By Morris A. Davis
For the past 20 years, downtown city living has undergone a renaissance. A seemingly endless influx of millennials with groovy facial hair has driven up housing rents and prices in many areas, and new development has had difficulty keeping up with demand.
This multiyear transformation of our downtown areas has led to speculation that the single-family, suburban model of development is dead. In this article, I will make a different prediction. Today, I predict that the next great housing boom will be in the suburbs. The millennials will move out of the city and into the suburbs and will commute farther distances, by car, than previous generations.
There are two major, profound forces at work that lead me to this conclusion. The first is that the millennials will start having children soon, and there are a lot of millennials. The table below shows the age distribution of the population in the United States in 2015. As you can see, the single largest adult demographic is the group aged 20 to 34. Over the next 15 years that group will begin to have children and research has shown that the strongest prediction of homeownership is having a child. If these millennials wish to buy a single-family house in a low-crime neighborhood with good schools, then where will they buy and who will they buy from? As the table shows, there aren’t enough people currently over age 65 to sell to them. Presumably, some of the people currently age 50 to 64 will need to sell. I guess some market participants and planners are gambling that a location swap will occur: Retirees will sell their houses to millennials and rent the units the millennials abandon. That scenario seems unlikely to me as research has shown that previous cohorts of seniors tend not to downsize. We might need to build a lot of new single-family homes to accommodate the millennials once they start having children.
The second major force in play is that a transportation revolution is at our doorstep — the self-driving car — and it will profoundly affect how we commute, where we live and what we are willing to pay for land. The theory is simple and arises directly from the canonical model of urban economics. Imagine a metro area where employment occurs at the center of the metro area and people live outside the center. All else equal — schools, crime, parks, etc. — the value of land reflects the cost of commuting to the center to work. The more costly it is to commute, the cheaper the land and housing. At some distance to the city center, commuting becomes so expensive that the land is valueless or nearly so. This distance defines the metro boundary. Urban economics predicts a “gradient” on the price of land that reflects commuting costs: Holding school quality, crime, etc. all constant as best as possible, the price of land starts out high near the metro center, in areas with low commutes, and gradually falls to the price of farm land at the metro boundary as commuting costs rise.
Commute costs include physical costs and time costs. In what follows, I assume the physical cost of commuting (gas plus wear and tear) is $0.55 per mile, which is approximately the IRS reimbursement rate per mile driven. But the more important commuting costs may be time costs. People that are commuting cannot work (except for phone calls) and cannot enjoy leisure (except for the radio). How costly is this? Consider a person making $80,000 per year living 15 miles from work and commuting 30 minutes by car each way. Each day, the person spends $16.50 on the physical cost of commuting and $20 in time costs (lost work or leisure opportunity), assuming the person values their commuting time at one-half their hourly wage. If the person works 250 days per year, the annual total cost of commuting is $9,125. The driverless car changes this equation by making time spent in the car productive time, either at work or leisure.
If a person does not mind riding in a driverless car — i.e. suppose they can work or enjoy leisure, such that the time cost of commuting falls to $0 — then the annual cost of commuting falls by $5,000, from $9,125 to $4,125.
This drastic reduction in commuting costs ultimately acts like a supply shock to the quality and quantity of land. Land where commutes are currently barely manageable will become more desirable because people can now enjoy their trip to work or be fully productive while driving. Once the driverless car has widespread adoption, more people will be willing to live relatively far away since commuting will not be as costly. Directly related, the value of land will change everywhere because commuting costs will change everywhere. The “gradient” of the decline in land prices from the metro center to the urban boundary will change: The value of the land that is farthest from the city but currently commutable will increase relative to the price of land in close-in locations since commuting costs will fall the most in far-away locations. A back-of-the-envelope calculation suggests the value of land where our hypothetical commuter currently lives might increase by $100,000 — assuming a cap rate of 5 percent on rent — assuming the driverless car reduces commuting costs by $5,000 per year. When I ran this idea past my colleague David Frame, he additionally conjectured that the value of land and housing near the metro center might outright decline due to the fact that the millennials will want larger plots of land that are only available in the suburbs.
There is also speculation that the driverless car will make commutes shorter. The theory of this builds on areas outside of my expertise, but as I understand it, (a) cars will be more tightly packed on the road, increasing available roadway space and (b) getting into and out of Manhattan or any densely packed employment area will take less time because driverless cars will not occupy valuable roadway space to queue up to park, which is a cause of current congestion. These points seem logical to me. As a force in the opposite direction, however, I expect there to be more cars on the road once the suburbs expand.
Carl Goldberg and I discussed this analysis and he asked me to consider two ideas that might serve as caveats. The first is that not all suburbs are created equal. Carl feels — and I suspect he is right — that suburbs that offer more “downtown-style” living, such as Montclair, will be more attractive to millennials than suburbs that are zoned with large lots and require a 15-minute drive to do anything useful, like buying groceries. The second is that house prices in many suburbs are still too high for the millennials to purchase: Their incomes are too low and they have too much student and other debt. This issue of affordability is very real and it is correlated with the lack of density and lack of downtown-style living options in some suburbs. Poor planning has created an affordability problem, and a desirability problem, that we need to fix soon.
So where does this all leave us? Let me sum up. Many of the millennials are of the age where they are going to want to own a single-family home in the near future. There are currently not enough housing units in easily commutable areas to accommodate all the millennials. Once the driverless car technology is finalized, the supply of commutable land will increase, the cost of commuting will fall and the price of land in the suburbs will rise relative to the price of land in close-in locations. With appropriate caveats, this will enable a building boom in the exurbs — if state and local governments allow it — and millennials will buy farther out and have longer commutes than previous generations.
The suburbs are back. Count on it.
Morris A. Davis, PhD., is the Paul V. Profeta chair in real estate and academic director at the Rutgers Business School Center for Real Estate.