Recency Bias and the Housing Market

In 2021, in the wake of the pandemic, the housing market ascended, prompting pundits to compare the housing market at the time to the bubble precipitating the 2008-2009 financial crisis — and project an impending bubble. In response, we explained why we believed that housing price increases were not indicative of a bubble in a series on the housing market.[1] First, we posited that it is normal for markets to increase and decrease without reaching the extremes of “bubbles” and “busts.” In addition, we pointed to recency bias as a potential logical fallacy that could lead some to predict a bubble even without the evidence to back it up:

“Recency bias, defined as the belief that recent events will reoccur, is a common cognitive error in which a disproportionate amount of importance is given to relatively rare events. In the context of the current housing bubble panic [as of September 2021], this is relevant for two reasons. First, on a large scale, some people view the past nine months of housing gains— as shown by the Case-Shiller National Home Price Index — as a housing bubble due to their experience of the housing bubble that precipitated the Global Financial Crisis in 2009. Second, on a smaller scale, others view the last nine months of data and assume that the trajectory of housing gains will continue. Though these applications of recency bias are different, they both lead to what we believe are erroneous conclusions.”

Since we wrote those articles, the Fed has substantially increased interest rates to combat rising inflation. In response, the housing market is slowing substantially and many of the same people who predicted a 2006-type bubble are now calling for 2008-type bust — and just as we felt comparisons to that bubble ascent were overstated, we also feel that today’s comparisons to the subsequent bust in are overstated.

In our blogs we mentioned three factors differentiating the two ascents: 1) Reduced housing supply, 2) financing quality, and 3) demographic tailwinds. We believe that the decline will also be softened by these factors —preventing a repeat of 2006-2011’s decline in terms of length and magnitude.

That said, we do not wish to understate the situation: the housing market will slow, and prices will decline. However, unlike what we saw 10-15 years ago, this will be healthy, and we believe it is extremely unlikely that a systemic problem will develop. In fact, we believe if people can look past their recency biases, they will find the housing market of the late 1970’s and early 1980’s to be a more apt comparison.

For the rest of the piece, we’re going to draw some comparisons between the three time periods mentioned. To make it easier to differentiate between them, we’ll utilize the following terminology:

• Period A: 1979-1983
• Period B: 2005-2012
• Present: 2020-Today

To recap, we posit that though some would compare the Present to Period B due to recency bias, Period A would be a more apt comparison.

In Period A, real home prices declined more modestly than Period B — From July 1979 to October 1982, there was a 12.2% decline, and from December 2005 to February 2012[2], there was a 35.7% decline.

In addition, we believe that Period A and the Present share three factors differentiating them from Period B[3]:

1. Housing Supply. We look at housing supply in two areas: a) housing starts and b) months of supply. Housing starts were elevated for more years going into the Period B peak than in Period A. In other words, in Period B, builders added houses to supply at higher levels for more months than they had in Period A, which produced larger excess housing supply when demand fell.
2. Financing Quality. Many mortgages issued during Period B were poor quality, which begat a swath of distressed sales when the mortgages when bad. These distressed sales lowered prices even further, perpetuating price declines beyond what would have been the case through the sale of merely organic supply. We do not find this condition today, and was not the case in Period A, so it is unlikely that the market would experience a similar cascading feedback loop of distressed sales forcing prices lower and then leading to even more distressed sales.
3. Demographic Tailwinds. In the Present, demand for housing should remain more robust than it did after Period B because there is a more balanced population pyramid and stronger household formation. This was also the case in Period A as the plethora of baby boomers forming households put a floor on housing demand.

Other similarities between the Present and Period A include 1) an inflationary environment highly driven by a spike in energy prices, 2) a hawkish Fed raising interest rates at a sharp trajectory, and 3) a sharp increase in the housing carrying cost.

Though the Present shares more similarities with Period A than Period B, it is not completely the same as Period A — after all, no periods are perfect replicas. Importantly, the most salient factors tend to favor today’s housing market vs. that of the late 1970s:

• Less Entrenched Inflation. In Period A, inflation was far more entrenched than it is in the Present, from which we can infer that the Federal Reserve will not need to tighten monetary policy as much in the Present than it did then — in the Present, inflation has been over 4% for merely 14 months, and Period A experienced essentially a decade of inflation up to that point in time.
• Low Home Inventory. Additionally, the inventory of homes for sale today is extremely low by historical standards, even as inventory rises a bit as some sellers try and get out in front of looming market declines.

Both factors lead us to believe that as mild as the Period A housing decline was as compared to the that of Period B, the decline occurring in the Present may be even more muted. In the end, it will really come down to how much the Fed needs to raise rates to halt inflation: if our null hypothesis of a more modest tightening plays out, this will further buoy our argument.  However, if inflation ends up being more entrenched than is currently believed, thus necessitating a more hawkish Fed, then that would add a further headwind to housing prices. That said, we still posit price declines would still be far from the those experienced in Period B.

Conclusion

It is human nature to look for patterns — and stories — that explain the world around us. This can often be helpful; however, sometimes, we try to see a pattern that doesn’t exist. Recency bias explains humans’ tendency to expect recent events to reoccur. In the case of the economy, some are inclined to predict a housing bubble because they remember the Great Financial Crisis in 2008-2009. Examining some key factors that differentiate a bubble from a regular market cycle — housing supply, financing quality, and demographic tailwinds — lead us to believe that the current decline in prices isn’t a bubble at all. Rather, it is more similar — and potentially less severe — to the housing market in the late 1970s/early 1980s.

[1] Balentine, Housing Déjà vu All Over Again, Balentine, Housing Déjà vu All Over Again – Part II
[2] St. Louis Fed, Real Residential Property Prices for United States
[3] We explained these differentiators in Housing Déjà vu All Over Again