Chart of the day, housing bubble edition

By Felix Salmon
September 19, 2012
new paper by Karl Case and Robert Shiller, looking at the results of a survey they've been handing out to homebuyers annually since 2003.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">


This chart comes from a new paper by Karl Case and Robert Shiller, looking at the results of a survey they’ve been handing out to homebuyers annually since 2003. The idea is a very smart one: if you want to get an idea of the behavioral economics of homebuyers, the best way to understand what they’re thinking is to simply ask them.

And this chart, in particular, is both very elegant and very informative. It’s elegant because you have a very close maturity match: the average duration of a US mortgage, before it’s refinanced or the house is sold, is about 7.5 years, which is close to the ten-year horizon in this question, which Case and Shiller ask every year:

On average over the next 10 years, how much do you expect the value of your property to change each year?

Now the number of homebuyers in America vastly exceeds the number of people who understand the mechanics of compound interest. If you asked instead “how much do you think your home will be worth in ten years”, and then presented that answer as an annualized percentage increase, I suspect that the answers — especially in the peak years of 2004 and 2005 — would be substantially lower. (Put it this way: if you bought a $260,000 home in 2004 and expected its value to rise at 12% a year for 10 years, then by 2014, you’re saying, it would be worth more than $1 million. I suspect the number of people answering 12% or more is going to be greater than the number of people who think the value of their home will quadruple in ten years.)

Still, that’s not particularly important, especially since the question has remained the same for the past decade: the trend here is real. And what’s fascinating is that the big fall in expected long-term home-price appreciation happened before the financial crisis, and that the crisis is actually completely invisible in this chart: expectations continued to deteriorate long after it was over.

And even given the fact that homeowners tend to overestimate annualized percentage returns over 10-year horizons, we’re now at the point at which the expected rise in home values barely exceeds today’s record-low mortgage rates. Over the long term, homebuyers still think it’s a good idea to buy a house. And they might be right about that. But they’re not buying because they think they’ll make a handy profit in ten years’ time.

Which brings me to one of the central themes of the Case-Shiller paper: the idea of a “speculative bubble”. If you look at the situation in the chart circa 2004-5, there was a huge gap between the cost of funds and the long-term expected return. And if people really believed house prices were going to rise that much in future, it made all the sense in the world to lever up, get the biggest mortgage they could find, and buy lots and lots of house. After all, the more levered you are, and the more house you buy, the more money you make.

Case and Shiller have a handy definition of a speculative bubble, in this paper: it’s a bubble with “prices driven up by greed and excessive speculation”. But here’s the thing: people don’t speculate on a ten-year time horizon, and the producers of “Flip This House” weren’t waiting around to see what properties would end up being worth once the kids had gone off to college. A truly speculative bubble, it seems to me, is a function much more of short-term house-price expectations than it is of long-term expectations. If you think you can buy a house today, sell it in a few months’ time, and make tens of thousands of dollars doing so, and if you intend to do precisely that, then you’re clearly part of a speculative bubble. But it turns out that home buyers were actually surprisingly modest in their expectations of one-year price increases — they expected prices to rise less than they ended up rising in reality.

On the other hand, if you buy a house now in the expectation that it’s going to increase in value substantially over the next decade, you might be a buy-and-hold investor, but it’s hard to characterize what you’re doing as speculation.

I’ve been disagreeing with Shiller on the subject of speculative bubbles for five years now, but I think this is important: just because you have a bubble, doesn’t mean you have a speculative bubble. The dot-com bubble was speculative; the rise in house prices in 2000 was not. There was a speculative bubble in Miami condos; there was not a speculative bubble in Manhattan co-ops. If you buy because prices are rising, that might be because you want to flip your property and make money — or it might equally be because you worry that if you don’t buy now, prices are going to run away from you, and you’ll be forced to move out of the neighborhood you love because you can’t afford it any more. It’s still a bubble, but it’s more of a fear bubble than a greed bubble.

Still, bubbles are bad things, and they’re liable to burst either way. And so I take solace in this chart, because it shows me that people are buying, these days, for the right reason — which has nothing to do with expectations of future house prices, and everything to do with simply paying a fair price for the shelter they’re consuming. House prices might not rise much over the next decade. But if they fail to rise, today’s house buyers aren’t going to be disappointed: they will still have lived in their homes while paying a perfectly reasonable sum to do so. Which is a much better state of affairs than bubble-and-bust.


We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see

Good article, Felix, but I would caution on one point:

“the average duration of a US mortgage, before it’s refinanced or the house is sold, is about 7.5 years”

This statistic is oft stated, but it should not be presumed to be a universal constant. When mortgage rates fall 2% every five years, with home values rising, people will regularly refinance. It would be insane for them not to do so. When mortgage rates rise 2% every five years, with home values falling, then refinancing activity will grind to a halt. Few people will have enough home equity to take cash out, and refinancing will mean an increase in borrowing costs.

I have a feeling that we’re about to see this statistic grossly violated, perhaps doubling or more. People are selling less frequently, and once mortgage rates stop falling they will also refinance less frequently.

Posted by TFF | Report as abusive

I don’t think your argument over the definition of “speculate” is really useful. I think of speculation as betting on asset price increase as opposed to income flows.

If people took out excessive mortgages because they “knew” they could sell the house for a profit at any time their ability to pay the monthly interest was impaired, is that speculation? If people pay a lot more each month than equivalent rent, hoping to make it back when they sell, is that? Maybe not, but those actions were bad for the economy, just as the short-term flippers were.

The big distinction for housing, IMHO, is buyers who think of shelter as a disposable, and those who think of it as an investment asset class. Obviously, those Manhattanites who are worried about being priced out are not relying on their asset going up in value, but I think people who bought for strictly defensive purposes were small.

Posted by AngryInCali | Report as abusive

The NY Fed pretty much settle the question of the housing bubble in their September, 2011 Staff Report No. 514 ( f_reports/sr514.html). The mortgage companies allowed “investors” to use Alt-A loans to finance speculative house purchases which were largely concentrated in the “sand states” (CA, AZ, NV, and FL) where the bubble was the worst.

The all of the homeowner psychology/expectations arguments are just speculation and do not explain how the housing bubble was financed and why it was so much worse in the “sand states”.

Posted by brookside | Report as abusive

I would say people are still being incredibly optimistic on future returns. By 2022 I’d be surprised if the average house will have increased in price by more than 15% total.

Posted by Harpstein1 | Report as abusive

An asset would approximately triple in value- not quadruple in 10 years at a 12% yearly appreciation rate. It takes about 12 years to quadruple in value at that rate.

Posted by AdamJ23 | Report as abusive

Felix, you obviously were not paying attention circa 2004-2006 to what actual buyers were actually doing in places like Baltimore. Prices in that city–which is not known as one of the bubble zones–were absolutely driven by real estate investors, who by 1Q 2005 accounted for 2/3 of all sales. 0358

There, as most everywhere, much of that speculation was built upon mortgage fraud.

This is just a fact. Analyze it.

Posted by Eericsonjr | Report as abusive

While a minor point, you might want to check the mechanics of your own compounding interest. A $260,000 house that increases in price 12% annually would be worth just over $800,000 in ten years. It would take compounding at 15% in order to reach the $1 million level you mention.

Posted by IanF | Report as abusive

The idea that there were all these “defensive purchasers” out their is just post-hoc rationalization. Maybe among graduate degree holders, this might have been 10% of the market, but college graduates and below are not making this sophisticated of purchasing decisions in enough numbers to mention.

I think people mentioning it is more a function of the social circles people who write about economics run in than any actual behavior.

Also as mentioned your math was wrong. Remember the divide by 70 shortcut!

15% is a little overly pessimistic with the population growth the US has, but I would be surprised if it was only 2-3% a year.

Posted by QCIC | Report as abusive

QCIC, by some metrics the US housing market remains overpriced by about 20%. Calculate cumulative inflation over the next decade (or wage inflation, if you believe that is the stronger driver of real estate prices) and subtract 20% — Harpstein’s figure seems realistic.

Not that free markets are known for being predictable and orderly…

Posted by TFF | Report as abusive