The spring housing market music is playing, and purchase application data and active listing inventory rose together last week. The fear of not having an increase in inventory this spring should be put to rest. The other focus should be where mortgage rates go; only a little happened last week.
Here’s a quick rundown of the last week:
- Active listing rose 8,260 week to week, down a bit from last week’s gain, but I’m not complaining — anything on the plus side is positive.
- Purchase application data rose 5% weekly, keeping the streak of more positive data than negative for the year.
- Mortgage rates, once again, didn’t move too much this last week; the bottom level range was 6.50%, while the top was 6.67%; we ended the week at 6.59%.
Weekly housing inventory
Since new listing data was trending at all-time lows in 2023, some feared we wouldn’t see the typical spring inventory increase. After the last few weeks, we can put that fear aside: we are finally getting the seasonal increase in active listing. The one thing that is different about this year is that it took the longest time in history to get the seasonal bottom in inventory, but better late than never.
Since 2020, the seasonal inventory bump has happened later than usual — not until March or April. I reviewed the reasons for this in the HousingWire Daily podcast in February. Now that I believe the seasonal inventory bottom is in, we can focus on the next stage: tracking the weekly data on how much inventory growth we can get this year before the seasonal inventory starts to decline.
- Weekly inventory change (April 21-28): Inventory rose from 414,010 to 422,270
- Same week last year (April 22-April 29): Inventory rose from 271,510 to 287,821
- The bottom for 2022 was 240,194
- The peak for 2023 so far is 472,680
- For context, active listings for this week in 2015 was 1,070,493
New listing inventory hasn’t recovered since last year’s big mortgage rate spike and we have been trending at all-time lows in 2023. Even though it does appear that 2023 will have the lowest new listing data ever recorded in history, we are seeing the traditional growth in new listing data for the year, which is a big positive in my eyes.
We have to remember that a conventional seller is usually also a traditional buyer, so new listings growing toward their seasonal peak throws cold water on the idea that no one will list their homes because they already have a low mortgage rate (the mortgage rate lockdown theory).
New listings:
- 2021: 67,137
- 2022: 71,023
- 2023: 64,769
For some historical perspective, back when housing inventory levels were normal, here are the weekly new listing numbers for 2015-2017:
- 2015: 86,902
- 2016: 80,940
- 2017: 87,327
As you can see in the chart below, new listing data is highly seasonal, so we don’t have much time left before we should see a seasonal decline in the data line.
The NAR data going back decades shows how difficult it’s been to get back to anything normal on the active listing side since 2020. In 2007, when sales were down big, total active listings peaked at over 4 million. We had high inventory levels while the unemployment rate was still excellent in 2007. This proves that the mass supply growth we saw from 2005-2007 was due to credit stress, not because the economy was in a recession; the U.S. didn’t go into recession until 2008.
The total NAR inventory is still 980,000. As you can see in the chart below, there is a big difference between these two different historic housing economic cycles.
People often ask me why there is such a difference between the NAR data versus the Altos Research inventory data. This link explains the difference and is worth a read.
Finally, the significant observation I see with the inventory data this year versus last year is that last year’s new listing data was higher than in 2021. Also, the volume of active listings was higher in 2022 this week, even though we were working from a lower level. This shows me that while active inventory is growing, we don’t see the same growth volume this year versus last year. This can change as the spring inventory increase is still early, but that is the big difference I see for now.
The 10-year yield and mortgage rates
Last week, mortgage rates didn’t move too much, which might seem strange given that another bank, First Republic, was on its way to failure. The market is a bit calmer now than when Silicon Valley Bank failed, evident in how the stock and bond markets traded this last week.
The growth rate of inflation from the PCE data released Friday morning wasn’t a big deal in my view, it’s more of the same as service inflation has been firm recently, but looking out for 12 months, the growth rate of PCE will be below 4%. When that happens, the fear of breakaway 1970s inflation should be put to rest.
New home sales beat estimates while pending home sales slipped month to month. The 10-year yield tested the Gandalf line once again, to bounce off that level and only go back toward the end of the week.
In my 2023 forecast, I said that if the economy stays firm, the 10-year yield range should be between 3.21% and 4.25%, equating to 5.75% to 7.25% mortgage rates. If the economy gets weaker and we see a noticeable rise in jobless claims, the 10-year yield should go as low as 2.73%, translating to 5.25% mortgage rates.
Of course, the banking crisis has put a new variable into this year. However, even with that, the labor market, while getting softer, hasn’t broken yet.
As you can see in the chart below, the 10-year yield has stayed in its firm economic range 100% of the time. We can also see how hard it’s been for the 10-year yield to break below the 3.37%-3.42% area with any conviction. Mortgage rates have been in a range between 5.99% – 7.10%.
My line in the sand for the Fed pivot has always been 323,000 on the four-week moving average. This has been my big economic data line for the cycle since I raised my sixth and final recession red flag on Aug. 5, 2022. While the labor market is getting less tight, it’s not broken yet.
From the Department of Labor: “Initial claims for unemployment insurance benefits decreased by 16,000 in the week ending April 22, to 230,000. The four-week moving average fell to 236,000.”
Purchase application data
Purchase application data has been the main stabilizing data line for the housing since Nov. 9, 2022, with 16 positive prints versus six negative prints, after making some holiday adjustments to the data line. For 2023 we have had nine positive prints versus six negative prints. This data line has been very rate sensitive, and we are working from the lowest bar ever in this index. This past week we saw 5% week-to-week growth in the data line.
The year-over-year decline in purchase application data was 28%, the smallest year-over-year decline since September of 2022. However, the year-over-year data will improve independently even if the data line stands flat for the rest of the year.
The year-over-year comps will get noticeably easier as the year progresses, especially in the second half. This data line looks out 30-90 days for sales, and we are almost done with the seasonality of this data line. I always weigh this report from the second week of January to the first week of May. Traditionally after May, volumes will fall; this hasn’t been the case post-2020.
After May, I will address this issue with seasonality and a possible growth push later in the year, as seen in previous years.
The week ahead: Jobs and the Fed
It’s jobs and Fed week; jobs data is the one economic data line the Fed wants to slow down. Not only do they want to see the unemployment rate get to 4.5%-4.75%, they also want to see wage growth slow down even more. This week we have the job openings report, the ADP jobs report, jobless claims, and the BLS jobs report on Friday.
This can be a big week for mortgage rates and the bond market if the economic data does get softer on the labor front. Already, we see a cooldown in the labor market, but it’s not fast enough for the Fed.
Continuing claims have been rising for some time, meaning the labor market isn’t tight enough for these Americans to find work quickly after filing for unemployment benefits.
Job openings as high as 12 million in 2022 are now below 10 million. We are still at historic highs here, but the labor market is getting less tight. Job openings getting to 10 million early in the recovery was a huge call of mine.
For me, the one data line that shows that we don’t have 1970s entrenched inflation is the wage-growth data tied to the BLS jobs report. It has been cooling down even though we have had a tighter labor market, as I wrote about in the last jobs report.
The Fed is also meeting this week and the market has already priced in another quarter-percent rate hike; this will happen while the government finds a buyer for another bank on life support.
With the Fed meetings, it’s not their actions as much as their words. Since the Fed has now publicly said they believe a recession will happen later this year, based on their models, their actions during the recession matter more than what is left to do here. I believe that is where the discussion around the Fed should go since this might be the last rate hike of the cycle.
So, the week ahead has a lot of juicy economic data lines for us to keep an eye on because, to me, the housing market moves with the 10-year yield this year; when it goes down, the market acts better, and when it rises, demand gets softer.