This is the third and probably final article that examines the recent housing bubble. The previous 2 articles (here and here) are related in that they both examine data that shows a bubble is underway. The obvious piece of data to look at involves the price increase of the commodity in question (houses). Somewhat less obvious but still important is to look at corroborating data, housing inventory in the 2nd article and construction sector employment in the first article. It seems people have a remarkable ability to deceive themselves when it comes to whether a bubble is currently happening. So corroborating evidence should help persuade.
In this article we’ll try to put ourselves in the shoes of a would-be investor in the midst of these bubbles and see how we’d react. Consider Dr. Shiller’s original graph. The peak in prices was very large. But let’s suppose you had access to a grapic like Dr. Shillers updated each year from 2000 through 2006. The question is “when would you sell?”
Somewhere between 2000 and 2001 the prices in the current boom reached the prices seen at the peak of the last two booms. Investors having access to the 2001 version of Dr Shiller’s graph should have started to get nervous then. However, if they sold their investment properties in 2001 they would have missed out on another 5 years of growth. Suppose you did sell in 2001, by 2003, when the housing market was even higher, would you still have confidence that the bursting of the bubble was near? Or would you be tempted to jump back into the market? Even if you can see the bubble coming, timing the market is hard. Being confident of your convictions is hard too.
And yet this is not the first bubble we’ve been through. Recently we’ve had 3 other bubbles: Oil in the late 1970s and in 2008, Technology/Nasdaq in 2000 and Gold in the late 1970s (and possibly now). I’ve plotted these 3 bubbles on the same scale in the graph below. It is useful to see all 3 on the same graph but some scaling is necessary since the oil high ($127/barrel in 2008) is lower than the gold low ($198/oz in 1970). The data is adjusted for inflation (all dollars are in 2009 dollars) and the relative values were set to each item’s respective value in 1996. Why 1996? Because that was a period of relative calm for all 3 items. What can we learn from the graph? (Click on the graph for a larger view.)
History Does Repeat Itself
The gold bubble in 1980, the Tech/Nasdaq bubble in 2000 and the Oil bubble in 2008 each share a similar feature, a rapid rise (over a period of 1 to 2 years) followed by a similarly rapid fall. I don’t have access to Dr. Shiller’s data, so I can’t graph it together with the above graphs but it is possible to compare them side-by-side. The housing booms of the 70s and 80s were twice as long as the gold, oil and Nasdaq booms. Shiller marks the 70s boom as 3 years and the 80s boom as 5 years long. This extra length may be due to houses being harder to sell than gold, oil or technology stocks. Also, the current boom is longer still, Shiller marks it as 9 years long. By the standards of the 70s and 80s booms, 2000/2001 should have marked the end of the current housing boom. Yet it continued for another 4 to 5 more years.
Another interesting feature is that after the gold and tech/Nasdaq booms the prices stabilized somewhat higher than their values before the boom. The oil boom we’re just coming off of looks like it too may follow this pattern, but it’s too soon to tell. The oil boom in the 1980s runs counter to this but it was an odd price spike for several reasons. The US department of energy has an excellent chart with their take on the reasons why. It also appears the 80s housing boom followed this same pattern stabilizing about 5% higher after the boom than before.
Are we now in another gold bubble? If we are, it is unlike the previous gold bubble. The current upward trend in gold prices has been going on for about 10 years; not classic bubble behavior. However, if you look closely at the right hand edge of the graph (click it to see a larger version), it appears that the increase in gold prices has suddenly quickened. I doubt we were in a gold bubble in 2008, but looks like we may just be entering one now.
One reason why Dr. Shiller’s graph may not have gotten much attention in 2006 was that the bubble hadn’t burst yet. After a bubble bursts, the damage is widely known and it is only natural to then ask what the historical prices looked like. But by then it’s too late. We need to ask these questions before the bubbles burst, hopefully before they get so large the bursting is catastrophic. Think about the what’s happening in the world today. Can you think of a market that may be currently experiencing a bubble? Would you have thought of gold if I hadn’t already included it above? What about health care? (That will be the topic of a future article.) Any other markets?
Thinking of possible markets ahead of time is hard for the general public, though somewhat easier for investment professionals. Getting the data and verifying whether there’s a bubble is even harder. But these sort of things are trivial for computers. Wouldn’t it be nice if someone somewhere set up a system that defines a possible bubble by certain price behaviors and then offers a service where computers monitor all of these indexes attempting to pattern match on just those bubble behaviors. Perhaps one day someone will create such an app and the damage bubbles cause could be mitigated.
NASDAQ data came from finance.yahoo.com
Gold prices came from the London Bullion Market Association
Crude Oil prices from 1974 to March 2008 came from one Department of Energy Site
Crude Oil prices from Apr 2008 to today came from another Department of Energy Site
Inflation adjustment data came from the Bureau of Labor Statistics