« Consumption vs. income inequality, revisited | Main | Know Thyself »
Was there a Housing Bubble?
The conventional wisdom is that there was a housing bubble which has now popped. The data, however, tell a different story. Remember, that the evidence for the bubble was that real house prices had increased tremendously since around 1997 leading to prices that were far above any seen in the past one hundred years. Here's Robert Shiller's famous chart.
The clear implication of the chart is that normal prices are around an index value of 110, the value that reigned for nearly fifty years (circa 1950-1997). So if the massive run-up in house prices since 1997 was a bubble and if the bubble has now been popped we should see a massive drop in prices.
But what has actually happened? House prices have certainly stopped increasing and they have dropped but they have not dropped to anywhere near the historic average (see chart in the extension). Since the peak in the second quarter of 2006 prices have dropped by about 5% at the national level (third quarter 2007). Prices have fallen more in the hottest markets but the run-up was much larger in those markets as well.
Prices will probably drop some more but personally I don't expect to ever again see index values around 110. Do you? If we don't see the massive drop back to "normal" levels then the run up in prices should be described as a shift to a new equilibrium - much as happened during World War II - see the chart. (It's an important question to ask what changed and why?). In the shift to the new equilibrium there was some mild overshooting, especially due to the subprime over expansion, but fundamentally there was no housing bubble.
Here's a nice picture from The Mess that Greenspan Made.
Posted by Alex Tabarrok on February 13, 2008 at 07:45 AM in Economics | Permalink
Comments
Save this post, so that you can look back on it in later years. It's a bold prediction to make right now...hope you're right!
Posted by: Kevin Miller at Feb 13, 2008 8:14:47 AM
I think the tough thing about the housing market is that it moves much more slowly than other markets (such as stocks), so it takes longer to see. If the tech stocks were soap bubbles that could pop very quickly, housing could be more like a lava bubble that takes more time. That isn't to say that housing hasn't reached a new equilibrium, only that with the size of the investment for most people, I don't think you could expect to see a sudden drop as you do in other markets. People don't sell their house only because they don't want to lose any more money in their investment, unless they're speculators. Most people get value from living in the house they chose. The long term question, I think, is whether or not people who don't currently own houses can afford them, and if there is enough incentive for home builders to lower prices of new buildings. I believe your equilibrium theory holds if new homes are priced at current levels, and the market continues to buy them.
Posted by: Aaron at Feb 13, 2008 8:20:59 AM
I don't think I know enough economics to have an opinion on this, but I have a couple questions for those that do:
1. many people seem to expect above-trend inflation over the next 5-10 years; how would this change the visual effect here? I assume it would compress the whole graph vertically, but would it have a disproportionate effect on later/earlier periods? For example, I'm trying to picture (and can't) what this graph would have looked like in 1970, shown in 1970 dollars instead of today's dollars. Would that postwar spike have seemed even more prominent?
2. i think i have a general understanding of hedonic effects, and some differences between a "standard" house today and at the beginning of the graph are obvious -- but it seems like the average American house has gotten nicer (better appointed but especially bigger) much faster than the standard of living has risen over the last ten years or so. If this is true and it's part of what was going on in the "bubble," what does it say about that graph? For example, should we be drawing it on a per-square-foot basis like we would for commercial real estate?
Posted by: Peter at Feb 13, 2008 8:26:20 AM
I don't think I know enough economics to have an opinion on this, but I have a couple questions for those that do:
1. many people seem to expect above-trend inflation over the next 5-10 years; how would this change the visual effect here? I assume it would compress the whole graph vertically, but would it have a disproportionate effect on later/earlier periods? For example, I'm trying to picture (and can't) what this graph would have looked like in 1970, shown in 1970 dollars instead of today's dollars. Would that postwar spike have seemed even more prominent?
2. i think i have a general understanding of hedonic effects, and some differences between a "standard" house today and at the beginning of the graph are obvious -- but it seems like the average American house has gotten nicer (better appointed but especially bigger) much faster than the standard of living has risen over the last ten years or so. If this is true and it's part of what was going on in the "bubble," what does it say about that graph? For example, should we be drawing it on a per-square-foot basis like we would for commercial real estate?
Posted by: Peter at Feb 13, 2008 8:27:35 AM
Was the WWII up-swing a shift to a new equilibrium, or a shift back to the pre-WWI equilibrium?
Take a look at the last 10 years of the NASDAQ as a reminder of how far things can fall.
Posted by: Winslow Theramin at Feb 13, 2008 8:31:22 AM
This seems an odd time to make this argument. If it drops back to 130, was it a bubble? I'd say so. In my judgement, real housing prices could drop 30% from the peak, which would be about 130. Just looking at the top graph, it looks like that last peak to trough took about 7 years, so I wouldn't be bringing this up until maybe 2014.
Posted by: matt wilbert at Feb 13, 2008 8:32:29 AM
"Prices will probably drop some more" way to step out on a limb here.
I fail to see a connection between any evidence you’ve presented and your claim ‘there was no housing bubble’. So prices will ‘probably drop’ while there absolutely was ‘no housing bubble’.
Correction from a bubble this severe is certain to be measured in years. What sort of further price depreciation are you forecasting that you can make such a claim and what are your assumptions? All I see is a graph showing history of a correction well underway.
Of course, what do I know? I’m just a Credit Snob who thought that tighter regulation was the answer.
Posted by: Mark at Feb 13, 2008 8:54:36 AM
I have to agree with some of the other commenters. Whether or not housing has reached a new equilibrium, it is a stretch to say that equilibrium will be high enough to credibly claim there was no bubble.
Is it also your position that lending to people at 13 or 14 times income was responsible behaviour too? I don't mean to be pointed but this really does put you out there. I suppose someone needs to be the anti-Roubini - though his predictions are looking pretty good lately.
Posted by: Finnsense at Feb 13, 2008 9:03:02 AM
It has all the signs of a bubble. There is significant inventory overhang, including new homes, resales, and just coming on line, repos or bank REO. Numbers of sales are declining. People in non bubble areas are keeping homes off the market unless they are forced to sell, creating a hidden inventory. Ratios of prices to income are at record high levels. It is significantly cheaper to rent to buy in bubble markets. Yes it's now conventional wisdom, but if you want to question it, some serious data would help. There is all the evidence you could want plus more at the blog calculated risk.
We won't be out of this until residential housing no longer seems like an 'investment' -- let alone a good investment. I would say right now, most people feel like houses will be nominally more expensive in 5 years then today. Until people think of a mortgage as an anchor and houses as consumption, we won't be out of it.
Posted by: Ziggurat at Feb 13, 2008 9:16:09 AM
Isn't it too early to say what's going on here? The index shows a decline - at least, the one that Schiller showed at the AEA seminar in New Orleans a month ago did. In that series, which had new monthly data included that I think was an update of the chart from his second edition book, there was a fairly significant downward movement in the price index.
Posted by: jason voorhees at Feb 13, 2008 9:49:18 AM
I think that country and local government slow growth policies and other regulations play a role in keeping supply down which can cause a price spike if interest rates fall.
Posted by: Floccina at Feb 13, 2008 9:50:02 AM
You're not really presenting an argument here, other than "personally I don't expect to ever again see index values around 110", which is simply your own intuition. Other than that, you simply point out that the bubble hasn't fully popped yet, but in housing it's normal to take years for prices to adjust.
Posted by: at Feb 13, 2008 10:00:03 AM
I think the run up was caused by a combination of low interest rates and tax law changes enabling people to move without worrying about capital gains.
I also agree with the commenter that the lack of liquidity in the housing market causes it to take a long time to adjust to new prices. This would process will quicken if we get higher interest rates.
Posted by: Jeff at Feb 13, 2008 10:06:50 AM
I think the biggest problem with Shiller's chart is that it's not adjusted for interest rates. Make that adjustment and the recent run-up, while still apparent, isn't nearly as sever.
But I too was hoping to see some argument from Alex as to why there was no bubble. Instead, all I got was a flat assertion, with no supporting argument. Help us out here?
Posted by: David in VA at Feb 13, 2008 10:10:52 AM
Alex, how much do you stand to make on the bets you placed on the prediction markets that housing prices won't go down much more, since you seem pretty confident of that?
Posted by: Cliff at Feb 13, 2008 10:22:27 AM
The second graph says that prices have increased in cities which are more attractive for tourists or retirees (Miami is on top. Detroit at the bottom).
This makes me think that part of the story might be that housing prices reacted to a large shift in demand (inflated by loose monetary policy), while supply couldn't increase as fast.
Also, low interest rates and massive investment in US government bonds from foreign governments may have changed the asset allocation of many investors: real estate is often considered as an asset riskier than t-bonds but less than stocks. If t-bonds pay too low yields, and stock markets go down, it is not absurd to imagine that a lot of investments in 2001-2002 went into real estates. And houses do not pop-up in a couple of days. Again, this would lead to a shift in demand (obviously, more attractive cities would attract more capital, and this would explain the situation depicted in the second graph), while supply could not adapt immediately to the new situation (Rome wasn't built in a day... nor Miami beach).
What next? Probably we now have an increase in supply, and a partial decrease in demand. Prices might stay up or go well down, but the entire story might be that of a perfectly rational situation, not necessarily the boom and bust of a housing bubble.
Posted by: Roberto at Feb 13, 2008 10:25:00 AM
That new homes are getting larger may be true in some parts of the country, but definitely not here in New York City, and presumably in similar congested urban areas. You are definitely getting less square foot per dollar for new homes in New York.
Posted by: bronxilla at Feb 13, 2008 10:27:01 AM
So let me get this straight, everytime an asset price drops in value over a year or so, there has been a bubble? Geez, there must be a few hundred bubbles going at any given time.
Posted by: cb at Feb 13, 2008 10:27:58 AM
Shiller did a good job catching the housing bubble before it popped, and explained it at length in the early editions of his book "Irrational Exuberance."
I think the main problem with your chart though is that certain things become more meaningful when adjusted for "inflation" and some less so. We know those "constant dollars" (a) have less meaning when comparing a horse-and-buggy economy to a modern one, and (b) are recursive (when inflation is housing and housing is inflation)
A better bet would be for you to chart median home prices against median income. That's an easy way to separate the inflation issue (it should apply equally to wages and housing).
So how have median homes fared against median income in the last 100 years?
(Not the "mea culpa on debt" I was looking for :-)
Posted by: odograph at Feb 13, 2008 10:31:13 AM
P.S. - If you do the income/house-price chart it might be interesting to see percentage home ownership on the same timescale. It was generally considered a good thing when the post-war boom (and low prices?) led to an ownership boom.
P.P.S. - How many libertarian and free marketers are with me on eliminating the federal deduction on house interest?
Posted by: odograph at Feb 13, 2008 10:35:59 AM
In order to maintain those higher values, it would mean a sectoral shift in demand, whereas a bubble would imply a temporary increase in quantity demanded.
The best I can reason is that:
1. The dropping of mortgage interest rates led to permanent decreased cost of ownership, shifting a great number of those who were on the bubble of renting to home buying. While a small number of those first-time buyers used ARM's and over-purchased, many others took advantage of low fixed-rate mortgages. This would shift demand permanently instead of just moving along the curve, raising prices.
2. The above phenomenon also created a bubble in addition to the long-term change, allowing people who already owned homes to relocate cheaply. Some buyers may have even moved more than once, driving up prices in the short term. Since the market has cooled, they will not settle, so this should lower prices.
3. Interest free mortgages and increasing prices led to increased investment in the housing market, which would only have a short-term increase in prices. As the market has cooled, investment has shifted away from the housing market, which should also lower prices.
But that last chart pretty clearly indicates that there WAS a bubble. The data ends in 2006, and in 2007 we see a decrease in prices. While these prices are slow to fall, I believe they will slowly drop until they catch back up with the inflation rate or another mortgage phenomenon takes place (mainly because housing is not just an investment but also a useful and necessary good).
Posted by: J at Feb 13, 2008 10:38:08 AM
Odograph said:
P.P.S. - How many libertarian and free marketers are with me on eliminating the federal deduction on house interest?
---------
I wish someone could've nipped this in the bud before it became a politically untouchable topic.
What we should be doing is making college-tuition & education expenses fully deductible. There's a lot more evidence of the positive effects of education for a society than home ownership.
Posted by: Finance Monk at Feb 13, 2008 10:40:49 AM
I agree with Floccina. The metro areas with the largest run-ups are the ones where local governments restricted supply. Randall O'Toole's Thoreau Institute blog explained three years ago how smart growth was fueling housing bubbles.
Smart Growth and Housing Bubbles
Dallas, Houston, and Atlanta - three relatively unregulated metro areas - have not experienced the housing bubble, despite growing the most in population during this new century.
Here's another passage from a more recent Randall O'Toole essay:
"While housing prices grew by more than 130 percent in California and Florida from 2000 to 2006, prices in Texas grew by only 30 percent. With few exceptions, the states that saw the biggest bubbles were ones that had passed growth-management planning laws. And with one exception, every state that has passed such a law also saw a housing bubble."
Posted by: John Dewey at Feb 13, 2008 10:41:12 AM
You know John Dewey, what they call the "zoning zone" often corresponds to areas hemmed in by geography. Here in OC there is zoning (though it varies greatly by city). The bigger factor is that flatlands "this side of the mountains" are mostly built, and the commute-paths through the passes are bottlenecked.
You can blame it on the zoning, or the course of history that makes your other choice a 2-hour commute.
I assume that the cities with low prices have open geography and relatively fast commutes down-town?
Posted by: odograph at Feb 13, 2008 10:47:27 AM
Don't forget to adjust the lower graph for inflation, those drops from mid 06-07 would probably be 10-15 index points after taking 4% more for inflation (that would bring real prices back down pretty quickly).
I'd argue that there is some real adjustment in home prices necessary post 1997, to account for the capital gains exemption in home prices, but I don't think it's 100%.
I do expect that we'll see values in the low 100s (110-130) with a likely overshoot for a few years, too. But I'd guess that inflation will be more of a factor than nominal declines.
Posted by: nelsonal at Feb 13, 2008 10:56:43 AM
