Brad DeLong's Semi-Daily Journal: A Weblog

OK this argument is basically assuming that rent is constant over time - independent of interest rates. As Brad knows nominal interest rates are a function of the inflation rate and the real interest rate. The natural (?) assumption is to assume that rents follow the inflation rate, and thus rent should not be constant over time. If this is so, then the income stream generated by the house needs to be compared to an increasing rental stream, not a constant one.

Basically, Brad's argument can be adjusted by talking about real rather than nominal interest rates. I'm not sure this is the case, I haven't seen the data from TIPPs (treasury bonds where adjust for inflation) over the past ten or so years. Of course, it could well be the case that over thirty years (before there were TIPPs) the real rate has decreased because inflation expectations have been reduced and there is less risk premium. And then of course, maybe rents don't track the general inflation rate - not sure about this one.

Presumably you also have to account for the fact that the cheaper interest rates are, the more attractive it is for people to borrow money to buy property to rent out, thus increasing prices even more. Or is that somehow included in the original calculation?

A related question: what happens to rents when/if the bubble bursts? Do rents fall too? I assume the obvious: yes. Presumably rents would be competing with a lower cost to buy, right? These things seem obvious to me, but it also *seemed* obvious to me that rents should track purchase prices upwards as well, so better to ask...

There's also the question of price-appreciation expectations. A landlord expecting house/apartment building prices to rise may be willing to accept a lower yield in rental income versus Treasuries, as tradeoff for holding a rising-value asset.
In my very-high-demand-for-housing part of the country, rising prices paradoxically seem to feed on themselves and *increase* demand, as people go further out on the ledge of debt service in order to own a potentially-rising asset, or 'get on board before it is too late'. Of such dynamics are bubbles made.
And rapidly-rising sale prices have paradoxically *decreased* rental rates, because some former renters have leaped into buying houses for the reasons above, and more people are enticed by rising sale prices to invest in rental property, increasing the supply of such units.
So the change in rent/buy ratios in my estimation appears to be driven mostly (not completely) by sale-price-appreciation expectations rather than current yield.

So we have a self-perpetuating dynamic, all driven by rising house sale-prices and the expectation of their continuing long-term. All very reminiscent of Cisco stock circa 1999, Japanese real estate in 1989, gold in 1979, ad nauseum. These situations have a habit of ending badly, when desperate-rising-price-chasers leave the demand pool.

In a number of locales, the rise in rental rates is held in check largely by rent control laws, and the occasional recessionary lack of demand. I used to live in a rent-controlled place where the law said the rent can be raised no more than 10% in 3 years, and that's pretty much what I've got (on an admittedly quite cheap apartment that was still cheap after two such periods). And that's also pretty much consistent with "benign" inflation rates. (I got raises only a little in excess of this during the time.)

Wanted to toss my two cents in on Brad's thesis apt rental thesis. His argument makes perfect sense until you run across situations that make a mockery of it.

My family has been in the real estate biz for over 50 years in the SF bay area. My father actually created the first condo-conversion in SF in '64 (hold the applause.)

We just sold three multi-unit apt bldgs in SF (about 250 units total) after having owned them since the 1960s. We really weren't eager to sell since we obviously had no mortgage and they paid quite decently, in the 7% range after expenses.

Most of the units were small 1-bedrm and studios rented to older folks, immigrants and newcomers on a tight budget. These weren't glamor properties by any means.

SF has strict rent control so we couldn't really raise rents much and couldn't bring them to market prices unless someone voluntarily left.

The last few years saw more vacancies than normal but the buildings were about 85-90% full and we were getting about 70% of what we consider market rent.

We thought of selling back in the dot-com days when money was flooding the area. But no interest. Why buy a funky old apt bldg when you could buy Webvan shares at $150?

Then something funny happened. Jobs left, office vacancies skyrocketed, population declined. But residential (both single-family and multi-unit) property started to move even with crappy economic conditions that still persist.

Earlier this year, for the first time in 2 decades, we got a couple of inquiries about selling our buildings. So we got an appraisal. We were shocked. The buildings had nearly doubled in value in six years, even though the economy was weak, vacancies were up and rent control made it tough to increase revenues.

We decided to sell. We asked for 10% more than the appraisal. We got 25% more. Shock again.

The new owners appear to be pleased. Yet even with low interest rates, they will not be able to make any money with these properties for a number of years at least because of the very high price paid. Our view is that income properties should throw off income. Apparently that notion is no longer in vogue.

To make a long story a bit longer, we know a lot of folks (long-time owners who know the market well) who are selling. We aren't really sure why people are buying, but we'll let them worry about it.

I know lots of big money players are stretching for income right now. That's why you see junk bonds trading for an historically low premium over Treasuries---anything to get an extra 1-2% of income. Don't think that will end well, but I'm no expert.

But I do know income property. And when buyers are stepping up and paying prices that guarantee a multi-year negative income stream (at minimum), I don't know what to think. This hasn't happened before.

I assume buyers like this think that asset inflation will make them whole and they will be able to make up the lost income that way. Maybe they will. Anything's possible. But for me, it's much safer to buy income property that throws off (get this) income. Then any appreciation is just gravy. Better not to have to rely on asset inflation to make money.

Brad's example certainly is lucid. But here in SF, buyers are paying prices far beyond what any reduction in rates would warrant. They are, in short, buying income property that will provide them no income.

And on the residential side, the math is even more extreme.

It's been a while since I studied economics, but if I understand Dr. DeLong correctly, this sounds suspiciously like the primary axiom of economics: buy low, sell high.

If interest rates are low, borrow. If they're high, lend.

The buying of "income" property that generates no incone is a sure-fire speculative buble. I rent out a house in the UK and I can tell you for nothing that even with low-ish interest rates (5% in the UK now) the rental income wouldn't come close to covering the cost of borrowing the money to buy the house (house price 200,000 market price rent 7500/year) i.e. buy-to-rent is a non starter if you are not simply expectinmg speculative price increases. In fact, now that everyone in the UK is waiting for the house prices to crash, no one is buying anything anymore, those people who are renting and were planning to buy, are staying put, and, low and behold, rents are starting to rise.
It would appear thus that the US market is about 6-12 months behind the UK market, so WATCH OUT!!!

Steve

sfjoe:

We'll see who's laughing when Arnold TERMINATES rent control!

Two points:

(1) Jim Follain and Dixie Blakely have a paper that shows that real rents for a constant quality apartment in the United States have been basically constant dating back to WWII.

This makes sense: rent is driven largely by locational, rather than financial, considerations.

(2) Piet Eicholz has a paper that shows that real property values in Amsterdam have cycled around a pretty much constant level for 200 years.

>>Richard Green wrote:

(1) Jim Follain and Dixie Blakely have a paper that shows that real rents for a constant quality apartment in the United States have been basically constant dating back to WWII.

(2) Piet Eicholz has a paper that shows that real property values in Amsterdam have cycled around a pretty much constant level for 200 years.<<

Since incomes and GDP have historically increased faster than inflation, does the above mean that the proportion of income spent on housing has been decreasing over the years?

Not in the States, judging by the paper Angry Bear linked to a while back.

In the good old days, the rule of thumb was to spend 25% of your income on housing. The paper (and the indication from my finance company seven years ago, when we were mortgaging our future) indicates that the current range is 30-33%.

That this leaves less margin for error, and makes buying more risky (to the buyer and the mortgage holder), appears not to be being priced into the market.

The problem is the "constant quality" business. While in most places the price of a house that has only had basic upkeep (no additions, improved kitchens, etc) has fallen relative to income, the fact is we keep building and rebuilding better quality, or at least larger, houses. The average house built now is about 50 percent bigger than it was 30 years ago. Thus if we use size as a proxy, the typical house will be in real terms 50 percent more expensive than the typical house of 30 years ago, even if "quality adjusted" prices have remained constant.

The notion that "speculation" is a special property of certain kinds of investment was Yglesias starting point. He misunderstands a good deal about the math of investment, as Brad has demonstrated, but also about the nature of investment, I think. Other than short-dated TIPS, there is an element of risk in just about any investment one can think of. Assumptions about the direction of asset prices and family income are implicit, for instance, in decisions about using a mortgage to purchase a home. If the value of the home falls, the homeowner may have to forego a lucrative job offer in another city. That is true at any time, not just now that prices have risen a lot. The homeowner has speculated on the value of the home. "Speculation" is not an absolute, but a matter of degree. Throw in the right investment math, and the degree of speculation involved in rental property at today's prices seems unremarkable.

So I ask again (the new barriers to posting here apparently didn't like my prior effort), why do we care what a recently minted philosophy Phd who misunderstands the economic concepts at issue has to say about the real estate market? Yglesias gets lots of things right. Can't we look the other way when he says silly stuff about real estate valuations?

Interest rates have a huge affect on the occupancy rate needed to make money. A midwestern builder borrows $1 million. At 8%, that is $80,000 per year. At a flyover country rate of $4000/y, 20 of the apartments built must be occupied all year to make money. At 5%, or $50,000 per year, only 13 apartments need to be occupied. We have seen a huge increase in numbers of apartments and a large increase in occupancy rate. It makes sense to throw up an apartment building at a locked in lower rate and wait for the population to grow into the vacancies. As vacancy rate affects rates, interest rates drops also mean that Matt should be getting a break on his rent even as the landlady is paying more for the building. A more likely scenario is the landlady pockets the extra money.

About thirty years ago, back in my hometown in upstate NY, my father had an opportunity to buy a very well located and basically sound home needing some repair at a very attractive price. But having come of age during the Great Depression, he was convinced that the rampant inflation of that day was unsustainable, that even that low price was too high, and that in the long run it's cheaper to rent. Today, you can't even buy the cheapest imported Korean car for what he would have paid.

Back in the late '90s I was thinking of buying a bigger home, but believed Greenspan really meant it when he said he intended to reduce inflation to a level at which it would not be a factor in any economic decision. (Remember that?) So I didn't. Boy, was I a sucker.

A friend of mine just inherited some prime real estate in Miami. Though it's shot up in value in recent years, and he'll have to pay high taxes to keep it, he's going to, because he believes that even if the real estate market cools, it will still appreciate 10% a year. (And this is a very level-headed guy -- a retired corporate comptroller.)

Few Americans alive today have ever seen any drop in home prices that wasn't just a temporary blip. Anyone who has thought that inflation has to stop someday, that the steadily rising debt loads will prove unsustainable, has been proven wrong again and again. This has been the consistent experience for more than half a century.

If you're sure there will always be a greater fool to buy it at a price 10%-a-year higher than what you paid, why worry about rents?

To be sure, there are people like me who know it was the same in Japan up until about 1992. But the Ben and Alan have has assured us that the Fed is on the job, and it can't happen here.

http://www.mitsuifudosan.co.jp/english/home/market_report.html

I thought Yglesias went to Yale.

I disagree with Brad's reasoning comparing the change in the rent/price ratio with interest rates. There is a similar argument used in the stock market where the earnings/price ratio is compared with interest rates (Fed model) and an argument is made that the earnings/price ratio should fall as interest rates fall. Even though empirically this is a good description of U.S. stock market since 1980, it does not hold before that period. Economically, it makes no sense.

Don’t believe my word for it. Read this paper by Cliff Asness

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=381480

Fight the Fed Model: The Relationship Between Stock Market Yields, Bond Market Yields, and Future Returns

[There is the unfortunate fact that stock market price/earnings ratios appear to fluctuate in response to changes in *nominal* bond yields, not *real* bond yields. This is a *big* problem for our models...]

Brad's argument makes perfect sense to me. Yet, if there's not a bubble in the Washington DC housing market, I'll eat my hat.

People here are overextending themselves for very marginal properties, others have turned into millionares in the course of three or four years. Prices have increased 100% in the last five years, meaning the average homeowner has watched their house double in value since 1999.

And I still don't see why people aren't cashing out their rental properties and converting them to condos. What if Matt's $10,000 in rental income could be converted not into a 100,000 sale to invest in bonds, but a $300,000 sale (which much more accurately reflects the current market, by the way)? Wouldn't that make a difference?

Brad's point is well taken but I am not sure if it fully explains the data. I live in DC in a very nice apartment building next to a very nice condo building. For the past 2.5 years, the prices of condo units have gone up dramatically up to the point where a small 1-bedroom unit on the 3rd floor with a view of the parking lot and garbage dump just went for $450,000 (up from about $350,000 when we first got here) whereas our top floor apartment with a view of the zoo still rents for $1350 a month, as it did 2,5 years ago. Interest rates haven't gone down much at all over this period. How can we make sense of such a large shift without assuming some sort of speculation?

I think a lot of this depends on your time frame. A major income property buyer/developer can possibly afford to lose money for 5-10 years (negative income flow) on a building. They are likely grabbing as much capital as they can while the capital is available at cheap rates. And capital is simply gushing at them right now.

For a small investor or individual looking at a home or investment property, it's a different story. It matters greatly when they buy and at what price.

Few individuals could afford to wait out a downturn and hope that rents increase over time to make up the losses. A bad real estate purchase could wind them up in bankruptcy, like any other bad investment. But right now, no one believes that real estate is risky---a philosophy that many in the commercial real estate business would beg to differ on.

If California were to experience even a mild 10-15% downturn (something which has certainly happened in the past after monumental runups), legions of homeowners would feel the pain.

Many wrongly believe that, if values decline, they would just wait it out and not sell---using the old stock market canard of "it's not a loss if you don't sell."

To some degree, that's true. But homes in this state are turning over every 5-7 years. Rare is the owner who pays off a 30-year mortgage, people are much more mobile and move around these days.

What many don't understand is that buying property is a leveraged transaction---you are only paying 5-10% of the price and the rest is borrowed. A 10-20% decline in value not only wipes out equity but forces you to dip into your pocket to get out of the deal. And with prices so high, it would be a big hit for many. Professional investors understand the risks and rewards of leveraged investments/speculations. Not many individuals do. Houses only go up in value, after all.

SFJOE,
On to the interesting part of your story! Where did you put the money? In new rentals somewhere? TIPS?

It is hard to explain the jump in prices in DC the past few years. In 2002, one needed only modest expected price appreciation for buying in DC to make sense, assuming one planned on staying for a little while. Now, you would need sustained growth of 7 percent a year in value--something that is not very likely (unless inflation in general comes back).

Richard Green wrote, "The average house built now is about 50 percent bigger than it was 30 years ago. Thus if we use size as a proxy, the typical house will be in real terms 50 percent more expensive than the typical house of 30 years ago, even if 'quality adjusted' prices have remained constant."

No. Most of the price of a house is the value of the site.

If you construct a 50% larger house on a site of the same size, you won't have a 50% more expensive house.

Furthermore, many new houses are constructed farther from city centers, and hence their site value is lower for the owner, if the owner has to commute farther to work.

kharris wrote: "Yglesias gets lots of things right. Can't we look the other way when he says silly stuff about real estate valuations?"

Well, Matt eventually conceded, and appears to have learned something new. That's worth something, isn't it? This is the kind of socratic dialogue you dream about in school... but only now can witness through the wonders of modern blogotechnology.

I thought Matt's first posting was more excusable than the second. Initially, he did not consider the effect of interest rates.

After this was pointed out, he should have retracted his claim that the significance of price/rent was invariant with respect to interest rates. As a non-economist, I would find it remarkable if any indicator did not depend on interest rates. It's also clear that the decision to buy vs. rent depends in an obvious way on the mortgage interest rate. Matt presented an interesting case (decision to keep collecting rent vs. selling) which does not depend in as obvious a way on interest rates. But it still should have been clear at that point that you cannot neglect the effect of interest rates, unless you could prove that the effect of the buy/rent decision is canceled out by something else, which seems highly unlikely and which Matt was not even trying to prove.

Even with all the explanations I have read here, I still believe Brad's reasoning is wrong. The link to the paper on the Fed model shows why it is wrong for the stock market and by using similar arguments you can show that it is wrong for the housing market.

The first part of Brad's argument as repeated below makes sense.

"Now suppose that Treasury bond interest rates drop from 6% to 5%. Invest $150,000 in Treasury bonds, and you only get $7,500 a year. All of a sudden going into the real estate business and renting to Matt and his ilk looks like a good deal: you get $3,000 more a year, and the hassle of listening to complaints about the holes in the floor is well worth it"

At this point even if you get $3,000 more a year than your investment in Treasury bonds, your future growth in rents (assuming they track inflation) is going to be smaller as Treasury bond interest rates drop from 6% to 5% because of inflation. The price you pay for the home should be the discounted value of all future rents and if you do the calculations you will find that with only a change in inflation there should not be a change in the rent/price ratio. Now if real rates change, it is a different story.

Let me know where I am wrong.

By the way, it's "Shaw" not "the Shaw".

As in, "You know, I was driving through Shaw yesterday on my way to work and spotted another crack house being converted into luxury condos starting at $500,000. Wow, we can't even afford condos in the ghetto anymore. Let's flee to New York City, where our friends are finding places on the Upper East Side for a nice $350,000."

[They are? How?]

sfjoe wrote, "I know lots of big money players are stretching for income right now. That's why you see junk bonds trading for an historically low premium over Treasuries---anything to get an extra 1-2% of income. Don't think that will end well, but I'm no expert."

Another sector is REITs. I've had money in Vanguard's REIT index fund for a few years now. It's just been going up like gangbusters. My guess is that it's due to investor's chasing yield.

Gavin's comment is correct. If rents are expected to grow at rate g, then value should approximate

V(0) = Rent(0)/(r-g)

where "0" is today and where r is the interest rate. If g and r move in tandem, the lower rate should not have an imoact on value.

However, if you look at CPI rent growth, which might be the signal people use for rental inflationary expectations, it has stayed pretty constant for awhile, meaning that homebuyers may be assuming that the lower r is not being accompanied by a lower g. This would prompt them to bid up prices.

http://flagship2.vanguard.com/VGApp/hnw/FundsHoldings?FundId=0123&FundIntExt=INT

Notice that the earnings growth rate for the Vanguard REIT Index is negative. This is a 3 year growth rate to the current month, and shows that REIT earnings have actually declined in 3 years. REIT's are have another superb year, but the increase in share price is not being driven by earnings. Rather, we have REITs being bid up in expectation of rising property prices and because the yields are higher than for other categories of stock. The earnings growth rate is -2.7%, while the p/e ratio for the index is 34.4.

Anne,

From what I've read, P/E isn't the right number to use for evaluating REITs. Rather, it's something like funds from operations. (Don't recall exactly.)

As my previous post indicates, I agree REITs have been bid up by yield chasers.

Gavin,

Isn't the criticism you mentioned of interest rate-based valuation models fixed by just using the real rather than nominal interest rate (i.e. TIPS instead of traditional treasurys)?

With real interest rates also very low right now, you might still conclude that real estate prices should be high. (Some would say short rates are even negative, depending on which inflation measure you use.)

Finally, Asness does some interesting work with a generalized Fed model, with an added parameter based on the ratio of vols of equities vs bonds. It does a very nice job of fitting the data.
(Figures posted here: http://infoproc.blogspot.com/#110331086983920757 )

The fit to the data is so good it makes me think that there *is* something to these models. Of course, in economics (unlike physics) if enough people believe in the model it will be predictive, even if "fundamentally" unjustified.

Richard,

You are right about the valuation formula you use. But you are not right about CPI-rent. The CPI-Rent inflation rate (you need to use the owners' equivalent rent to get at comparable properties) has come down sharply in the last 4 years. So, the g in your formula has actually come down sharply as has r. Through the end of 2001, the difference between these two narrowed sharply (a positive effect on prices). Since then, they have widened somewhat. Looking at it in the context of last twenty years, housing is a bubble. I am surprised that someone like Brad would swallow that justification about "nominal" interest rates.

The Upper East Side in Manhattan is surprisingly not that bad. At least by DC standards right now. A friend of mine purchased a co-op on the Upper East Side a few months ago for about $325,000. A quick search of the NYTimes classifieds turns up over 100 hits for 1 bedroom co-ops and condos on the Upper East Side between $200,000 and $400,000.

We're actually thinking of moving to New York and heading for the private sector. Non-profit and goverment salaries get a lot less attractive when home prices are this high.

Steve,

A model comparing rent/price ratios for houses or earnings/price ratios for stocks to nominal yields makes no economic sense. That said, I did read the paper describing Asness' extention adding the parameter based on the ratio of volatilities of equities vs. bonds. But as he emphasizes, just because it empirically fits the data it is not a good model to use to predict future returns.

Real interest rates have fallen from where they were in 1999 but not as much as nominal interest rates. If you use nominal rates to explain the trends in the rent/price ratio you will be stretching to explain only part of the fall. But if you use the smaller fall in real rates, the housing market seems even more in a bubble.

Srinivas,

I too am surprised that Brad did not mention the difference between nominal and real rates.

Hi Gavin,

I definitely feel that the housing market is in a bubble right now (more so in some cities than others), although I think Brad's calculation is OK if you use real rates. (Even then, housing is subject to a big decline if real interest rates go up.)

I'm not so sure about equities, though. I'm not sure why Asness thinks his vol-modified model doesn't have predictive power - despite its creaky foundations it may influence the behavior of real traders. It would be interesting to see a fit using the real bond yield Y in place of the nominal yield.

I would not be surprised if traders right now are discounting future inflation increases (given that the Fed was recently worried about deflation), so that the disconnect between nominal and real interest rates is not as big as you might think.

Todays Palo Alto's throwaway paper reported a 25 % drop in year to year Palo Alto home sale prices.
Other cities in Santa Clara Co and San Mateo Co were
up, steady or small drop in price. The Mercury News
some time ago reported my zip code dropped 40%.
Not sure of the expected volitility of this data.
I still want to know where SFJoe put all that money after the sale of that apt house. I have income property I would sell if I could figure out where to put the money.

Question from a non-economist: the theory suggests
that low interest rates will cause high real-
estate prices. But what happens if interest
rates are currently low, but people expect them
to rise in future ? Surely then you're best to
sell your property before the rise in rates
causes the price to drop ? And then if everyone
is expecting rates to rise, there should be a
rush to get out of real estate first, so the
price drop could actually come *before* the
rate rise. And with the current fiscal and
trade deficits, we must surely be heading that
way soon ?

liberal: Workplaces are not necessarily in the city center. Many companies have relocated out, or new office parks are built outside the cities. That's part of the suburban sprawl story. But be that as it may, if you want a reasonable choice of employers or office space conveniently located w.r.t. your customers, that almost always entails commuting, regardless where you live.

Richard Green writes that the average house now is about 50 percent bigger than 30 years ago, and so will be 50 percent more expensive even if 'quality adjusted' prices have remained constant.

liberal takes issue, stating that most of the price of a house is the value of the site. (Certainly true in the markets where we're seeing bubble prices.)

Another weakness of the "50%-larger equals 50% more expensive" argument is that construction productivity has risen quite substantially, while structural quality has fallen. According to a professional home inspector I know, nowdays even very expensive houses are cheaply built.

“... construction productivity has risen quite substantially, while structural quality has fallen.”

So very true and the two are connected. Two by fours have actually shrunk over the past 60 years and the spacing between joists has increased in some places. Wallboard has completely replaced plaster even in very expensive homes. In the 1940’s wallboard was considered unfit for everything except warehouses. Now it’s difficult to find a plasterer, they have all pretty much retired. Even as far back as 1960, the real craftsman who could do “run plaster” work had to be brought out of retirement to do the Lincoln Center in New York. Carpet over plywood has replaced hardwood floors except in more expensive custom homes and even there you find wall-to-wall carpet in the upstairs bedrooms. However kitchens and bathrooms seem to have improved significantly in almost all respects. If you are lucky enough to have a custom home then modern windows can be far superior to those in the past. Double-glazed “low E” glass windows with exterior cladding are terrific. Modern electrical wiring (if done right) is far superior to the old knob-and-tube cotton covered wire. But water heaters don’t last as long because they are generally not copper. So all in all, the picture is mixed. While older homes can have more charm, style, and better interior finishes, modern homes have better layout, use of space and sometimes better materials. For example plywood-sheathing beats the old fire stop cross members. Cedar shingles are a firetrap. Modern roofing materials are much better. Of course you have to find a competent roofer.