A brief history of local real estate, part 5: days on market.
In the latest thrilling episode in a series that's analyzed local real estate since April 2000 by sales price, absorption, sales and inventory, this week we'll see what light the market indicator "days on market" can shed on the history of local real estate. As always, we'll use data from the latest edition of my monthly newsletter, What's happening now in local real estate.
Days on market (DOM) is the average number of days homes are on the market before they go into contract. Because DOM indicates how fast the market is moving, it should tell us whether the market favors buyers or sellers, and by how much. And since the latest month's DOM is known well before that month's sales close and confirm the latest price trend, it's potentially a very valuable indicator, giving us the latest word on real estate prices well before they're revealed. If DOM goes up, the market is slowing and prices must be falling, and vice versa.
At least that's the idea. Because as we've learned, real estate indicators don't always pay attention to their job. Let's see if DOM has a stronger work ethic.
Here's average days on market for five local sub-markets since April 2000.
We can see that DOM varies greatly, not only between price ranges and between housing types but over market cycles of boom and bust and from season to season within the same year. If low days on market reliably indicates a hot real estate market, then 2000, early 2002 and late 2003 through early 2006 must have been smokin' and, in fact, they were and did, at least in most price ranges. And if high DOM tells us real estate is cold, then late 2001, late 2002 and late 2008 must have been frigid and, in fact, they were. Somehow buying a home seems a lot less risky when the economy is booming or, failing that, not catatonic. Available credit also helps a whole lot, as we've discovered, while economic-, terrorist- and war-induced anxiety have a way of making homebuying look a lot riskier.
Now let's do what we do and look for a correlation, either positive or negative, between sales price and our featured indicator of the week. We'll start with low-end SFR. Days on market is on the left axis, sales price/sq.ft. on the right.
Well now, this looks promising. DOM for affordable SFRs plunges in early 2004 just as price shoots up, exactly what you'd expect to see (although seeing exactly what you'd expect isn't always a good sign, or at least a sign of good science). Then, in early 2007, days on market begins rising just as the low-end SFR market starts running out of gas, although price coasts on fumes for several months, proving the truism that "prices are sticky, until they aren't". And by mid-2007 this end of the market is stalled on the railroad tracks with a mile-long freight train bearing down at eighty miles an hour.
Yes, pretty much what you'd expect—if you ignore a few stretches of positive correlation. For instance, both price and DOM begin declining early in 2008, as investors and first-time buyers start pouring into affordable (and increasingly so) neighborhoods. At first the influx isn't enough to stabilize price, but by late 2008 bank-owned homes are suddenly and inexplicably scarce while buyers aren't, an imbalance that pushes price up and DOM down. This is why it's impossible to time the market. Prices fall fall fall and everyone knows they'll fall fall fall forever or at least for a long long long time and then suddenly prices stop falling and then suddenly they bounce right back up and walk around acting like nothing happened. And note that I used the word "suddenly" three times when describing this phenomenon.
S-u-d-d-e-n-l-y. As in who saw that coming?
Next let's move up a price range.
Pretty much the same scenario, at least through early 2007, but then midrange gets its second wind, slowly loses it but keeps putting one foot gamely in front of the other until finally hitting the wall in late 2008. Not a bad performance from a price range often accused of getting soft and flabby.
Now let's go up another notch.
There's a neat, even dramatic negative correlation between DOM and price until mid 2003, when the top end of real estate (and, not coincidentally, the stock market) recover from the Iraq invasion jitters. But from then until the Lehman Brothers collapse of late 2008 the correlation is as much positive as negative.
But if we turn away and pretend we don't see this "anomaly", then wow! we've finally found a real estate market indicator that shows up to work on time and eats its lunch at its desk. Most of the time. Which might be as good as it gets.
In two weeks, A brief history of local real estate, part 6: bid.