The first thing I learned when I started looking at buying houses is that there are NO HOUSES! Everyone says there’s a massive supply shortage. There are not enough houses for all the people! It’s an extremely scarce resource that if you miss out now, you may never own a house.
I saw this scary graph over and over again – this is inventory for San Diego. Look at that dropoff! Wow, this makes most buyers terrified that something about the world has changed. There are no houses left. We went from 10,000 houses to 1,000 houses?! There are 1/10th the number of houses available?
Well, not really.
Definition of Inventory
Eventually I decided to look up how this graph is made which led me to the definition of inventory. According to NAR, “Inventory is calculated monthly by taking a count of the number of active listings and pending sales on the last day of the month.”
So it’s a snapshot of how many houses are LEFT at the end of the month. Which means it’s affected by both supply AND demand. It is not a true measure of supply.
Inventory = Supply – Demand
Maybe this is well understood by economists, but regular buyers (and the RE professionals who show them this) seem to use it as a stand-in for supply. Which it is most definitely not. It’s just as dependent on demand.
There’s one other weirdness in the way it is measured by NAR. This graph is counting only those houses in active or pending at the end of the month. A house goes from not active –> active –> pending –> closed.
So the bean counter doing the math has to catch the house in either “active” or “pending” state at the end of the month. Which makes this number not just dependent on how many sellers wanted to sell their house, but also very dependent on how quickly the house sold.
How Quickly = Time Spent Active + Time Spend Pending
Note that some data only uses active listings, not both active and pending. These would see the same effect in that houses that used to be on market for a month would get counted, but since they are only sitting for days, they don’t get counted at all. In those cases it’s dependent on just how long the house was active.
How Quickly = Time Spent Active
Let’s take a look at these two factors in the normal times.
Time Spent Active is easy – it’s tracked as “Days On Market”. You can see here that it used to be pretty reliably > 30 days before 2020.
Time Spent Pending is a bit tougher. It’s not tracked by SDAR/NAR as far as I can tell. All I know is in California it tends to take 30-45 days to get through the contingency process. Let’s just call that 40 days.
So on average a house that a seller wanted to sell would spend 30 + 40 days in active or pending, making it very likely to show up in two months’ inventory numbers.
Say a house is put on the market Jan 5, it’s likely to go under contract Feb 5th and close the sale Mar 15th. This mean it gets counted as inventory in Jan and Feb.
However in our current market, houses are being snapped up in a weekend and waiving contingencies and closing in record time. Many are also being bought in cash, speeding the process up even further.
If a house is bought in cash with all contingencies waived, this could close in 2 weeks. A cursory look at cheap houses that might be likely to be cash buys shows that this is very possible in California.
This house would actually have been counted in inventory at the end of Jan. But only one month and not 2 as we were seeing before. Timed right, it could sell within a month, and not get counted at all.
Around this time friends of mine sold their house to a service called Sundae. They bought the house directly from them to sell at a later date. Their house went directly to sold. It was never in active or pending. So there’s no way inventory would catch this house.
This means there are houses that sellers are trying to sell but never show up in inventory. They are part of supply but due to high demand, we never see them in “inventory”.
I don’t believe this timing issue accounts for all of the shortfall. But I do think that it makes inventory a bad way to talk about supply. It’s not just supply, it’s a measure of the difference between supply and demand, and a poor one at that when you throw in the issue of houses being over or undercounted depending on how long the sale took.
And maybe the realtors know that it’s a measure of supply and demand, but the buyers by and large do not. Almost everyone I’ve talked to believes this shows we have “way fewer houses for sale” than there used to be. So it’s misunderstood at best, but deliberately misleading at worst.
So gut check, how can we check how many houses sellers did want to sell? Not exactly, but as it turns out, they do track “number of houses sold” and surprise surprise, there is no sudden drop in houses being sold!
Number of sales in San Diego follow the same seasonal pattern and overall look a bit higher in 2020-2021 than 2018-2019.
So why does it feel like there are so few listings? Because there are.
When Sundae buys a house, it doesn’t get listed. When Zillow or Opendoor buys a house, it doesn’t get listed. When an investor approaches a seller directly, it doesn’t get listed. Realtors are sharing “pocket listings” to other realtors before the home even goes active. We too got shown houses before they went live. So it’s true that there are fewer listings, fewer houses available for us plebs to buy, so it does still suck, no doubt about that.
Inventory does mean something. But it’s good to understand it, because knowing that there aren’t actually 1/10th the number of houses for sale makes me feel a lot less anxious about the bigger picture. It helps me be patient and wait for demand to slow down.
Disclaimer: I’m an idiot first time home buyer. I’ve never taken an econ class in my life. I’m just sharing what I see and learn as it happens. I am 100% certain I will get things wrong, so don’t take any of this as the golden truth.