DeLong Smackdown Watch: Paul Krugman

Paul Krugman writes:

Permanently high home prices? - Paul Krugman - Op-Ed Columnist - New York Times Blog: Brad DeLong is of the belief that home prices won%u2019t fall back to pre-bubble levels, because "America is filling up" and "we will wind up with higher prices for scarce positional goods--chief among which is location, location, location."

The trouble with this argument is that it's an argument for rising rents as well as rising prices -- and if you believe the BLS data, that just hasn%u2019t happened nationally. Below are the Case-Shiller home price index and the BLS index of "rent of primary residence", both adjusted for overall CPI and expressed as indexes with Jan. 1987=100. Bottom line: rents have hardly risen at all in real terms.

Now, maybe the BLS is wrong. But for what it's worth, the data say that essentially all the rise in real home prices came from a rise in the price-rent ratio, which suggests that things will go right back to where they were.

The "rents" argument applies to the "filling up" part of the argument, but not to the "foreigners want to invest in America and that will keep interest rates low for a long time" argument.

Abd I am less confident in the BLS than Paul is. Don't get me wrong--they do a superb job on a shoestring--and maybe I spend much too much of my time in unusual and unrepresentative places, but I am almost as unhappy with the rental price index as I am with the software price index.

Comments

[jerry]

My one data point is that for the first time ever, ever, my rent did not increase this year. Hasn't gone down. But it didn't increase.

But I really wanted to point you to this: John Conyers, citing Yoo's interview for Esquire, is calling Yoo once again before the judiciary committee.

http://www.guardian.co.uk/world/2008/apr/08/usa?gusrc=rss&feed=networkfront
Former Bush administration lawyer John Yoo asked to testify before Congress

"Overall you have made such extensive public comments on these and related matters, it is difficult to understand why you would continue to decline to present your views to the committee," Conyers wrote to Yoo.

[jerry]

I do have another data point. Duncan Black is buying a house! (Congrats to Duncan Black.)

[matt wilbert]

I think the "foreigners want to invest in the US argument is dubious" too. Foreigners may want to invest in the United States, but I doubt they are going to want to keep investing in fixed-income assets at low rates--it hasn't worked out too well for them so far. Maybe they will, but I wouldn't want to bet on it.

[Gavin]

I agree with "matt wilbert". The "foreigners want to invest in the US argument is dubious". Foreigners were able to invest in the U.S. for decades. Why should the current round of investments lead to a booming housing market unlike any other in 50 years?

Also if Brad doesn't like the BLS rent index you can look at the data at HUD

http://www.huduser.org/datasets/fmr.html

[Rich C]

I don't see why the stability of rental rates doesn't address the foreigners-buying-our-condos argument: if non-US purchasers were buying up properties, wouldn't that reduce the supply of rental units, hence driving up prices?

Also, on the software price index, isn't this just a reflection of the bet Bill Gates made back in the early 1980's, that intellectual property rights would keep software prices stable (via monopoly), while hardware prices would fall precipitously? Is there a reason to think that software prices are falling, or that the quality of software is rising so quickly, that this is not an accurate picture?

[RedCharlie]

Does the BEA index consider Open Source? I imagine not, as would have to measure how many people are downloading and using Ubuntu/Fedora/Knoppix/Suse/etc. I would be surprised if the BEA even attempted to measure this. But thanx to Linux, I no longer pay the M$ tax, so my personal software cost has gone down considerably.

I use RHEL at work as well, but we pay for that. The lion's share, however, goes to Oracle. If any two companies are responsible for the software cost remaining resilient, it's M$ and Oracle.

[RedCharlie]

Does the BEA index consider Open Source? I imagine not, as would have to measure how many people are downloading and using Ubuntu/Fedora/Knoppix/Suse/etc. I would be surprised if the BEA even attempted to measure this. But thanx to Linux, I no longer pay the M$ tax, so my personal software cost has gone down considerably.

I use RHEL at work as well, but we pay for that. The lion's share, however, goes to Oracle. If any two companies are responsible for the software cost remaining resilient, it's M$ and Oracle.

[RedCharlie]

Does the BEA index consider Open Source? I imagine not, as would have to measure how many people are downloading and using Ubuntu/Fedora/Knoppix/Suse/etc. I would be surprised if the BEA even attempted to measure this. But thanx to Linux, I no longer pay the M$ tax, so my personal software cost has gone down considerably.

I use RHEL at work as well, but we pay for that. The lion's share, however, goes to Oracle. If any two companies are responsible for the software cost remaining resilient, it's M$ and Oracle.

[Lord]

As far as national rents, no I don't see them increasing much, but you were looking at LA and the 20 composite. LA rents are rising over 5% and have traditionally increased at inflation + 2%.

[RedCharlie]

Just FYI, I hit the "post" key once.
I did hit my back arrow once after that, but still, that doesn't explain my post above appearing three times!

[N]

How do you define a bubble? A drop from 200 to 100 is not the same thing as a drop from 200 to 50 (the nasdaq was more like a drop from 200 to 50 vs. a drop from 200 to 100). Also, at some points rents may increase.

You may need to look at this by region of country.

[Jonathan King]

I think Krugman has already done most of the smackdown, but I think it is worth noting the "running out of room" argument also has big problems explaining why prices were going up over 15% per year in *so* many markets, including some (like Las Vegas and Phoenix) where the marginal cost of building new units was still quite low. In other words, I don't think anybody doubts that (say) the better DC suburbs or Manhattan or SF are always going to be more expensive than most other places, but there was no particular reason for prices in these places to increase that quickly when incomes weren't and population increases were also fairly modest. In other words, when it's obvious that a bubble mentality is taking over in many, many places, I think it's dubious to argue along the lines that "but city X is different" without the demographic data to back this up.

[jerry]

Well Phoenix (and Las Vegas?) prices increased presumably as Californians fled from San Diego, LA, and San Francisco prices. Phoenix is a quickly growing "city" and that population came from somewhere. (For some very broad definitions of city, not necessarily my own.)

Phoenix prices used to be much much lower than California prices, and still are, but not nearly as much. When the Californians came because they couldn't pay San Francisco prices, they could still pay much higher than the typical Phoenix price and so bid these wastelands up.

I say nuke the place from orbit and start over in a millenium.

I am hopeful that prices will drop as Krugman predicts. I wonder though, have stock P/E ratios gone back to the historical trend?

[joe]

IMHO Rents can be artificially low for the short term due to the real estate bust. They are in my Bay Area town. I see lease to own signs and for rent on properties that were once for sale.

Home owners are forced to rent when they can't sell a property and they're forced to rent at a loss if they have "investment" properties. They also take in renters to try to cover the mortgage.

There's unusually high competition for renters until these "rental" properties" are foreclosed or the owner realizes 10% annual appreciation isn't going to happen and they stop the hemorrhaging.

[dissent]

San Francisco rents are soaring.

[Randolph Fritz]

Software development is the slowest design practice; unlike computer hardware it hasn't gotten any faster. This is why there is such hope for AI, and why not achieving it means a different future.

[Ken Houghton]

I'm talking my position as a potential seller in a market where (unlike, say, most of the Midwest) real estate is often an investment as well, but there is a problem with the Krugman argument--one that is clear from his graphic, actually.

Which is that--unless you lived on the east end of Lon Gisland, and bought in 1989--you've seen two equilibrium points in that time period: one ca. 1988-1991 and the second ca. 1996-1998.

I can imagine a return to 1997ish levels, or even a slight increase from them (chained GDP is up 32.1% from 1997q4 to 2007q4), but a return to 1987 levels plus that increase seems a more major step.

And even that would be a 50%+ decline from current levels here in the NYC area. Which is at least a 60-70% decline from the peak.

I've been pessimistic--I called Krugman's call for a 30% decline optimistic at the time--but if he really expects a return to ca. 1987 levels--well, Menzie Chinn's model appears hopeful by comparison.

Unless you're blending bankrupting both coasts (Bos-Wash corridor to at least Richmond and probably Raleigh; California, Oregon and the other Washington) and losing almost all of what remains of the 4% p.a. increase (unadjusted for inflation; probably less than 1% real) that the rural Midwest [the rusted-out belt, as it were] has seen and then coming up with a mean 50% decline with a rather impressive standard deviation, I don't see how 1987--or even 1997--fits with Krugman's previous predictions.

Which, unfortunately, is not to say that he's wrong. But I suspect, sadly, that there will be more people willing to rent than buy for quite a while.

[a]

"foreigners want to invest in America and that will keep interest rates low for a long time" I guess we're looking at different economies. New loans - to businesses, municipalities, and individuals - are currently so hard to come by that the U.S. government is having to step in. (Look at municipalities, student loans, businesses. Where are all those foreigners willing to invest in America?) So I guess you're making a prediction about the future. Well, only time will tell.

[James Wimberley]

Brad: the citation ends after the graph, but you format your response within the indent.

[baileyman]

Delong argues for higher land prices, that's all. That foreigners will keep demand high is to say that OUR foreigners will want to own more than one place here in increasing numbers. Perhaps US citizens who are THEIR foreigners will do the same there? Isn't this just another way of saying that rising inequality around the globe has unusual effects, one of which is to drive up the price of land in global cities because the top end wants multiple places around the globe?

[Oskar Shapley]

Brad is unhappy with the software price index?! He should be unhappy with the hardware price index and all the terrible hedonic corrections in the CPI.

Top of the line computers cost $1000, that's a marketing law, so the index should be constant. Apparently the BLS is calculating the price of CPU POWER, not computers.

Using CPU power as measured output IS A BAD CHOICE. Computers are not twice as good because they are twice as fast. Doing such a chained index is naive.

What they should be doing is to take logs of computer power, which would be closer to the real usability progress that users receive.

And software production is a design process, not a manufacturing process. It can not be automated and is subject to Baumol's cost disease.

[John Howard Brown]

I think there are two keys to the behavior of the software index. First substantial monopoly power exists in operating systems and some forms of application software. Second, unlike computer hardware, productivity in software coding is improving little if any. Instead, as programs add "bells and whistles" the among of coding and testing goes up almost exponentially. It is true that some unmeasured "quality" improvements are a part of the newer versions of software. However, most of the added features in, say, Word are completely irrelevant to my usual word processing activities. Any benefits provided are not worth the diversion from getting my work done to learn.

[Joe]

Despite the current credit market trainwreck, I think there is a good argument that the U.S. mortgage industry that comes out of the recession will be able to provide housing loans that, all other things equal, are slightly cheaper than were available in 1980, 1990 or even 2000. More economies of scale, IT, global access to funds, a widely-held belief that inflation will remain low and other things are responsible. My point is that cheaper access to capital could allow people to spend a bit more on housing and increase the average cost of housing over what it was pre-bubble.

[save_the_rustbelt]

Move to Detroit, you can buy your own city block, real cheap.

Does Brad understand that the entire world is not like California?

[MarvyT]

Rents are more sensitive to supply and demand economics than home prices. When I owned a condo in San Diego there was no relationship between rents and the market value of those properties. Long term psychology seems to determine property values. If owners feel that they will get a good return in 5 to 10 years, they will rent a property at a loss. Renters can move every year so rents have to remain short-term competitive. In real estate markets where property price appreciation is historically slow (small towns in the mid-west), property values tend to track rents more closely. Property values will only return to the rent index if the long term psychology of property ownership has changed.

[João Carlos]

Professor,

With relation to computers X software prices:
1- hardware follow the Moore's Law;
2- software is human work, line by line;
3- there is a monopoly at software, to be more specific the operational system, Micro$oft.

Any economist worth its salt will say that under these conditions hardware prices will fall while software prices will stay constant.

So, what is the problem with the BEA price indeces for information thecnology? Or better, you have any real reason to be less confident in the BLS?

If the problem is that the data show your hypothesis is wrong, I am sorry, real Science will drop the hypotheses and not the data.

[Measure for Measure]

I call this one for Krugman.

1. Rental data from other datasets gives the same story.
CBO analysis:
http://krugman.blogs.nytimes.com/2007/12/28/housing-how-far-is-down/
(which admittedly *might* use the BLS rental data, though they cite the Office of Federal Housing Enterprise Oversight, the Dept of Commerce and the BEA)
and
Fortune/Economy.com / Property& Portfolio Research:
chart: http://money.cnn.com/2007/11/06/real_estate/home_prices.fortune/index.htm
methodology:
http://money.cnn.com/magazines/fortune/price_rent_ratios/
(which still *might* use the BLS data, though I suspect not).

2. Foreign buyers might lower the spread between mortgage rates and tbills, but I doubt whether that would sustain half of the increase in housing prices over the past few years. They may also buy lots of non-residential real estate, which could have indirect effects. But I doubt whether foreign buyers are engaged in a price-distorting degree of absentee landlordism.

These econ-blog debates are terrific, btw.

[n]

I heard a famous investor (Rogers - not sure) say something like stuff trades at 30% of its high in a bear market.

Sometimes stuff stays down for years (look at a stock such as Intel for an example). Or consider how long it took gold prices to come back from all-time highs.

It will be interesting to see how real estate works vs. stocks or other markets. Real estate has key differences, so maybe it will not drop as much as other assets (and will not stay down as long).

[John Drake]

Having looked for either a place to buy or to rent in the Bay Area last June, and I found rents were a bargain compared to purchasing. I'm renting a house for less than half of what it would cost to buy, and the owner has to assume market risk, and seismic risk. SF could be a different story because I didn't look there. As I had great flexibility, I looked in the North Bay, East Bay, and Peninsula so I got a good idea of what the rental and RE market is like for a good spread in the Bay Area. Go to Craigs List and take a look at what places rent for. For example in Novato I found you could rent a house identical to one listed at over $900,000 for $2,500. A no brainer. That house is over priced by a factor of two.

As for the notion of "filling up," look at Japan. That is pretty crowded place and RE market there has not recovered from it's 1989 peak. If Japan isn't "filled up" I don't know what is.

[Maynard Handley]

"Software development is the slowest design practice; unlike computer hardware it hasn't gotten any faster."

This is simply not true.
Even within the commercial software realm, there are substantially better tools today than there were ten years ago. If it takes just as long to ship each successive rev of a piece of software, that's often because each successive rev is more ambitious. Fifteen years ago, say, to ship an MPEG decoder for the Mac was to write a real piece of software requiring at least a few man-months; today this is a task that would take a day, because you can put the UI together using visual tools, while all the hard work of decoding the MPEG is delegated to calls in the OS. It's shifting the goalposts to not realize this profound difference.

There are, however, two extra issues.
On the plus side, if you are writing open source software, you can often get up to speed astonishingly quickly because you aren't constantly re-inventing the wheel. You get some of this as OS's offer an ever richer selection of APIs, but there are advantages to having access to the source in that you can create a better match between what is available and what you need --- for example you may not actually need JPEG decode, but your project requires an IDCT, and by looking at the JPEG code you can figure out very rapidly how to write an IDCT that matches your task.

On the negative side, there is the question of how far backwards and how widely you want to go in terms of compatibility. If you insist on being able to read even poorly formed files from twenty years ago, if you insist that your software run of the version of the OS three generations ago, if you insist that your software work even of craptacular hardware that doesn't actually conform to the more difficult parts of a particular specification, well of course your job is that much harder. But the choice to operate that way is yours, it is not a fact of nature or of economics. It is a business choice.
Microsoft has choice bend-over-backwards compatibility and, while it served them well for many years, it now appears to be a millstone around their necks.
Apple has frequently chosen to ditch the past (in the sense that compatibility with the past has been limited to special environments which are only kept around for a few years), and of course makes no claim or attempt to run on any random piece of PC hardware. This model has had people complaining in the past, but overall it seems to have done reasonably well for them.

[Noel Maurer]

Regarding the CPI, rents, and the Bay Area:

The /exact same apartment/ in Mountain View that I rented for $1,200 in 1992 currently rents for $1,667 in 1992 dollars. (That's $1,724 and $2,395 in 2007 money.)

The /exact same apartment/ in the Mission District (with parking!) that I rented for $995 in 1997 currently rents for $1,473 in 1997 dollars. (That's $1,250 and $1,850 respectively in current dollars.)

So Brad's observation about real rents is certainly accurate for the central part of the Bay Area.

Of course, there were still empty lots in that part of M.V. in 1992, and I still had to make friends with the local crackhead in 1997. Now, both areas are somewhat nicer places to live. Real rents in central Boston, unlike the Bay Area, seem to be about what they were 15 years ago.

So I draw no conclusions.

[Kelly]

I'd like to submit this partial quote from the local paper business section;
--------------------------------------------
Star-Telegram Staff Writer (Fort Worth, TX)
Existing-home sales in North Texas were down 25 percent in March compared with a year ago -- the largest percentage drop in 14 straight months of declines, according to preliminary figures from the North Texas Real Estate Information System.

But the median price of a home in March was up 2 percent from a year ago, to $149,790, according to NTREIS.

Home builders are seeing a similar effect on new home sales, which are down 31 percent in the first quarter, according to a report released by Metrostudy last week.
-----------------------------------------------
Now my question is this; Are Krugman and DeLong debating about the average price of homes or are they debating whether the housing market is near it's bottom.
I've noticed that while the home sales business is CLEARLY in serious trouble, the price of homes has not gone down much to speak of, if you use these "average" numbers.
If you look at the selling price of a given size home in a given neighborhood however, what I see anecdotally is a pretty significant drop.
I guess the point would be that I'd have to know where the sales price data they are both using is coming from, like the rent price data that is being questioned.

[Joel]

Rents aren't going up because, unlike banks, owners have to live with their deadbeats. Owners are a lot more careful about with whom they do business because owner's mistakes aren't just pieces of bad paper.