Steering clear of the roller-coaster housing markets, at the city and neighborhood level
Stock analysts often speak of the “beta” indicator as a measure of risk. Why not apply the same thinking to U.S. housing markets? It is useful for a property investor to know if a location has a history of incredible booms followed by epic busts, relative to the national average. A bit of consistency is not a bad thing, is it?
One size does not fit all
There is no such thing as “one” national housing market. After all, some places in the U.S. have experienced some very sharp peaks and troughs over the years, and others weathered the global recession with very little effect on housing prices.
At Spatial Laser, we often hear about the “sand states” and their high volatility. Examples include metros like Las Vegas, the inland areas of California, and Arizona, which experienced sharp pricing drops in the last cycle. Meanwhile, the Austin metro has continued to rise in a stepwise fashion, with each pause followed by yet another price surge. If we compare these cities to look for historical downside risk, it would be the cities following the Las Vegas pattern.
We can quantify that using the beta measure. The idea is to compare each city against the national average trend, a basket of “all” metro area prices indices. Combining all metros is the same idea as the using S&P 500 as a benchmark for the entire stock market. Markets that move up and down in tandem with the national average have a beta of 1. A city with beta closer to zero, meanwhile, is less correlated to what most other cities are experiencing. These cities move to their own rhythm and offer useful diversification benefits. They may go up and away – like Austin, or, are the slow and steady markets like Syracuse. Conversely, a beta of 2 means double the ups and downs of the country as a whole.
Finding the “safer returns”
Volatility is one thing, but appreciation is another. Who wouldn’t want the best of both worlds? On the chart below, we show historic risk (beta) plotted against the average annual appreciation of each market since the last market bottom (generally around 2009). Of course, past performance does not guarantee future returns, but this makes for an interesting start to the discussion. Austin, Dallas, Fort Worth, and Denver are all showing solid returns combined with limited price volatility. To show the relative footprint of each market, the size of the dot indicates the population of the metro area. From a portfolio theory perspective, it makes sense to take a deeper dive on these upper-left quadrant markets that seem to offer a better combination.
Drilling down to the micro level
If we can do it at the city level, why not do it at the neighborhood level? Sure enough, we can. The idea is the same – we measure the price movements of each neighborhood compared to the national average trend.
This is very important because of the large differences within a single metro area. While the metro area sets the general tone of the market, some sub-markets and neighborhoods are significantly more volatile than others. This is one example of where the micro analysis made possible by Spatial Laser becomes critical in planning an investment strategy.
Areas that are high-beta have probably been exposed to greater market swings due to factors such as: 1) greater swings in the supply-demand balance of homes due to inexpensive land and more new housing starts, 2) housing where affordability is more sensitive to interest rates, 3) proximity to industries that experience sharper cycles in growth and labor force size 4) greater exposure to cross-state migration from other cities at different points in the market cycle.
Meanwhile, the low-beta neighborhoods are insulated from big swings and offer a more steady path to growth with less downside risk. There is a belt of “low beta” mature neighborhoods about 15 miles from the Dallas central business district which seem to show the best resilience and greatest degree of decoupling from price movements nationally.
Putting it all together: one way to identify winning locations
The past 20 years of growth, and the volatility of that growth, can be combined into an index to compare areas. Based on this Dallas-Fort Worth example, areas around White Rock Lake, Richardson west of 75, and Northwest Dallas are all areas that have been standout performers for the last two decades on both fronts.
About Spatial laser
We are a real estate analytics company, and we use machine learning to help investors understand market conditions and investment returns at the hyper-local level. We bring institutional-grade site analysis to the single-family investment world. We combine many sources of data to project net operating income, holding costs, risk factors, as well as supply and demand, and display this interactively in a cloud-based mapping system.
Our cloud-based product is currently in development. While we build it out, we help investors on a customized basis and seek to learn more about how they can use data to make faster and better decisions.
Steven McCord is co-founder and Chief Executive Officer of Spatial Laser.
For single-family homes, about three-quarters of recent sales in the Dallas-Fort Worth area fell roughly into a $100 - $150/sf range*. Prices are generally kept in check by competition from new supply, as the Metroplex is surrounded by developable land with plenty of new construction. While prices in these areas do increase, they do so at a moderate pace.
One of the big exceptions to this is the central area, where limited-to-non-existent new supply meets consistently strong demand. This results in elevated pricing in the $200+/sf range and beyond. In the central area, the number of homes is constrained by the near absence of developable land. At the same time, proximity to centers of employment and amenities drive strong buyer demand.
The price-jump areas
In our other articles we have talked about investing in homes for cashflow. One of the other approaches to investing in single family homes is to buy in specific areas that are experiencing rapid price appreciation. In the Dallas context, this means buying in areas where the high-priced zone is expanding. Demand for close-in housing exceeds the supply, causing spillover of that demand into neighboring areas and the expansion of the high-priced zone.
This causes a “jump” in pricing where a neighborhood very quickly goes from “average” to “above average” in pricing. The result is that is common to see renovation of older homes in these transitional neighborhoods as buyers seek modern, updated interiors.
So how do we identify these areas of change?
To do this, we need to establish boundaries between “typical” and “high” home prices on a per square foot basis. Out of the 90,000 sales that occurred in the last 18 months, the upper quartile (top 25%) starts at the $150/sf mark, making this a convenient boundary. Similarly, if we look at natural breaks in the distribution of home prices, these also occur around the $150/sf mark and $200/sf mark.
We took thousands of home sales, put them as dots on a map, and used a spatial analytical technique to fill in the space between those points. The calculation is designed to fill in the map by giving the highest weight to the nearest sales points, and to give the more distant points a lower weight. This is called inverse-distance weighting (IDW)**.
Then we took the first six months of 2018, the first six months of 2019 and compared the two.
Zooming in on several areas reveals places where high prices may be creeping outward. This is noticeable near Love Field, especially southwest of Love Field near the UT Southwestern Medical Complex. It is also noticeable in Northwest Dallas where the high-priced zone may have moved by several blocks westward over the course of a year.
The result is that this analysis gives us the first clue on areas to watch. By looking at sales that have already occurred, we are observing lagging indicators that describe the past. The next step is to use other analytical techniques that describe the future.
For more information, or full-resolution maps, please contact us.
Steven McCord is co-founder of Spatial Laser.
*Based on the inter-quartile range of 90,000 homes sold in the last 18 months.
** We tested with another technique, which fills in the gaps using a regression-based formula called kriging and got similar results.
Investing in rental property doesn’t have to be a research headache. The traditional approach is all about evaluating deals one-by-one. This is great for specialized properties and is the norm in the commercial world, but falls short in the high-volume single family residential market where there is a sea of opportunities. Evaluating deals and then driving the neighborhoods to get a sense of their viability is a time-consuming task. It’s a process is begging for a little automation and help from a data-driven approach. Why spend time vetting deals when 80%+ of them can be easily eliminated, allowing you to focus your time and energy on the other 20%?
Much of this legwork can be simplified. We start by building a model that estimates rents for every location in the metro area using machine learning. From there, we look at holding costs, and finally, risks.
Doing this at the city level, or even ZIP code level gives a very general sense of how locations differ, but does not provide much specificity. Within a single ZIP code, there are often very large differences between neighborhoods or even opposite sides of the street. The solution is to bring this down to the micro level using much more granular data. Conventional wisdom says that a good balance of risk and return is in “B” neighborhoods that are reasonably priced but have important stability characteristics such as:
An investor looking for high yield while still remaining in a “B” neighborhood would filter out areas that don’t meet those criteria and get this result. Right away, we now have an idea where to start looking.