isn’t symmetrical. Rents are outpacing below-median incomes and are more normal in neighborhoods with above-median incomes.

But, even looking at the averages and medians, part of what is happening here is that higher rents lead to much higher prices. The 2005 price peak was about half and half a cyclical/credit boom and the inflationary effect of rising land rents. Today, the elevated price level isn’t cyclical or credit-driven at all. It’s all inflated land prices due to the scarcity of urban housing. More or less, the difference between the orange line and the blue line is causing the difference between the black line and the red line.

Figure 2

This does cause an affordability constraint for some buyers who are trying to buy homes with a highly leveraged mortgage. It is harder to buy a home whose price is inflated by inflated land rents, because the price/rent ratio on those land rents can be north of 30x. This surely has marginally reduced homeownership in the last couple of years. But, it really doesn’t affect the fundamentals of home prices much. Prices have flattened, but high interest rates aren’t going to bring prices back down. Only more construction will do that.

But, again, what happened when homeownership slowed down? In Figure 1, after prices were driven higher for nearly a decade while homeowners reclaimed the stock of single-family homes, prices stopped rising when more single-family homes were once again available for investors and renters.

And, here is another issue where understanding that investors are, first and foremost, described as an absence of family buyers. This is a subtle point that it is hard to catch. Potter notes that the rise of build-to-rent activity was related to this affordability issue - families on the margin of mortgage access have had a harder time buying since mortgage rates increased.

I’d say that the typical housing economist would say the same thing about build-to-rent - that it was spurred by low interest rates and that high rates would slow it down.

Both can’t be true - or, rather, if both were true, then new single-family home completions would be very volatile, like they were in the 20th century. But they aren’t, and they aren’t going to be. Completions have been flat as a pancake, and, short of some cataclysm, the only significant trend shift from here will be upward, regardless of what interest rates do.

So, what is it? Did high mortgage rates reduce homeowner household formation or build-to-rent activity? Potter is right, and the housing economists are wrong. In Figure 3, the black line is the number of owner households formed and the blue line is the number of new homes built. The cumulative difference between the black and the blue lines is the black line in Figure 1 (the cumulative number of homes transitioning between owned and rented). The red line in Figure 3 is the estimated number of new build-to-rent single-family homes.

Figure 3

The entire growth of the build-to-rent market has coincided with high mortgage rates since 2021. Builders are at capacity. Where homeowners are buying, they claim that capacity. Where homeowner purchases decline (and high mortgage rates probably were responsible for some of the temporary decline after 2021), more of that capacity is available for investors and renters.

Figure 3 also shows build-to-rent completions in orange, on the right scale, to highlight patterns. It is recently a mirror image of the trend in homeowner household formation. But, it’s a much smaller scale. Ten fewer homeowner households might be associated with 1 more build-to-rent completion. And, maybe the sharp drop in build-to-rent completions in the most recent quarter is related to a new resurgence of homeowners. This data is a bit noisy, so only time will tell.

Generally, as capacity to complete new homes rises, the new homes will marginally be rentals because those are the households that haven’t been forming (because investors get outbid by families). But, if there is some change like a sudden decline in mortgage rates that temporarily raises demand from potential homeowners, expect investor activity and build-to-rent activity to pull back a bit.

This makes sense if you understand that under our current capacity constrained condition, investors are mostly the absence of homeowners. They are the full moon, that, if we allow them, can help illuminate this market in a dark time. They aren’t the cause of things. They are, in the aggregate, the absence of things. All the papers Potter cites - because it’s the only kind there are - are trying to determine if the full moon is crowding out the sun. They are trying to measure how far out of my reach the 5-year old is pushing the box. They are confused.

If we can manage to allow either multi-family or single-family rentals to grow from here, over the next decade or so, there will be a new set of data that will thoroughly associate more investor activity with lower rents and prices. In a way, that will be a spurious correlation. It will be the result of more building, and at some point, the capacity to build will finally outpace the potential growth of homeowner households. So, rising supply will be associated with rising investor activity and with more affordable housing.

The evidence will be overwhelming, if we can manage to allow it, and if anyone is willing to ask a question in which the correct answer is allowed to exist.