For older American consumers, senior housing costs are rising slowly but steadily as the population ages. Government regulation is known to affect prices and price growth in other sectors like residential real estate. Yet we know little about how regulation affects senior housing costs.
A Place for Mom looked at one type of regulation — land-use regulation — and found no link to the growth of senior housing costs. The finding raises questions about whether other types of regulation could affect senior housing costs, and how to measure their impact on the pocketbooks of aging Americans.
The older adult population in the U.S. will grow more than 50% by 2030 and will more than double by 2060 . Half of today’s older adults will need assistance with activities of daily living . Declining family sizes  and lack of publicly-funded options for the middle class  create an elder care gap pushing more seniors into private-pay senior housing, which costs nearly $50,000 per year in the US. Those costs grew annually by 2.4% in 2015, and are expected to grow faster in coming years as demand increases .
Given rising demand, price growth will be driven largely by whether supply can keep up. One driver of supply growth is the local regulatory environment.
Lawmakers institute federal, state, and local regulations ostensibly to protect people and the environment, and to promote healthy levels of economic competition. Yet these regulations can increase the costs of production or reduce supply, leading to higher costs for consumers.
A good example of how regulation affects prices is land-use regulation. When lawmakers put a cap on the allowable height of buildings or minimum amount of green space in an area, that limits real estate supply, which could raise prices. Another way regulation can influence prices is by delaying construction while builders obtain permits and fill out other required paperwork.
None of this means regulation is inherently bad. It just means we should pay attention to both the potential benefits andcosts of regulation.
The costs of regulation are often unintended. So we usually have to measure the impact on consumers after regulations are already in place. Regulations also grow out of political compromise and bureaucracy. What’s more, regulatory constraints vary by location. All of this makes it harder to understand the regulatory environment itself, much less its influence on prices.
For these reasons, economists have to develop clever metrics of regulatory constraints. One such metric is the Wharton Residential Land-Use Regulation Index developed at the Wharton School at the University of Pennsylvania . This index summarizes the strictness of local land-use regulation and its enforcement into a single score.
Using the Wharton index, economists have shown that areas with stricter residential land-use regulations have higher home prices and faster home price growth. Rapidly growing prices in an area also make people more willing to have roommates [6, 7].
Because senior housing is a subset of residential real estate, the link between land-use regulation and price growth should hold. But does it?
Before we discuss our findings, you need to know a bit about the major types of senior housing. Senior housing costs cover not only rent but, for some types of housing, also care. Below are descriptions of the three main types of private-pay senior living.
Independent Livingis for or independent seniors seeking community. Rent typically includes individual apartments, transportation services, up to three meals per day, and weekly housekeeping.
Assisted Livingis for seniors who need daily assistance with personal care. Rent typically includes private and semi-private suites, medication management, assistance with personal care, all meals and housekeeping, and 24-hour access to caregivers.
Memory Careis for seniors with moderate to advanced cognitive impairment. Memory care is like assisted living with specialized memory care services and programs, specially trained staff, and secured environments.
Senior living costs increase with the level of care. Independent living charges include rent only. Assisted living charges are about 1.5 times greater than independent living with a third of the costs going to care. Memory care is about twice as much as independent living with costs split evenly between rent and care.
Independent living is the type of care most similar to real estate. So if there is a link between land-use regulation and senior living costs, we’d expect it to be strongest for independent living.
A Place for Mom matched the Wharton index to thousands of senior housing transactions in over 750 municipalities in the United States. Our findings show:
There’s no evidence land-use regulation impacts the growth of senior housing costs. Case closed, right? Well, no. The land-use regulation index doesn’t capture regulatory components important to the senior housing industry, which we outline below:
We are unaware of any validated local index that encapsulates each of these types of regulation. If we want to understand the benefits andcosts of regulation to senior living consumers, maybe we need to build a new index.
We asked Dr. Albert Saiz his thoughts on these findings. Aside from co-authoring the Wharton index with Dr. Joseph Gyourko and Dr. Anita Summers, Saiz is the Director of the MIT Center for Real Estate and Co-Chairman of the Samuel Tak Lee Real Estate Entrepreneurship Lab. According to Saiz, another explanation for our findings is that areas with stricter land-use regulation are also more likely to be lenient about or even subsidize senior housing construction. This would cancel out the negative effect of land-use regulation on affordability.
First we look at how land-use regulation relates to senior housing costs. For all care types, there is a clear upward trend in inflation-adjusted price for positive values of the Wharton index, which indicates higher than average regulatory constraints.
The effect varies somewhat by care type. Given its similarity to the residential real estate market, independent living costs should increase more quickly with regulation pressure, and they do. There isn’t enough data to trust the apparent downward trend for assisted living or memory care at the highest index levels, evidenced by the wide gray uncertainty bands around the curve.
It’s no surprise senior living costs are higher in more regulated places. Income, thus the price of land and living costs, also tend to be higher in those places. What matters more is the relationship between regulation and (inflation-adjusted) price growth.
We measured price growth in two ways. First, we compared the percent growth in costs for the same property offering the same care type in 2015 versus 2012 (so-called same-store growth).
The plot below shows no clear relationship between same-store price growth and land-use regulation. The smoothed average shows a complex rend, but the uncertainty around that trend (represented by the gray bands) is very high.
We also compared the median care-type-specific price in each city in 2015 versus 2012. Again, no clear evidence of a link between price growth and land-use regulation; the uncertainty around the trend remains high.
If land-use regulation makes senior living costs growth faster, seniors should be more willing to share a room in more regulated areas to offset the cost increases. Of course, there’s no evidence land-use regulation drives faster senior housing cost growth. So it’s no surprise we can’t find a relationship between land-use regulation and willingness to share a room.
Although there is an apparent upward trend for memory care, the width of the uncertainty bands around the trend shows we can’t actually pinpoint its magnitude or direction.
The relationship between land-use regulation and home prices is robust. Yet we didn’t see the same thing for senior housing. Maybe regulation doesn’t influence senior housing costs much. Or maybe we aren’t measuring the regulatory factors most relevant to senior housing and care. Until we build and validate an index of senior housing and care regulation, we can’t know for sure. With over 20% of the population set to be older than 65 by 2040 , not knowing can hurt us.
We obtained the Wharton index data from the Wharton School website. There were some misspelled place names (e.g., Palmer, Alaska), which we corrected. There were also some duplicate places, so we average the index between them.
We obtained the senior living transaction data from the A Place for Mom transaction database. A small number of senior living transactions have unrealistically high or low costs, even on an inflation-adjusted basis. We excluded senior living transactions either less than a third of or greater than three times the median inflation-adjusted care-type-specific costs for the most granular geographic level (starting with zip code and ending at the national level) with at least ten move-ins.
To calculate same-store cost growth, we first took the median care-type-specific transaction cost for each property, care type, and year. We then compared prices in different years for the same property and care type. We did something similar for city-level cost growth, except the first step was to take the median by city instead of property.
The plots were generated using the geom_smooth function in the ggplot2 package for the R statistical programming language. The function applies a loess smoothing (or a generalized additive model if there is insufficient data for a loess smooth) and plots the smoothed average along with its 95% confidence intervals. We used the default bandwidth (i.e., the span argument) of 0.75. For the plot of willingness to share a room, we specified logistic generalized additive models using the method and method.args arguments of the function (specifically, we set method = “gam” and method.args = list(family = “binomial”)).
We also collected a land developability index measuring the amount of developable land area by county, which we could potentially control for in the analysis. One might also employ a multilevel modeling strategy to reduce bias in the estimates of land-regulation effects due to variation in prices among properties and cities, some of which are better represented in the data. Finally, one could build a complex model with many potential predictors and measure the predictive importance of land-use regulation. Given the results presented here, we doubt the substantive results would change.