War-Era-Like Housing Market, Crash Ahead

A for sale sign is seen near a house in South Pasadena, Calif., on April 24, 2020. Frederic J. Brown/AFP via Getty Images

Commentary

The Nobel prize winner Robert Shiller warns about the risk of the U.S. housing bubble bursting. In 2013 Robert Shiller (together with colleagues) was awarded the Nobel Prize in Economic Sciences for his empirical analysis of asset prices.

Although he has been famous for stock market studies, his century-long housing price series is the standard dataset used in long cycle studies nowadays. Even though he did not collect and compile the old annual data, integrating them into one single series together with his monthly series (the Standard and Poor/Case-Shiller housing price index) already greatly facilitates most real estate researchers.

For the last century, there have been studies on long-cycle. Housing is one of the typical sectors here because a housing cycle is generally a multiple of a few business cycles. Most of the housing price series over time and across the globe suggest some 15 to 20 years for a housing cycle, That is, the peak-to-peak or trough-to-trough duration. In the good old days, when the housing price series was unavailable, other indicators, including land price or construction activities, were used. In some cases, growth instead of level exhibits better cyclical regularity.

If the current housing price behaviour were to compare with a particular history, then the 1940s would be appropriate. The 1940s is better than comparing now with the 1970s despite recent inflation development being highly comparable to that era (see my recent piece in The Standard). Even though these three periods—the 1940s, 1970s, and now, all have inflation trending upward, now the economy is overheated like in the 1940s when the U.S. unemployment rate was one percent (refer to my piece a few weeks ago about Fed chair Eccles). Now, and in the 1940s, unemployment is low, unlike in the 1970s when unemployment was high.

The fate of overheating or bubbling is simple: A burst.

Based on Robert Shiller’s dataset, the accompanying chart suggests a so-called “history repeating” in about 77 years time gap. However, the match between the new and old series is far from perfect. Only euphoria and panic parts exhibit a better relationship, and normal times are not highly correlated. Interestingly, literature documented the U.S. housing market cycle to be 18 to 19 years. This period is close to (but slightly shorter than) one-quarter of 77 years. A long cycle is multiple short ones.

The reason for such a regular pattern is not well understood. Some contribute this to wars, which exhibit similar cycles. However, there have been no world wars for a long time, but frequent small-scale wars spring up everywhere all the time. Moreover, comparing one type of cycle to another does not explain anything. More reasonable speculation is along the microdynamics from excess demand to excess supply. Yet it is never easy to argue why such dynamics exhibit stable duration, which involves the transmission of information, development of herd culture, and other factors.

Anyway, history is predicting a very bad future. Had the recession happened right now, the scenario would have been better. Yet the strong employment figures suggest that both overheating and rate hikes will be pushed to the extreme. The bounce back will be strong too.