Market Publications

U.S. Housing – Déjà vu All Over Again?

After one look at the below chart of U.S. housing prices, can there be any doubt that we have another housing bubble?

Figure 1. A Common Version of the U.S. Home Price Chart
US National Home Price Index
Source: Robert Shiller

To that, we say “not so fast, my friend!” In our writing on various financial topics, we come across a similar pattern:

  1. The media hypes a “crisis” or “opportunity.”
  2. As a result, investors/consumers pay heed to media’s “advice.”
  3. The investor/consumer is left worse off as a result.

The latest media episode du jour (among many at this time) is the reporting of another housing bubble and/or housing crash. A quick Google search on “housing bubble” turns up:

Something to which we have alluded bears repeating: the financial media are not interested in giving you facts; Rather, they are interested in selling you a story. This, of course, by no means dissuades them from publishing contradictory pieces Case in point, Bloomberg three weeks apart:

So, in page one of a Google search, we have (in no particular order):

  1. A housing bubble is looming.
  2. A housing bubble is here.
  3. A housing crash is coming.
  4. A housing crash is here.

Like the financial media, we share a goal of being informative. However, unlike the financial media, we strive to be correct, which is always easy retrospectively and not so easy prospectively. So, what is our take on the situation?

While we recognize the tremendous price appreciation since last summer is in excess of historical precedent, we do not believe there is a bubble. Moreover, we would suggest that price gains, after a period of cooling off, are likely to continue moving higher in the coming years.

Our thesis is a three-legged stool:

  1. current prices do not constitute a bubble – perception is not reality,
  2. a pervasive bubble mentality is not yet present, and
  3. fundamentals justify both price appreciation to date and further price gains looking forward.

In this piece, we will address the first two legs of the stool. Stay tuned for Part 2, in which we will address the third leg.

Current prices do not constitute a bubble.

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, 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.

Exceeding Prices Found from a Bubble Peak Does Not Indicate a Bubble.

A common refrain we have read is that prices are higher than even during the housing bubble that preceded the Global Financial Crisis. To that, we say, is that such a surprise? Viewing another recent bubble can help us understand both the amount of time it takes prices to recover after a bubble crash and show that meeting or exceeding a prior peak is not necessarily indicative of a bubble.

For example, let’s compare the housing price trajectory to the 2000 dotcom (i.e., NASDAQ 100) bubble (Figure 2).The NASDAQ 100 rose 127% in the final year of the bubble and then experienced a 79% decline in ~2.5 years. The peak reached at the height of the bubble was regained in just over 15 years.

Figure 2. The NASDAQ 100 bubble, which was much larger than the housing bubble, took 15 years to reach its prior high.
NASDAQ 1999 - 2015
Source: FactSet and Balentine

The 2002-2006 housing bubble was much smaller than the NASDAQ bubble; housing prices rose 42% over a three year period rather than 127% in one year. If we consider that it took NASDAQ 14 years to reach its previous high after the bubble, it is reasonable to expect that 15 ½ years after the housing bubble peak in March 2006, prices would be above their prior high.

Using the Wrong Type of Chart Can Lead to Misleading Conclusions.

Economists use charts to interpret the relationship between prices of goods and other factors, such as time in an effort to discern patterns that can help them predict market outcomes.

In Figure 1, we plotted month vs. housing price to view the way that the prices have changed, and may continue to change, over time. This is an example of an arithmetic relationship, which assumes that each unit, or point representing month + price value, changes at the rate of a certain percentage. In other words, the percent change between unit values is consistent. In this case, the percent change between months and prices is assumed to be equidistant, which you can tell by the way the y axis plots price values rather than percentages. Once we have plotted the values, we can draw a trend line using the slope of the graph. We can use the trend line to predict future values.

Not all relationships between data points can be accurately represented arithmetically because the percent change between values is not consistent. In these cases, we must use a different kind of graph to chart the relationship between them. Logarithmic charts use the percentage of change to plot data points, so, the scale prices are not positioned equidistantly. In this case, the y axis plots percentages rather than prices and the x axis plots months.

The reason this is important is because Figure 1 is an arithmetic chart; it assumes that percent change is consistent. It is useful to look at the price of houses over time; however, because it does not reflect the different percent of change between prices, it distorts the most recent price gains at the expense of earlier price gains.

For example, consider that the percent of change from 60 to 80 is equal to the percentage of change from 180 to 240, 33%, yet the distance from 180 to 240 on the arithmetic graph is larger than the distance from 60 to 80 on the graph. This is misleading because it makes it seem that the increase from 180 to 240 is much more dramatic. When we analyze the data on a logarithmic graph, which represents percent of change between data points, and when we add in more data, back to 1968, we see a more accurate picture of the rate at which housing prices are increasing (Figure 3).

Figure 3. A More Accurate Depiction of the U.S. Housing Price Chart, both in scale and duration.
US National Home Price Index
Source: Robert Shiller

When viewed from this perspective, the chart looks less like a bubble and more like a consistent upward trend, which can be seen even more effectively in Figure 4 when we insert a trendline.

Figure 4. Including a Trendline Completes the Picture.
US National Home Price Index with Trendline
Source: Robert Shiller and Balentine

The trendline has 96% correlation with the actual data and demonstrates that although current prices are higher than they were at the peak in 2006, many would be surprised to find out that prices are under where they are expected to be based on the underlying trend. In other words, put differently, the 50+ year trendline suggests housing prices should be higher than they are now.

Let’s looking at this from a slightly different angle. Figure 4 demonstrates times when U.S. home prices have been above and below trend expectations. But it’s hard to get a precise grasp for how much deviation from trend there is. All asset markets deviate around their trends. If these deviations are relatively modest, that is in the normal course of market ebbs and flows. If, however, the deviations are egregious, then that is the hallmark of a bubble (or, in the bearish case, the hallmark of an overcorrection). Again, for comparison purposes, let’s start with the NASDAQ 100 to see what a massive bubble deviation looks like (Figure 5).

Figure 5. The NASDAQ 100 bubble in 2000 demonstrated a massive trend deviation.
Percentage Deviation from Trendline
Source: FactSet and Balentine

Here we see a clear deviation from the trend far above anything reasonable or in line with historical standards. Now, let us take a look at the housing market back to 1968.

Figure 6. Though not as egregious as the NASDAQ, housing prices were a bubble in their own right.
Though not as egregious as the NASDAQ, housing prices were a bubble in their own right
Source: Robert Shiller and Balentine

Two takeaways come to mind right away: 1) housing prices were not as large a bubble as the internet bubble, and 2) housing prices still remain well below the trend, even with the recent coronavirus-induced spike (which will be addressed later).

A pervasive bubble mentality is not yet present.

Zooming out for a moment, defining a general market bubble is trickier than it would seem. All industries are going to have their fair share of booms and busts, so what makes these different from the idea of a bubble?
Empirical research of media publications shows more opinions at this time calling housing a bubble than not a bubble; this is not a bubble mentality; rather this is a wall of worry that will provide the opportunity for prices to ascend higher, not fall lower.

This is a key point – it is only after many skeptics throw in the towel and concede defeat before a bubble can truly form. There are the classic, oft-quoted 1600s tulip mania, 1700s South Sea Company, 1920s U.S. stock market, 1980s Japanese economy, 1990s dot com stock market, and 2000s global housing, all of which featured a key tenet:
the majority of the investors went along with the crowd psychology that takes over as “new paradigms” are justified in the face of facts to the contrary, thereby relegating the skeptics and contrarians to a relatively small subset.

Now, this is not to suggest that there all deviations from trend are a problem. All asset markets deviate around their trends, providing shorter-term traders opportunities to buy low and sell high around the trends. The problem is where a normal deviation becomes a feedback loop, where price appreciation begets further price appreciation. Such a dynamic is driven by broad euphoria carries prices beyond values supported by fundamentals or by the typical economic laws of supply and demand so long as there is a “greater fool” to whom to sell it. In other words, rational people will ignore valuations that they do not necessarily believe as long as they believe there is someone else who will buy it for an even higher value.

Of course, market participants don’t view themselves as paying up; rather, the rationalizations of “it’s different this time” pervade. The human tendencies described above lead to a fairly predictable outcome: After an idea takes hold and a market for the product forms, people believe both that they are among the few who have spotted the trend early, and that they will be smart enough to pull out at the right time. Of course, this goes on until the pool of buyers dries up and the “greatest fools” are left holding the bag. Shortly before this, the financial media and then the mainstream media are running with the story as their top headlines. Case in point in Figure 7, the cover of Time Magazine in June 2005, merely nine months before the price peak:

Figure 7. When it hits the magazine covers, generally the end is near
Time Magazine Cover - Home Sweet Home
Source: Time Magazine

Suffice it to say that such a bubble mentality is far from ubiquitous today. In fact, we would posit quite the opposite; people are looking for any reason to scoff and call for a bubble as asset prices rise.

We will delve into this dynamic more in Part 2 of this blog, in addition to looking at: 1) the third leg of the stool – fundamentals justify both price appreciation to date and further price gains looking forward, 2) why bubbles happen, and why we think bubble callers are out en masse, and 3) where to go from here for housing.

Contact Us

Looking for guidance managing your wealth? Balentine is committed to providing the education and advice our clients need to realize their goals.