[The author is] suggesting that, while cities continue to fight the battle for development in dense hubs, they also question the de facto exemption granted to low-density suburban areas from the onus to produce more housing. The dormant suburban sea is so vast that if the taboo on densification there were broken, even modest gradual redevelopment – tearing down one single-family home at a time and replacing it with a duplex or a small apartment building – could grow the housing stock immensely. Distributing the necessary amounts of new housing over vast low-density suburban areas instead of just concentrating them in dense hubs would dilute the local impact on neighborhoods. It would make a large increase in housing more palatable vis-a-vis neighborhood character, and more gradual. Of course, building in these areas could have different implications for congestion than building in dense hubs, but the affordability crisis in America’s expensive coastal cities is so acute that the tradeoff between worsening affordability and congestion should be evaluated with fresh eyes.

In order to nurture new residential development in the dormant suburban interior, local land use policy would need to undergo a revolution. The construction industry and the financial ecosystem would need to evolve as well, and infrastructure would need to be greatly upgraded. The very first step, however, involves grasping America’s new metropolitan landscape and realizing just how much of it has gone dormant. That is where the problem is, as well as the opportunity.

....

In “The Great American Single-Family Home Problem,” New York Times reporter Conor Dougherty drives home the point that the high cost of housing in America’s expensive coastal cities can be addressed by densifying the suburban landscape. At first glance this may seem obvious; combatting price appreciation calls for building more homes, and better to do so through infill than by exacerbating sprawl. But the notion of densifying the suburban landscape is a more radical departure from the status quo than it may seem.

In the past, virtually every patch of land in the metropolitan U.S. continually sprouted new housing, but this is no longer the case. In recent decades, residential construction has become increasingly confined to the periphery of American metro areas, while a growing swath of the interior has fallen dormant and produces new homes at a negligible pace. At the same time, a tiny fraction of the land area, scattered in small pockets throughout the metropolitan landscape, is responsible for a growing share of new home production, primarily in large multifamily structures. I refer to this increasingly spiky new pattern of housing production as “pockets of dense construction in a dormant suburban interior.”

The story is best told through maps

The following map colors the residentially developed footprint of the Los Angeles metro area as of 1960 according to the dominant type of new housing built there during the 1940s and ‘50s. It distinguishes between single-family homes in light blue, 2 to 49 unit structures in orange, and 50+ unit structures in red, and assigns each Census tract a color based on the type of housing that accounted for the most new homes during the period. 

The map reserves an additional color, dark blue, for areas that produced no new housing at all, or that produced it at a negligible pace, defined as less than 0.1 new homes per acre per decade. This pace is equivalent to less than one new home per acre per century, which means that transitioning a neighborhood from a stereotypical suburban density of 4 homes per acre to a borderline walkable one of 10 homes per acre at this rate would take more than 600 years.

....

In the 1940s and ‘50s the map was mostly light blue, indicating that the vast majority of the metro area’s developed footprint was producing new housing at a meaningful pace, mostly but not exclusively in the form of single-family homes.
In the 1940s and ‘50s the map was mostly light blue, indicating that the vast majority of the metro area’s developed footprint was producing new housing at a meaningful pace, mostly but not exclusively in the form of single-family homes. - Although a number of areas show up in orange, implying a local abundance of moderately sized multifamily structures, only a tiny fraction show up in red, and even fewer show up in dark blue.
By the 1960s and ‘70s, light blue areas still accounted for most of the developed footprint, but they were losing ground.
By the 1960s and ‘70s, light blue areas still accounted for most of the developed footprint, but they were losing ground. - A growing area saw new housing dominated by 2 to 49 unit structures, and the number of red Census tracts in which large multifamily structures accounted for most new homes increased as well. In addition, a scattering of Census tracts begin to show up in dark blue. Some of them – those which appear undeveloped (white) in the previous map – indicate extremely low-density residential development in formerly undeveloped areas, but most of them reflect previously developed suburban areas whose new home production essentially stalled.
The pattern of pockets of dense construction in a dormant suburban interior began to emerge more clearly by the 1980s and ‘90s.
The pattern of pockets of dense construction in a dormant suburban interior began to emerge more clearly by the 1980s and ‘90s. - Much more of the metropolitan landscape was dormant by this period, showing up in dark blue, while clusters of dense new home construction appearing in red had multiplied. It is visually apparent that the emerging dark blue sea of dormant suburbia expanded mostly into areas that were formerly light blue, though it also ate away at areas that were previously orange. Quantifying the change in the share of land colored orange just by looking at the maps is difficult, but it had actually decreased slightly by this period.