This article has two parts. The first is an analysis of a terribly misleading position presented by Mark Perry and Alan Reynolds. (Links to each will follow.) A position has been has been taken by Reynolds and basically endorsed by Perry. The second part of the article is a presentation of data that has either been ignored by Reynolds or dismissed as irrelevant. Part1 was the refutation of the opposition’s arguments; the second part (here) is the presentation of the author’s arguments. Thus another Great Debate© is joined.
Part 2. The Real Housing Story
I will avoid repeating much of what was presented in Part 1.
Let’s first look at Mark Perry’s graph:
Perry writes the following about the graph:
…the chart above provides further evidence of the significant disconnect between the Case-Shiller Home Price Index (10-city composite above for Boston, Chicago, Denver, Las Vegas, LA, Miami, NYC, San Diego, San Francisco, and Washington D.C.) and the FHFA U.S. House Price Index. As Alan Reynolds pointed out, the FHFA U.S. House Price Index based on 50 states increased by 1.07% in the third quarter of this year vs. the previous quarter, the largest quarterly gain since the 1.30% increase in the fourth quarter of 2006. Based on the FHFA House Price Index, there’s no threat to the U.S. economic recovery.
Let’s look at the two trend lines for the pre-bubble years:
The decline in prices to reach the two trend lines is 23% for the FHFA curve and is 39% for the Case-Shiller-10 curve. However, the short period of time establishing the baseline may be misleading. We have no Case-Shiller data before 1987 but there is a longer history for FHFA:
The longer baseline establishes a smaller distance to current prices (18%). It is not unreasonable to assume that the Case-Shiller data might show less of a price difference if that data also extended back another decade. It is reasonable to conclude that the Case-Shiller “over-valued” amount may be considerably less than the 39% indicated by the short baseline.
At any rate, the FHFA national data indicates homes are still over-priced by 18%.
CoreLogic indicates a similar situation in their national home price index, as shown in the following graph:
Case-Shiller also reports a 20-metro housing price index. The two Case-Shiller indexes are shown in the following graph along with the FHFA curve. Note: the Case-Shiller Comp-20 data starts only in 2000. The FHFA data has been normalized to the same value as Case-Shiller Comp-10 as of January 1, 1987.
In his Op Ed in Investor’s Business Daily Alan Reynolds wrote:
A dip in the Case-Shiller moving average of home prices in 20 cities for August to October is said to be “troublesome headwind” for the economy in 2011, and “markets such as Sacramento, Las Vegas and parts of Arizona and Florida are at risk of more declines.”
Some of those cities may indeed account for a significant share of the Case-Shiller index, because that index covers only 20 cities (and Sacramento, the centerpiece of the story, is not one of them). However, a few troubled cities in a few states do not represent the entire nation.
We have established that, though the problem may well be most serious in “a few troubled cities”, the FHFA national numbers clearly show a bubble not yet collapsed for the entire country. Furthermore, the Case-Shiller Composite 20 metropolitan areas have a combined population of 104.4 million. See table:
This is about 1/3 of the U.S. total population. So if the CS-20 represents 1/3 of the country and FHFA represents the entire country, then we have all we need to calculate what the housing bubble looked like for the “other 2/3”. That has been done and is shown in the following graph.
It is clear that, relative to January 2000, the “other 2/3” bubbled more slowly than the 1/3 represented by CS-20. But, curiously, they did eventually reach a relative price level at their early 2009 peak that was higher than the CS-20 relative prices at the same time. And the “other 2/3” still remain overvalued relative to the CS-20 supposedly dominated by “a few troubled cities.”
Caveat: A reader has pointed out, that although the FHFA represents all states, it does not represent all mortgages. They are largely conforming mortgages and do not include jumbos. Another reader has submitted that FHFA does not include FHA and VA mortgages, which are largely on the lower end of the conforming group. I have not independently verified either representation. So the treatment here should be considered as an approximation rather than an exactly controlled measurement.
There is one final thing that reinforces the degree to which homes are overpriced on a national scale. The graph below shows the bubble in the ratio of home prices to income. The average ratio was about 5.5 for the years 1975-2001. To get back to that ratio from the current level requires a decline in price of 19% at constant household median income.
Yes interest rates have been low and that increases home affordability. However, in the years 1975-2001 mortgage rates ranged from around 6-7% up to 12-14% and the ratio of prices to income fluctuated very little. We are now below the low end of that earlier range with mortgage rates near 5%. It just isn’t reasonable to justify the high ratio based on an interest rate argument when the current rates are so close to the prior range minimum.
When one looks at the 120 year history of home prices adjusted for inflation, as done by Robert Shiller, there is a long historical norm established. The following Shiller data, graphed and annotated By David Rosenberg, tells the story:
We are currently one standard deviation above the 120-year average. And the irrelevance of interest rates to home prices during a housing market depression is obvious when one looks at the ultra-low interest rates of the 1920s and 1930s accompanied by home prices one standard deviation below the historical average.
There is no way that a thorough look at the data can lead one to rationalize that the housing market is poised for recovery. Sorry Mr. Reynolds and Prof. Perry – I’m not buying what you’re selling.
People who represent themselves as knowledgable have two obligations to be credible:
- They must use broad and representative data.
- They must analyze that data critically.
This has not been done in this instance by Mr. Reynolds and Prof. Perry.
The Great Debate©: Will Housing be a Drag on the Economy in 2011? Part1 by John Lounsbury
21% Decline Forecast for December 2010 Existing Home Sales by Steven Hansen
New Home Sales Remain At Historic Lows by Steven Hansen
The Great Debate©: Residential Construction is Dead – Or Is It? by William C. Wheaton and Gleb Nechayev