This chart is the historic housing price index in Savannah (bold blue line), and I added in a couple trend lines to illustrate trends in the data.
See the grey line? That’s the trend line of increases in housing price index from 1984-2000. See how much slower the growth was back then? That’s because mortgage interest rates were in the double digits back in the 80s and remained in the high single digits through the early 90s. High interest rates always equal lower valuations of assets.
The yellow line is the most important. That shows the average growth in housing prices in Savannah for the period of 1995-2003. This is important because interest rates were relatively low (about 5-6% on average), and this was before the 2003-2006 boom in prices. This boom was a result of shoddy lending practices that we don’t see anymore. Check out how big that housing bubble got, and check out how much below historic trend lines housing prices got by 2012. We still haven’t fully caught back up to where we should be.
This is the biggest reason that I don’t foresee a drop in prices – unlike 2007, real estate in Savannah is not in a bubble. Will we see prices stagnate and maybe stay at the same level for a period of time? Probably. Will buyers be able to get more concessions out of sellers? I think so. Are you going to be able to steal deals? Not unless you find a juicy off-market deal and a seller who doesn’t know any better – or unless we totally gorp up the response to the virus and don’t execute the necessary economic stimulus measures. I think we could be doing better on that end, but I think we’ll make it okay, especially after seeing that $2.2T stimulus bill pass. It was foreclosures that drove the drop in prices from 2008-2012, and we won’t be seeing very many of those because of the stimulus bill that was recently signed into law.
Bottom line is this — if you’re an investor looking to steal some deals like it’s 2012, it’s probably not going to happen. Your best window of opportunity will probably run until about August. My recommendation? Be a little more choosy, bump down the year one rent estimates in your modeling, and keep more cash on hand than you normally would – but other than that keep buying. If you’re truly worried then keep your cash on the sideline until early May, I think by then we’ll have a good idea of how this will play out.
We’d love to hear your thoughts and feedback or answer any questions, especially if you’re a local investor or thinking about investing in Savannah real estate. Please reach out anytime – we’ve got plenty of time on our hands these days!
Methodology on the COVID impact on downtown Savannah average market rent modeling
My independent variables (e.g., the variables that I will manipulate) are the percentage of SCAD students who leave Savannah, the unemployment rate, and the percentage of AirBNB owners that move their units to the long term rental market. These are what I see as the biggest drivers of changes in supply and demand in the rental over the next few months.
My dependent variables (the variables that will be changed based on differences with the independent variables) are market rent and excess rental supply.
I pulled a lot of different statistics to help me build this model. They include the following data points taken yearly from 2010-2019: average rent in the 31401 zip (Zillow), SCAD enrollment and dorm capacity, Savannah metro unemployment, population, per capita income, GDP, and housing starts, 31401 zip building permits, new apartment units, and population.
I then performed a linear regression for a few of these statistics and their relationship to market rent in the 31401 zip and I averaged the output of each regression to come up with a predicted fair market rent, which I adjusted up or down depending on whether there was an excess of supply or not. How did I determine that?
I used building permit data to get a clear picture of the number of new rental units coming on the market every year vs the growth in population, and I established that the ideal ratio of units/population is .26 housing units/person. Now, you might be thinking that there would be about 4 people per apartment in that scenario, but that’s not really the case since a good portion of the population is homeowners and there’s just no way to model that. So how did I come up with .26? Basically, I took the average ratio of the years 2011-2018 and used that — I figure that the free market would come up with the best ideal ratio and I should just stick with that. So I can now use that constant of .26 to model how increasing supply or decreasing population would lead to a shortage or surplus in the market, and I adjusted market rent up or down 10 cents per unit of excess/shortage.
I then used these models to predict market rent for the years 2010-2019 to see how close my model was to the actual values. I saw that my model was off by an average of 3.12%, and 6.47% on the worst year, where the model was $100 off the actual market rent for that year. Not too bad for a guy who got a B+ in stats, I guess!
Author: Pat Wilver