The volume of mortgage applications fell during the last week of June as compared to the prior week which had contained an adjustment to account for the Juneteenth holiday. The Mortgage Bankers Association (MBA) reported a decline of 4.4 percent in its seasonally adjusted Market Composite Index. The change ended a three-week streak of volume gains. On an unadjusted basis, however, that index did move 6.0 percent higher.
The Refinance Index decreased 4 percent from the previous week and was 30 percent lower than the same week one year ago. The refinance share of applications ticked up to 27.4 percent from 27.2 percent.
The seasonally adjusted Purchase Index declined 5.0 percent but was 6.0 percent higher before adjustment. Applications were down 22 percent from the same week in 2022.
“Mortgage applications fell to their lowest level in a month last week as rates for most loan types increased. As mortgage-Treasury spreads remained wide, the 30-year fixed rate increased to 6.85 percent, the highest rate since the end of May,” according to Joel Kan, MBA’s Vice President and Deputy Chief Economist. “Purchase applications decreased for the first time in a month, as homebuyers remained sensitive to rate changes. Rates are still over a percentage point higher than a year ago, and housing affordability is still a challenge in many parts of the country. However, the average loan size for a purchase application declined to $423,500 – its lowest level since January 2023. This was likely driven by reduced purchase activity in some high-price markets and more activity in some of the lower price tiers as buyers searched for more affordable options.”
The recent low of $423,500 cited by Kan for purchase loans was about $4,500 below the average a week earlier. Overall loan sizes declined from $384,200 to $378,800. The last time loan sizes rose was in mid-May.
Other Highlights from MBA/s Weekly Mortgage Applications Survey
The FHA share of total applications increased to 13.0 percent from 12.9 percent and the Veterans Administration’s share decreased to 11.7 percent from 12.2 percent. USDA applications were unchanged from the prior week at a 0.4 percent share.
The 6.85 percent average contract interest rate for conforming 30-year fixed-rate mortgages (FRM) represented a 10-basis point increase from the prior week. Points rose to 0.65 from 0.64.
The rate for jumbo 30-year FRM increased to 6.95 percent from 6.91 percent,with points slipping to 0.64 from 0.69.
FHA-backed 30-year FRM had a rate of 6.68 percent with 0.98 point. The prior week the rate was 6.63 percent with 1.08 points.
Fifteen-year FRM rates averaged 6.30 percent, up 7 basis points. Points increased to 0.91 from 0.69.
The average contract interest rate for 5/1 adjustable-rate mortgages (ARMs) dropped to 6.00 percent from 6.28 percent,with points increasing to 1.23 from 1.02.
The ARM share of activity increased to 6.2 percent from 6.1 percent.
The volume of mortgage applications fell during the last week of June as compared to the prior week which had contained an adjustment to account for the Juneteenth holiday. The Mortgage Bankers Association (MBA) reported a decline of 4.4 percent in its seasonally adjusted Market Composite Index. The change ended a three-week streak of volume gains. On an unadjusted basis, however, that index did move 6.0 percent higher.
The Refinance Index decreased 4 percent from the previous week and was 30 percent lower than the same week one year ago. The refinance share of applications ticked up to 27.4 percent from 27.2 percent.
The seasonally adjusted Purchase Index declined 5.0 percent but was 6.0 percent higher before adjustment. Applications were down 22 percent from the same week in 2022.
“Mortgage applications fell to their lowest level in a month last week as rates for most loan types increased. As mortgage-Treasury spreads remained wide, the 30-year fixed rate increased to 6.85 percent, the highest rate since the end of May,” according to Joel Kan, MBA’s Vice President and Deputy Chief Economist. “Purchase applications decreased for the first time in a month, as homebuyers remained sensitive to rate changes. Rates are still over a percentage point higher than a year ago, and housing affordability is still a challenge in many parts of the country. However, the average loan size for a purchase application declined to $423,500 – its lowest level since January 2023. This was likely driven by reduced purchase activity in some high-price markets and more activity in some of the lower price tiers as buyers searched for more affordable options.”
The recent low of $423,500 cited by Kan for purchase loans was about $4,500 below the average a week earlier. Overall loan sizes declined from $384,200 to $378,800. The last time loan sizes rose was in mid-May.
Other Highlights from MBA/s Weekly Mortgage Applications Survey
The FHA share of total applications increased to 13.0 percent from 12.9 percent and the Veterans Administration’s share decreased to 11.7 percent from 12.2 percent. USDA applications were unchanged from the prior week at a 0.4 percent share.
The 6.85 percent average contract interest rate for conforming 30-year fixed-rate mortgages (FRM) represented a 10-basis point increase from the prior week. Points rose to 0.65 from 0.64.
The rate for jumbo 30-year FRM increased to 6.95 percent from 6.91 percent,with points slipping to 0.64 from 0.69.
FHA-backed 30-year FRM had a rate of 6.68 percent with 0.98 point. The prior week the rate was 6.63 percent with 1.08 points.
Fifteen-year FRM rates averaged 6.30 percent, up 7 basis points. Points increased to 0.91 from 0.69.
The average contract interest rate for 5/1 adjustable-rate mortgages (ARMs) dropped to 6.00 percent from 6.28 percent,with points increasing to 1.23 from 1.02.
The ARM share of activity increased to 6.2 percent from 6.1 percent.
As rates for most loan types increased, mortgage applications fell to their lowest level in a month last week.
For the week that ended June 30, mortgage applications fell 4.4% from the prior week, according to data from the Mortgage Bankers Association.
“As mortgage-Treasury spreads remained wide, the 30-year fixed rate increased to 6.85%, the highest rate since the end of May,” said Joel Kan, MBA’s vice president and deputy chief economist. “Purchase applications decreased for the first time in a month, as homebuyers remained sensitive to rate changes. Rates are still over a percentage point higher than a year ago, and housing affordability is still a challenge in many parts of the country.”
Reduced purchase activity in some high-price markets and more activity in some of the lower price tiers might be responsible for this average loan size drop, explained Kan.
Last week, mortgage rates increased for most loan types, with the 30-year fixed rate increasing to 6.85% from 6.75%, per the MBA’s data. On the other hand, the jumbo rate was higher than the conforming rate for the fourth week in a row. The MBA data showed that for jumbo loan balances (greater than $726,200), the rate jumped to 6.95% from 6.91% last week.
At Mortgage News Daily, mortgage rates were higher on Wednesday, at 7.08%.
Purchase applications decreased for the first time in a month, the purchase index decreased by 4.4% from one week earlier and was 22% lower than last year’s level on an unadjusted seasonal basis. Refinancing applications decreased 4% last week compared to the previous week and were 30% lower than the same week one year ago. However, the refinance share of mortgage activity increased to 27.4% of total applications from 27.2% the prior week.
The Federal Housing Administration loans’ share increased to 13% from 12.9% the week prior. The U.S. Department of Veteran Affairs loans’ share decreased to 11.7% from 12.2% the week prior. And the U.S. Department of Agriculture loans’ share remained unchanged at 0.4% of the total applications.
Housing inventory finally broke under 2022 levels last week. To give you an idea how different this year is from last year, last week in 2022, active listings grew 30,940 while this year they only grew 5,848. Mortgage rates rose last week after the better-than-anticipated jobless claims data but even with higher rates, we also had a third week of positive purchase application data.
Here’s a quick rundown of last week:
Active inventory grew by a disappointing 5,848 weekly
Mortgage rates went above 7% again after better labor data
Purchase application data showed 3% growth week to week
Weekly housing inventory
On May 15, I went on CNBC and talked about how inventory growth in 2023 resembled a zombie from the show The Walking Dead, slowly trying to rise from the grave. Since May 15, that trend has continued to the point that inventory in America is now negative year over year.
We have often discussed that the housing market dynamics changed starting Nov. 9, 2022, and today you can see the final result of that dynamic shift as inventory is now negative versus the 2022 data — all before July 4th. I recently recapped this crazy period on the HousingWire Daily podcast, going into detail about what happened in housing over the last year.
Weekly inventory change (June 23-30): Inventory rose from 459,907-465,755
Same week last year (June 24-July 1): Inventory rose from 441,106 to 472,046
The inventory bottom for 2022 was 240,194
The inventory peak for 2023 so far is 472,688
For context, active listings for this week in 2015 were 1,183,390
Seeing negative year-over-year inventory before July 4 would be a big deal if last year wasn’t so crazy. However, I need to put some context into what happened in 2022. In March of 2022 we had the lowest inventory levels ever recorded in history. Then in a short amount time, we had the biggest and fastest mortgage rate spike in history, which facilitated the biggest one-year crash in home sales in history, which helped inventory grow faster than normal in 2022.
So the fact that housing demand stabilized and inventory is now negative year over year needs the context that 2022 was a once-in-a-lifetime event. As you can see in the chart below, 2023 inventory growth is very slow compared to 2022.
The other big story with housing inventory is that new listing data has been trending negative year over year since the end of June 2022. A traditional seller is also a traditional buyer, and certain homeowners have refused to buy their next home with mortgage rates above 6%.
We had new listings growth from 2021 to 2022, but that’s not the case this year. This is another variable contributing to slow inventory growth, which has now turned negative in the weekly listings.
Compare the new listings data last week to the same week in recent years:
2023: 62,466
2022: 91,530
2021: 80,289
My concern lately is that we have seen four straight weeks of mild declines and are about to head into the seasonal decline period of new listings. This is one data line I will track like a hawk because it will be a negative for the housing market if this data line makes a noticeable year-over-year decline trend in the second half of 2023.
The 10-year yield and mortgage rates
For those who have followed the weekly Housing Market Tracker articles, I always focus on jobless claims data as it’s the critical data line at this point of the economic cycle for me and my forecast in 2023 for mortgage rates.
Last week we had a big move in the 10-year yield because jobless claims came in better than anticipated, and bond traders were caught off guard selling bonds on the news and sending mortgage rates above 7% again. As you can see in the chart below, that big spike was really about jobless claims getting better.
The following day, the PCE inflation data showed a cooling down in headline inflation year over year. Core PCE inflation is a bit more sticky than headline inflation, however, bond yields fell after that report and bounced back at the end of the day.
In my 2023 forecast, I wrote that if the economy stays firm, the 10-year yield range should be between 3.21% and 4.25%, equating tomortgage rates between 5.75% and 7.25%. As long as jobless claims trend below 323,000 on the four-week moving average, the labor market stays firm, which means the economy remains healthy. Jobless claims have stayed below this range all year, and job openings are still at 10 million.
I have also stressed that the 10-year level between 3.37% and 3.42% would be hard to break lower. I call it theGandalf line in the sand: You shall not pass. The setup for the 10-year yield to stay in the range is intact.
The counter to my 10-year yield range would be if the economy here or worldwide starts to accelerate higher; that would be a valid premise to get the 10-year yield above 4.25%. Considering our economy this year, the 10-year yield and mortgage rates look about right to me.
Now the one thing that has changed in 2023 is that since the banking crisis, the spreads between the 10-year and mortgage rates have worsened, making mortgage rates higher than I anticipated versus the 10-year yield, which is not a positive for the housing market.
We haven’t seen anything in the data showing that it’s been improving recently. This is a big deal as we have seen housing inventory not get much traction with higher rates and hopefully in the future, lower rates can entice some sellers to move.
On jobless claims data, I always stress using the four-week moving average with this data line because we do have times when this data line can get hectic week to week. Therefore, I only believe the low jobless claims print once I see weeks of this data line improving. So, it will be critical over the next two weeks to see if this decline was a one-time blip in the data, which we have seen from time to time. As you can see below, that was a significant drop week to week, which looks abnormal to me.
Purchase application data
Purchase application data has surprised people with three weeks in a row of growth, while mortgage rates have been near 7% during this period. This now makes the positive count since Nov. 9, 2022, 20 positive prints vs. 11 negative prints. The year-to-date numbers are 13 positive vs. 11 negatives after making some holiday adjustments to the data line.
What do these numbers mean? They just mean that housing data has stabilized; nothing in the data shows decent growth after that first good move from November to February. However, the fact that housing demand has stabilized is a big deal because last year, we did have a waterfall collapse in the data, as shown in the chart below. The only downside to this is that we haven’t had the housing inventory growth I would like.
Now the year-over-year decline was down to -21%, which was the lowest since Aug. 24, 2022. However, we all have to remember that the second half of 2023 will have much easier comps, so even if demand stayed the same the rest of the year we will have some positive year-over-year data at some point.
Be careful in reading too much into the better year-over-year data we will see in the future. The most recent pending home sales print came in as a miss from estimates, but the existing home sales data is still trending in the range I thought it would be in since I believed that first big print we had a few months ago was going to be the peak for year. When demand is coming back in a big way, purchase apps will be positive for a majority of the weeks as we are working from such low levels today historically.
The week ahead: Jobs, jobs and jobs data
Yes, it’s jobs week once again and with four labor reports coming up on this short holiday week, we’ll be able to see if the Federal Reserve is getting what it wants — a softer labor market. Recently, Fed Chair Powell once again stressed that the labor market is too tight and that softer labor is the way to get inflation down to the Fed’s 2% core PCE target.
Well, we have four reports this week: the job openings data (JOLTS), the ADP jobs report, jobless claims and the big one on Friday — the BLS job report — so we’ll see what happens.
So much of my COVID-19 recovery model was based on the labor dynamics being much different now, since I was the only person talking about job openings getting to 10 million in this recovery. Today as I write this, we are still at 10 million job openings, as the chart below shows.
I have a firm belief that the Fed doesn’t fear a big job-loss recession as long as job openings are this high. What they have enjoyed seeing is wage growth cooling down, as shown in the BLS job reports for 18 months now. So, for this week, we always focus on jobless claims data over everything else, but be mindful of the job openings data since the Fed wants to see this go down, and the wage growth in the BLS jobs report data.
Federal Reserve analysts have published a paper describing what they call the Twitter Financial Sentiment Index, or TFSI. The tool aims to gauge how investors and consumers feel by tracking social media posts about finances and credit markets. The Fed stresses that the document’s conclusions are tentative and preliminary.
A financial advisor can help you build a long-term investment plan.
What Is the Twitter Financial Sentiment Index?
The TFSI is a new tool in development by a group of economists at the Federal Reserve. While preliminary and, as the authors stress, still tentative, this tool measures investor sentiment and the consumer marketplace based on information gathered from social media posts.
The research is titled More than Words: Twitter Chatter and Financial Market Sentiment, written by Federal Reserve economists Travis Adams, Andrea Ajello, Diego Silva and Francisco Vazquez-Grande. It’s part of a discussion series run by the Federal Reserve in which economists explore new ideas, so this research doesn’t necessarily reflect the positions of the Federal Reserve or Board of Governors themselves.
In this case, the economists behind the TFSI wanted to explore whether “social media activity [can] carry any meaningful signal on credit and financial markets’ sentiment.” Essentially, when people post about the market online, does this accurately reflect their opinions? And furthermore, can economists pull any useful data from what is, effectively, a massive, real-time survey?
To answer that question, they turned to Twitter. The economists built a real-time sentiment index that pulled more than four million single tweets from 2007 to April 2023, searching specifically for posts that contained words and phrases pulled from financial and market dictionaries. So, for example, their system might flag a tweet with the phrase “bonds” or “assets” to include in the index.
Using a natural language processor, which is software that analyzes text for what the author intended to communicate, the index gives each tweet in its database a positive or negative flag depending on how the post talks about the market. Then, in aggregate, the index produces an overall current sentiment of the market. If most of the recent tweets are talking about the market confidently, the TFSI registers a positive sentiment. If lots of people are tweeting about selling or hoarding cash, the TFSI registers a negative sentiment.
The authors say that this binary approach of positive and negative works better than trying to assess how positive or negative a given tweet seems. Their goal isn’t to judge the strength of an individual poster’s emotions, but rather to judge the overall emotional state of the market at large. And, they say, it appears to work.
What Exactly Does the TFSI Measure?
Among other sources of data, economists rely heavily on surveys and price trends to make predictions and policy assessments. Surveys like the famous University of Michigan Consumer Sentiment Index gather data by directly asking people about their financial situation and choices. Price trends measure the current prices in a market and compare them to historic patterns to make predictions about what will happen next. In both cases, economists effectively look for massive amounts of data from which to pull trends.
The TFSI takes a similar approach. It is, in effect, an always-on survey, in which the authors look for patterns in how people talk about their finances and market issues. As with all matters related to social media, though, the question is whether this information is reliable. When people post on Twitter, does it reflect their true position? According to the economists involved, the answer is yes.
What the TFSI Reveals
“We find,” wrote the authors of the index, “that the Twitter Financial Sentiment Index (TFSI) correlates highly with corporate bond spreads and other price- and survey-based measures of financial conditions.”
The authors also state that they “document that overnight Twitter financial sentiment helps predict next day stock market returns. Most notably, we show that the index contains information that helps forecast changes in the U.S. monetary policy stance: a deterioration in Twitter financial sentiment the day ahead of an FOMC statement release predicts the size of restrictive monetary policy shocks. Finally, we document that sentiment worsens in response to an unexpected tightening of monetary policy.”
Among other correlations, they say, the TFSI has a few key uses.
First, it is quite adept at predicting next-day stock market returns. Strong real-time sentiment tends to correlate with gains in the next 24 hours, while a negative sentiment tends to precede losses. “This fact,” the authors write, “speaks to the ability of tweeted sentiment to reflect information that will later be included in stock prices once U.S. markets open.”
Second, and of more interest to economists, the TFSI “correlates highly with market-based measures of financial sentiment.” This includes indicators like bond and corporate bond spreads, as well as survey-based metrics like the Michigan sentiment index.
Potential Uses of the TFSI
This makes the new index a potentially useful tool for monetary policymaking. Based on how people discuss monetary policy and financial sentiment, the authors suggest that the TFSI can help “predict the size of restrictive monetary policy shocks.” In other words, it can “predict the market reaction around the FOMC statement release. We also find that the TFSI worsens in response to an unexpected tightening in the policy stance.”
Essentially, it can help the U.S. central bank measure how much the economy will slow down after it reduces the money supply (typically by raising interest rates).
Of course, there are limits to even the best tools. The TFSI is a new metric, and as such its results are still preliminary. It remains to be seen whether this will remain a valuable tool, especially once social media posters can access the TFSI itself, which could create a sort of feedback loop where index results begins to influence the index’s underlying data.
And the TFSI is a linear tool. It can signal whether people feel good or bad about the market, and the strength of that general sentiment but doesn’t provide context or lateral details such as whether they feel good about some issues and negative about others.
Still, in its early applications, it looks like the TFSI might have found a use for social media after all.
Bottom Line
Federal Reserve analysts have developed a tool for gauging investor and market sentiment around the bank’s policies and pronouncements. Called the Twitter Financial Sentiment Index, it measures the economy by listening to millions of tweets. According to the paper, the tool “helps predict the size of restrictive monetary policy surprises, while it is uninformative on the size of easing shocks,” when the FOMC eases its federal funds rate.
Investing Tips
A financial advisor can help you build a comprehensive investing plan. Finding a financial advisor doesn’t have to be hard. SmartAsset’s free tool matches you with up to three vetted financial advisors who serve your area, and you can have a free introductory call with your advisor matches to decide which one you feel is right for you. If you’re ready to find an advisor who can help you achieve your financial goals, get started now.
Investors use a wide range of metrics and indices to make their decisions. Whether you’re buying assets that you’ll hold for years to come or looking to make a profit day trading, it’s worth familiarizing yourself with some of the most common, like the Consumer Price Index.
Eric Reed
Eric Reed is a freelance journalist who specializes in economics, policy and global issues, with substantial coverage of finance and personal finance. He has contributed to outlets including The Street, CNBC, Glassdoor and Consumer Reports. Eric’s work focuses on the human impact of abstract issues, emphasizing analytical journalism that helps readers more fully understand their world and their money. He has reported from more than a dozen countries, with datelines that include Sao Paolo, Brazil; Phnom Penh, Cambodia; and Athens, Greece. A former attorney, before becoming a journalist Eric worked in securities litigation and white collar criminal defense with a pro bono specialty in human trafficking issues. He graduated from the University of Michigan Law School and can be found any given Saturday in the fall cheering on his Wolverines.