COVID-19 Impact on Housing and Mortgage Markets: Part 2 - Homebuyer Pool

Posted By Mark Watson on May 01, 2020
This blog post is from 2020 and may be outdated. For the latest insights, visit our main blog feed

Our longtime clients know that one of the things that differentiates us from other mortgage forecasters is a focus on the demand side of the mortgage purchase market.  Specifically, we begin our forecast process with a consideration of the homebuyer pool1 - the segment of households that is "ready, willing and able" to buy a house.  This process provides a framework to estimate the coronavirus impact on purchase market originations.  It also explains how we have arrived at a purchase originations forecast of $752 Billion for 2020 - a staggering 37% decline from 2019 to 2020.

Last year, we began with census data showing the breakdown of the 122.9 million households in the U.S.  Nearly 65% were homeowner households, and of those, 56.3 million had a mortgage. (See Figure 1).

2019 Households by HO Status Final
However, not all of those households were truly “able” to buy a house in 2019.  We made deductions from that aggregate pool of households to derive an adjusted homebuyer pool.

We figure that recent homebuyers weren’t likely to buy another new house and that very recent refinancers weren’t either.  Moreover, many other households would be disqualified from getting a mortgage for their new home for financial reasons:   unemployment, delinquencies/foreclosures on existing homes, homes underwater on existing mortgages, and other credit problems.  Some of these disqualification categories would have overlapped; we made allowances in our model to account for this.  Some of these disqualification categories would have affected the household segments differently.  (Obviously, the non-homeowner segment – the renters –  wouldn’t be impacted by previous years’ home purchases or mortgage problems.) We also made allowances for that.

The resulting adjusted homebuyer pool (Figure 2) gave us a starting point for setting  our initial forecasted estimate for 2019 purchase mortgage counts.  It gave us a count of those households “able” to buy a home, 96.1 million.
2019 Adjusted Homebuyer Pool Final

However, not all of those “able” to buy a home were “ready and willing” to do so.  As it turned out, only 4.3 million households purchased a home in 20192.  That gives us a “ready and willing” factor of 4.5%, which is typical for a year when the economy is robust and households are fairly secure financially.

Now let’s turn to 2020.  In a few short weeks, impacts from the coronavirus have wrecked much of the economy.  Whether or not the recovery is fast or slow, U-shaped, V-shaped or XYZ-shaped, the home purchase market is already being severely affected and will continue to be affected for the rest of the year.  Many forecasters are struggling to estimate the impact, but our model provides an analytical framework to do just that.

In the five weeks ending April 18, 26 million people have filed initial claims for unemployment3.  If the unemployment rate gets to 20%, that will mean nearly 32.6 million people are out of work, not very far from where we are today.  In some households, multiple workers may have lost their jobs.

We estimate that this new unemployment impact will take 27 million U.S. households4 out of the homebuyer pool.  Thus, our count of “able” households will drop by almost 26% to 71.3 million (Figure 3).
2020 Adjusted Homebuyer Pool Final

Furthermore, we’ve now entered a period where even “able” households are going to be nervous about buying a new home.  And despite the extraordinary efforts of the Federal Reserve to provide needed liquidity in a period of mortgage payment forbearance, we expect lenders to tighten credit availability.  (Technically, this factor should impact our “able” calculation, but we couldn’t quantify it there.) 

With those considerations, we estimate that the “ready and willing” factor will drop to 3.8% in 2020.  This is about a 15% decline from the 4.50% factor of 2019.  That gives an estimate of 2.712 million purchase loans for 2020, a 37% decline from 2019.

What about home prices?  With such an enormous decline in home purchases, shouldn’t we expect to see falling home prices as well?  Not necessarily so. It’s all going to depend on the supply/demand balance in each market.  Initially, we expect the supply of homes available for sale to decline sharply as many sellers of existing homes de-list and new construction slows. We are already seeing that happen in numerous parts of the country as early data from March come in.

As a first approximation, we are assuming that on a national level, the demand decline will roughly be matched by supply declines and that home prices will not change much.  However, we believe it likely that prices will ultimately decline somewhat, but not by as much as they did during the 2007-2012 Housing Bust.

With prices unchanged in 2020, our 37% decline in purchase loans means that total purchase origination volume will will also decline 37%, to $752 billion.  That will be worse than the biggest annual decline during the Housing Bust, which was 32%, the drop from 2007 to 2008.

With so much information around this crisis changing on a daily basis – both on the medical side and on the economic side –  there is a large range of uncertainty around our purchase market estimate.  Whether or not you agree with our model's assumptions, this framework provides a method for decision-makers to analyze the current uncertainty in the market and prepare for the possible outcomes.


1 Part 1 of our coronavirus impacts blogs was not labeled “part 1”, but this article is the follow-up to the first blog in our COVID-19 impact series.

2 According to 2019 HMDA data from the FFIEC compiled by iEmergent, lenders originated 4,329,115 loans for $1.199 trillion for purchasing 1-4 family unit homes. 

Initial claims for unemployment insurance from the Bureau of Labor Statistics (units in millions):  Week ending 3/21 – 3.307, 3/28 – 6.867, 4/4 – 6.615, 4/11 – 5.237, 4/18 – 4.427.  Total:  26.453. 

4 Not that this affects our calculation, but we estimate that the unemployment impacts on the segments of households with mortgages, the households without mortgages, and the non-homeowner households (renters) will be 23%, 10% and 67%, respectively.

Subscribe to Get Fresh Insights

Fill out my online form.