Purchase apps are very confusing this year, but new listings are key

I have been tracking purchase application data for over 14 years, and 2025 is likely the most confusing year for market participants that I have ever experienced. After today’s report of 22% year-over-year growth, I feel compelled to try to make sense of the data. Let’s dig in.

Purchase application data this week, and what it has looked like over the last 25 weeks

• Week-to-week growth: +3%
• Year-over-year growth: +22%
• 25 consecutive weeks of YoY growth
• 12 consecutive weeks of double-digit YoY growth

This data line has been quite confusing so let me clarify what is happening. As we can see in the chart below, we have clear year-over-year growth in the data from 2025 compared to 2024.

The year-over-year growth data

When we look at year-over-year growth, we are starting from a historically low level; last year was particularly negative. Specifically, when mortgage rates increased from 6.63% to 7.50%, purchase applications showed predominantly negative data. In fact, we saw no year-over-year growth in the first half of 2024. Additionally, in the first half of 2024, the week-to-week data had 14 negative readings, two flat readings and only two positive readings.

Mortgage rates have been lower this year compared to last year. However, that isn’t the most significant variable in 2025 for purchase application data growth: it’s really the growth in our new listings data. Most sellers are also buyers, which means they’re filling out purchase applications as they prepare to sell their homes. With new listings data returning to somewhat normal levels this year compared to last year, we can connect the dots to the year-over-year growth. Of course, those sellers still need a buyer for their home to sell and buy their next property.

Now, the week-to-week data isn’t as robust as the year-over-year data, and it is not as strong as when rates dropped from 6.64% to 6% in the past.

What’s happening with the weekly data

Year-to-date weekly data:

• 13 positive
• 10 negative
• 5 flats

This is a divergence in the data in 2025. For example, last year, when rates went from 7.5% to 6%, we had:

• 12 positive week-to-week data
• 5 negative
• 1 flat

In essence, a decrease of 1% to 2% weekly purchase application data percentages in some weeks could have led to flat or negative weekly data for the year. This year, we have had five weeks that were flat week to week. If those flat weeks were down just 1%, we would have more negative weekly data than positive data.

Additionally, purchase applications typically take 30 to 90 days to be reflected in sales data. Only recently have our weekly pending sales data shown year-over-year growth, and that growth is only relative to a historically low comparison. In fact, it’s the lowest benchmark we’ve ever had. This is why I often say that I have only seen noticeable improvement in housing data when mortgage rates drop from 6.64% to 6%. In that range, the weekly data shows more positive growth.

Always remember that most sellers are homebuyers, but many sellers need to find a buyer for their home before they can sell and buy another. The existing home sales report is coming today, I don’t expect much to happen in today’s report.

Remember, the year-over-year comparisons are extremely low for today’s report and will remain so for the next four months. Therefore, any growth or even a slight decline is essentially a low-comparison story. I have been tracking this data for over 14 years. While it has consistently trended with existing home sales since the 1990s, this year has presented a historically abnormal setup. This is precisely why we track live data on a weekly basis in our Housing Market Tracker article.

I understand the confusion on purchase apps. When you compare the year-over-year growth in the data and express it as a percentage relative to sales, it may seem like existing home sales should be much higher. However, I want to emphasize that this data is not best analyzed using percentages; it is more useful for identifying trends. We can clearly see a divergence between the year-over-year data and the weekly data. Additionally, we can connect the year-over-year data to the increase in new listings.