A certain amount of housing prices are due to supply and demand - more people means more demand. It can be hard, if not impossible, to keep up with that demand, which means more expensive housing. Of course there are tons of factors that play into housing affordability but supply (new units being built) and demand (new residents moving to an area) are easy to find data that can show the impact of these factors.
According to the Census estimates, the Austin-Round Rock Metropolitan area grew by 53,086 people from July 1, 2017 to July 1, 2018. A little less than a quarter, or 23.6%, of those individuals moved to live within the Austin city limits. That also means that about 40,000 people moved into the surrounding areas. During that same period, construction was completed on 12,453 housing units. Using the Census average household size of 2.48 people, that means we gained enough housing for a little over 30,000 people! This, in theory, should have helped slow down housing cost increases.
Austin has been experiencing a similar growth rate since 2010, but there have not been enough housing units completed to keep up with the change - 9,449 were completed between July 1, 2018 and June 30, 2019. Using the same average household size, that made room for just 23,400 people, which means that well over half of the new residents moved to the area, but outside of the city limits. The low number of new units likely increased the cost of housing by not meeting demand.
If units continue to be finished at the same rate, we will see 13,750 more units come available by June 30, 2020, which should be able to provide housing for about 34,100 new residents at the current average household size. While these are not huge numbers, and may not lower average rents it may help slow the relentless increase of housing costs.
Here's why this is dev-related to me. Once you have the data, it's fairly easy to figure out how many houses have been built within a certain period. Look at the number of building permits and a status of "Final" along with their Status date. (The Status date looks to have been started in 2007, and retroactively applied through 2008, so even this method only works for about the last 10 years.)
More difficult is estimating the units that may be completed soon. To do this, I first created a field called "Time to Completion" that finds the difference between the Completed Date and the Issue Date. That allowed me to looked at the average completion time in days. I narrowed the time range down to the last two years because construction methods change, crew availability changes, and I thought that two years would be a large enough time frame to have a broad average while also reflecting a state that may be similar to what we're facing now. (As I'm a developer, and not a construction or real estate professional, I might be wrong about that.)
If you look at the average completion time for all Building Permits related to new housing units, from July 1, 2017 to now, it is 331.7 days. I did exclude 16 permits that were listed at taking over 2000 days to complete. That's 5 1/2 years! Something seems to have gone really wrong with those, and I'm comfortable calling them outliers for our current purposes. That may seem like a good number to use, but if you think for a second longer - does it take the same amount of time to build a 3,000 unit complex as a single house? I hope not! I looked at the relationship between number of units and time to completion and, yes! Common sense works here, it takes longer to build more units. You can see it in a dashboard I made on Tableau Public.
I added a line to my scatter plot with Housing Units and Time to Completion. We could use the nice formula that Tableau generated based on number of units alone, but the line goes way above the reality for larger complexes, which means we would seriously over-estimate completion time for larger projects, like those with more than 150 units. What else might be at play? Next I checked the time to completion by permit class and wow! Now we see a range from 297.5 days to 989.9. That's from about 10 months at the shortest to over two and a half years! That seems like a better estimate.
After exploring these options, I decided to use the average Time to Completion by Permit Class. I created a new field that calculates an Estimated Completion Date based on the Issue Date plus the average Time to Completion for a given permit's Permit Class. If you are a more visual person, go download my Tableau workbook and play with it!
The final step was to account for actual completed date and estimated completed date at the same time. I created one more field that looks at the project status, decides what date to use (actual or estimate) and then allows us to view all of the building permits together, whether they have been completed yet or not.
If more people are moving to Austin and the area every year, and are increasing the demand for housing, then we can never really expect the housing costs to go down as long as the economy is doing well. It's awkward, but what incentive do housing developers have to make houses that will sell for less? Apartments that command lower rents may be profitable to management companies because they could tap a separate market, but at some point the profit may not be worth it to those companies. That is where local government steps in - they can (and do!) create incentives for developers to build affordable units.
A related issue is land. You can build on empty land, or you can tear down existing units. The former is very limited within city limits, and the latter often results in simple replacement of older (read: cheaper) housing with newer (read: more expensive) housing. Changing the land development code is one way to encourage greater density, which will allow for more housing within city limits. Existing options, like two detached homes on a single lot, are becoming more popular as a way for those willing to live in smaller homes to stay within the city. This, then is another area where local government can help control housing costs. And we'll get into that next time.