For many years, I have been adamant that following the Mortgage Bankers Association‘s purchase application data is critical for predicting what will happen in the housing market in the coming year. 

Housing market news can be over-hyped depending on the source, and this is especially true since the housing crash. Now it seems every soft print is thought to be a harbinger of another imminent collapse of the market.  

To help you navigate the sea of both positive and negative hyperbole that can make up the housing market commentary, last year, I started a weekly purchase application tracker on Twitter.

The context for this metric, however, is even more important than the nominal value.  Here are some guidelines on how to interpret this data line.

1. Seasonality: The second week of January to the first week of May is the period that makes up the bulk of the upfront housing demand.  

Purchase application numbers during these “heat months” should be viewed as the most crucial for the calendar year. By the end of February, we have a pretty good idea of how the year will shape up. 

2. The trend is your friend: The year over year comparison data is a more important predictor of future demand than the week to week data lines. 

While the week to week variations in the data can make for sexy headlines, they are not the most meaningful predictors of where the market is headed for the year. Trust only the year over year data.

3. Total mortgage application headlines need context: If you’re looking for direction just on spring season year over year demand trends, be mindful that the total application headline counts refinancing in it. 

You can have big swings on the year over year data on the entire mortgage application side and not have it mean much to the spring selling season. I have seen this mistake happen a lot because people forget about the refinancing aspect of the data.

Last year, we had a mini refinancing boom from buyers of homes in 2017, 2018 and 2019, as they had bought houses with higher rates. Now that a lot of them have refinanced, the whole headline year over year data – especially past April – may look bad, but it doesn’t mean home demand is falling. 

4.  Don’t assume: The percentage of growth or decline in purchase applications does not predict if sales will rise or fall by the same percentage. 

We can have 5-7% growth year over year, in purchase application data, and existing home sales can still be down for the year.

In the last two years, we have had slightly negative growth and an increase in inventory with positive purchase applications for the majority of the years.

We did have some negative prints during the heat months of 2019, something that hadn’t happened since 2014 when sales were down year over year. Remember this is a survey. The survey is suitable for trend direction, not exact sales.

The purchase application index in 2014 was at an all-time low when adjusted to the population. The index has been rising ever since. The best growth year was 2016 when we had over 25% growth year over year in the heat months. We didn’t have a 25% growth in sales that year.

Last year, due to the negative prints during the heat months, overall growth was only 3.7% year over year. The purchase application data from 2014-2020 doesn’t look anything like 2002-2005. 

From this, we know that we don’t have an overheating housing cycle.

Again, be careful to jump to a conclusion from weekly data alone. Slow and steady wins this race.