• Relative to their expected population growth, cities’ current housing plans were classified as either “aggressive” or “modest.”
• Similarly, relative to past population growth, cities’ past ability to add new housing units were also scored as aggressive in meeting past growth needs, or modest.
• Cities were classified into four groupings: Aggressive promises, aggressive prior delivery (Seattle, San Francisco); aggressive promises, modest prior delivery (Boston, Los Angeles); modest promises, aggressive prior delivery (Atlanta, Denver); and Modest promises, modest prior delivery (Washington, D.C., New York).

A lack of new housing supply has been blamed for a raft of housing issues plaguing many large U.S. housing markets, including rapid appreciation, deteriorating affordability (especially at the lower end of the market and for renters) and very limited inventory. In response, mayors’ offices nationwide have formulated plans focused on building new housing and rehabilitating and preserving what housing already exists.

But are these plans enough to keep up with population growth, or even feasible given a city’s prior track record at meeting its housing needs?

An examination of past efforts and current plans in 38 large cities nationwide suggests mixed results to come. Some of the fastest-growing (and least affordable) cities, like Boston and Los Angeles, are dreaming big in their current housing plan; but if their past construction serves as a reliable predictor for the future, it might be a challenge for them to meet their goals. Other cities, like New York and Seattle, are setting goals commensurate with their past performance – but in the case of New York, is it ever enough?

Aggressive vs. Modest

Every five-to-ten years, mayors typically announce plans aimed at tackling the housing availability and affordability issues unique to each city. To understand how these plans stack up against anticipated demand, Zillow classified each city’s housing plan along two dimensions:

1. Record of Delivery: how successful the city has been historically in keeping up with population growth
2. Future Promise: whether the current plan will keep up with anticipated future population growth.[1]

These two dimensions give us four groups (figure 1).

• Aggressive Promise, Aggressive Prior Delivery

Both Seattle and San Francisco pushed hard to deliver new housing in the 2000s and currently have aggressive plans. But compared with Seattle, it seems as though San Francisco is currently dreaming bigger – but with a weaker track record of execution. It’s hard enough to build in the land-scarce, notoriously regulation-rich city of San Francisco, but it remains a very desirable place to live and its population growth is rapid and expected to continue.

Dallas is currently setting the most aggressive promise, but also has the best record of delivery, a testament to the can-do spirit of Texas growth.

• Aggressive Promise, Modest Prior Delivery

Cities that fall into this category are more concerning. Boston, for instance, added less than one housing unit for every two additional adults between 2000 and 2010, a relatively weak record for a fast-growing city during the peak of the housing boom.[2] The city has promised to build 53,000 housing units over the next 15 years. Assuming that the city’s population continues to grow as it has over the past four years, the city’s current plan is to build even more aggressively than during the height of the boom. This might be possible – and kudos to Boston if they are able to achieve it – but given their recent history, it seems to be a lofty goal.

Los Angeles’ latest housing plan calls for the construction or rehabilitation of 115,000 units, or about one housing unit for every two additional adults. But during the height of the housing boom, the city delivered just under two housing units for every five additional adults, leaving them with the most adults per household among large cities analyzed. While the promise is aggressive – and desperately needed – it seems unlikely that the city will be able to execute on it without additional support or resources.

• Modest Promise, Modest Prior Delivery

Cities falling within this category also have potential cause for worry. Despite failing to keep up with yesterday’s population growth, today’s building plans remain modest – a sign that past trends are likely to continue. Major cities in this camp include Washington, D.C., San Antonio, Sacramento and New York.

New York City, for example, currently plans to build 200,000 housing units over a ten-year period, a huge number, but one that only translates into one anticipated new housing unit for every three-to-four adults added to the city. While New York has been able to deliver on past promises of building many units, population growth has always outpaced new development in the Big Apple. In setting a relatively low bar, it seems likely that New York will continue to be an increasingly expensive place to live – a place in which young adults are increasingly forced to double-up or live in a home with other adult friends.

• Modest Promise, Aggressive Prior Delivery

Atlanta and Denver are cities that have rather modest housing development goals, despite having a strong track record of delivering lots of new homes in the past. In Denver, the city aims to build 6,000 housing units over ten years, roughly one new unit for every twenty new adults the city expects to see in the coming years. But while this is a really low goal, it may turn out okay in a city like Denver, as opposed to a city like Boston.

Denver currently has a relatively low number of adults living to a household (1.12 adults per household, versus 2.01 in Boston), likely due in part to the city’s more aggressive history of housing development. In addition, Denver has been able to make more planning decisions at the metro level, so perhaps the city expects surrounding cities and suburbs to be able to absorb the demand that the city itself cannot sustain. However, much of Denver’s astonishing population and economic growth has come in more recent years. As the city and metro continues to grow and rents and home values rise, adults may increasingly double-up. In ten years’ time, the mile-high city might wish they aimed higher.

The Best Laid Plans…

In addition to comparing housing unit growth with population growth, we want to understand how current plans to add housing stack up to what the city has previously been able to build, on average (ignoring the changes in population). We contrasted the number of housing units built annually between 2000 and 2010 with the number of housing units that will need to be constructed each year according to the current housing plan of each city (figure 2). The 45-degree line in figure 2 serves as a benchmark, representing perfectly equal rates of construction during the housing boom (2000-2010) and the anticipated current plan.

Cities located to the left of the line, like Boston and Los Angeles that made aggressive promises, will have to pick up their building rate to meet their goals of constructing more homes per year going forward than during the height of the housing boom (2000-2010). New York, despite making modest promises in terms of keeping up with population growth, joins Los Angeles and Boston in the promise to step up the rate of build. Will that even be possible?

On the other hand, cities located to the right of the dividing line, including Sacramento, Denver, Las Vegas and Houston, have goals that should be achievable given that they have built more units in the past.

Of course, it is not quite so simple. For example, Chicago’s position far from the line on the graph implies that the city’s housing goal is pretty high relative to what they built between 2000 and 2010. But Chicago also lost more than 60,000 adult residents during this period – and as the city’s population fell, Chicago might have slimmed down its past building ambitions in response. In recent years, the population has been steadily increasing. To both ensure that current housing meets the needs of new residents and that units currently in the housing stock do not become dilapidated, Chicago’s goal might be reasonable, and one that’s likely achievable since Chicago wasn’t overly challenged to grow during the 2000s.

Putting it Together

Atlanta, Denver and Kansas City clearly have a record of delivering sufficient housing for their growing populations, and their current plans are within their means. Dallas, San Francisco and Seattle also have big plans for the future (and have previously built quite a bit), which may come as a surprise for those familiar with recent home prices and rent trends in those cities, particularly in San Francisco and Seattle. But unless San Francisco and Dallas really ramp up construction in the next few years, it may be difficult for them to follow-through with their plans; it seems likely, however, that Seattle will be able to deliver on their promise of adding 50,000 units to the housing stock.

On the other hand, Boston, Los Angeles and San Diego have all set pretty large construction goals. If past performance is a predictor of the future, it may be difficult for these cities to deliver on those goals. Similarly, New York City’s plan may only begin to scratch the surface of the housing issues that the city faces.

Even with these numbers, it is difficult to definitively say if a city is producing enough housing, largely because the type of housing built is just as important as the total number of units added. There is a need for more affordable housing in many cities. New construction should also include the right mix of housing – whether apartment, condominiums, townhouses or single-family homes – to accommodate each city’s population and demand.

As abundant job opportunities and a preference for city living continue to attract individuals to these cities, it is important for city planners to keep housing policy front of mind to ensure that these cities remain desirable and affordable places to live.

Methodology

To track the record of delivery, we use the number of housing units added to the housing stock for every new adult in the decade from 2000 to 2010.[3]

The vertical line divides cities characterized as having modest records of delivery relative to their more aggressive peers. This is set at five units for every additional six adults. While this may seem like a high threshold, the period between 2000 and 2010 represented the peak of housing construction and housing was quickly being built everywhere. Without knowing the housing needs in the year 2000, if we assumed that keeping pace required one new unit for every two adults – the stereotypical household – we would have considered all the markets we studied except for Los Angeles to be overbuilt during this decade. As many of these cities are locally understood to have too little housing today, we took the relative approach instead when classifying cities’ records of delivery.

We define the strength of a city’s future promise – if its plans are modest or aggressive – similarly: The number of units the city has promised to build[4] for every adult added as the population grows over the time frame of the housing plan.[5]

The threshold for a current plan to be classified as modest or aggressive is much more moderate, requiring one new housing unit for three additional adults. Again, this threshold is defined to divide large cities in a balanced way. If we insisted on the one unit per every two new adults rule, we would discover that only five cities – Dallas, Los Angeles, San Diego, San Francisco, and Seattle – are making “aggressive” future promises.

[1] While we investigated 38 city plans from the largest U.S. cities, we’ve limited this portion of our analysis to the 19 who are currently challenged by growing populations. Cities that are still struggling to recover from the great recession of 2008 were largely excluded from the analyses, as their housing problems are primarily a product of trying to keep their housing stock from dilapidating.

[2] Of the cities studied here, the typical city built about two units for every three adults between 2000 and 2010.

[3] Due to the high variance in the record of delivery, the scale has been logged in Figure 1.

[4] Housing plans are often rather vague, so we made some assumptions:

1. If a plan involves “preserving”, “rehabilitating”, or “restoring” housing units, then these units would have otherwise been demolished or no longer part of the housing stock within the city.
2. If a plan involves constructing a certain number of new units, that whole number is added to the housing stock. For example, if a city promises to build or construct 4,000 new units, then this means there will be 4,000 units added to the housing stock. This does not mean that 2,000 units will be demolished to make space to build a few high-rises that collectively contain 4,000 units, resulting in a net of 2,000 units.

[5] We used the adult population growth between 2010 and 2014 as a proxy for future expected growth. Using an annualized population growth during this time period, we multiply the growth rate by the number of years of the plan to determine the anticipated adult population growth over that period.