Economic Papers

what are we out now in manhattan and where were we at peak?

So you guys think that the 15X ratio they mention as historical signals the bottom, as they state in this paper? Steve demands 12x, looks like NYC was 21x... will you feel comfotable buying at 15x now?

I've said for a long time that Steve's 12x was off, and the number is actually 15-20 depending on other factors (rates, taxes, etc.).

That being said, that doesn't mean you buy at 15... its just saying thats the equlibrium. You'll have to overshoot it to see a bottom.

and if you're cautious you'll wait until rents stabilize.

15 vs. 18, very narrow range.

This article suggests a 25-50% decline is in store for NY metro area. It reports NY as having a ratio of median (not comparable) sales to median rental of 21, while the national historic norm, is 15. It also takes 1995 as the last pre-bubble year, suggesting that we are returning to 1995 plus inflation. Why is this bullish?

Note also that the 15x ratio is for MEDIAN sale::MEDIAN rental. For comparable apartments, the ratio must be lower, since the median sale is higher quality than the median rental.

Steve's 12x is off; it is too high a price for an investor to pay and expect to make a normal return on equity invested without assuming (1) mortgage interest rates are low relative to inflation, (2) rents will increase faster than inflation over a long period of time, (3) extraordinarily low repairs, taxes and utilities, or (4) a bubble will enable capital gains by selling to someone willing to overpay even more.

I have no problem with 15x or being the NATIONAL average, which is what that paper says. However, that 15x takes into account the recent bubble. Factor out the bubble, and the figure is lower.

12x in NY according to the data.

That said, in places where rents are cheap - Houston, for instance - where there is an oversupply of rental units, that ratio can be higher. It can also be higher at times of low interest rates.

A buyer who hopes to be able to sell within 10 or 20 years and match inflation should assume:

1. prices are likely to return to trend (1995 plus inflation) and then rise with inflation.
2. prices are likely to return to a level at which an investor can expect to make money on its equity buying a condo and renting it, without taking into account possible capital gains and largely without treating the investor's labor as free. Historically rents rise with inflation in successful cities, but they can go up much faster short term, and they can plunge permanently in unsuccessful cities, so investor perceptions of NY's job market will affect their projections of future rents.
3. prices are likely to approximately the cost of building comparable new units, assuming that land sells for its value in its current use (i.e., for rental value as a tenement or parking lot).
4. prices will deviate from 1-3 if investors are optimistic or pessimistic, as they usually are. It is very hard to predict whether optimism or pessimism will prevail in the future. However, you need to be able to survive the inevitable bouts of pessimism.

Correction: The appendix makes clear that Baker et al have adjusted, somewhat arbitrarily, for the difference in quality in sales/rentals.

So their 15 x ratio is neither for median:median nor for comparable apartments. It is not empirical evidence that any unit anywhere should sell or ever has sold for 15x its rental value. Or 12x.

There is a bias in the authors, as well, as their objective seems to be to support sales prices and prevent them from falling further. That, of course, is impossible!

stevejhx and financeguy

...so currently NY is at 20x. I want to ask your personal opinion,I know there are people who disagree with you, but would like to hear your thoughts regardless.

If I was to buy an apartment that would rent for $4,000, for a two bedroom,(and ofcourse there are many other factors, I know I know! every neighborhood, apartments are unique, but just in general (ex-herlema and brooklyne)), how much should I be paying for it and know I made a smart investment decision (not emotional one ie. "depending on how much I love the place" etc)? thanks

Steve: Dean Baker is a perceptive economist who was one of the first to call the bubble, years ago. He is also the leading proponent of dealing with the collapse of the bubble by allowing former homeowners to rent their homes long term at market rates after foreclosure. Your accusation that he is trying to talk prices up is just silly. Read his blog "Beat the Press;" you might learn something.

jjun:
Short answer: $400,000-$480,000 depending on how optimistic you are. Which exists in lots of places, but not NYC.

Not so long answer: If you are buying for an indefinite long term, I'd expect a ratio of around 8-10 x annual rents for the same unit (somewhat higher using Baker's aggregate data). If you might need to sell, I'd take into account that if Wall St jobs don't recover fast, the ratio could could easily go lower than that and rents could drop too. There are no guarantees. NYC prices have dropped for decades at a time in the past and could do so again.

Long answer:

I'd think about the apartment as if you were an investor. If you bought it for cash, rented it for $4k (plus or minus whatever changes you expect in the future) and paid all the expenses, including repairs, vacancies, and something for your time, could you make a reasonable return -- say 3-5% over inflation -- on your purchase?

Reasonable here depends on how sure you are about your projections of future rents and expenses. The less sure you are that you know how much rents are going up or what future repairs/utilities/taxes are going to be, the higher a return you should demand.

The numbers you should use, ideally, would be future numbers: not current rents/expenses, but what you expect them to be over the life of the building.

Historically, rents in successful cities go up with inflation, but when cities have problems, there is no floor. So if you think NY is going to recover quickly, you might assume rents and expenses go up with inflation, but if you think that finance is going to be permanently smaller, you might expect rents to drop commensurately while taxes will need to rise to make up for the lost Wall St income.

You can ignore any gain or loss when you sell, because you'll be selling to someone making exactly the same calculation as you -- so capital gains and losses will just be (unpredictable) changes in estimates or differences between your calculation and your buyer's -- unless you think you have some way to increase rents or decrease expenses more than everyone else or to predict how people are going to be viewing the future when you sell.

As a practical matter, if you assume rents are going up with inflation (that's high, I suspect), that expenses will too, that the bubble is not coming back and that this prediction is moderately likely to be right -- you might use current rents and expenses and calculate a price that would give you 8-10% on your money. That'd be a fair price; a deal will be cheaper.

One reality check on your guess: a bank making a non-FNMA mortgage loan on the property would be investing in the same property, but with less risk and less work -- so ordinarily, it should demand a significantly lower return. If you accept a lower return than the bank, you are betting that you know significantly more than they do -- and while you may, you should be sure you're not kidding yourself. So if banks want 7% on Jumbos, 8% is probably too low for the equity holder.

Another reality check: Try to estimate what it would cost to create a comparable unit -- by building, renovating, selling a unit that is currently rented, converting a rent-stabilized or non-residential building. If any of those are profitable at current market prices, investors are likely to do it, thus tending to increase supply and decrease price. If FIDI offices sell for 350 psf and can be converted to residential for 350 psf, no price above 700 psf is stable (these are numbers I've seen in blogs, not based on any knowledge of real costs).

Of course, for coops, all this will be a theoretical exercise, but coop prices generally track condo prices close enough for these purposes.

As a rule of thumb -- to see whether it is worth making this kind of careful calculation -- most run of the mill apts should sell for between 100 x monthly rent and 10 x annual rent.

If it is selling for double or triple that, as would be typical in most of NYC today, you know the following: (1) probably some current landlord holding to rent could make more money by selling (which will tend to increase supply and decrease prices). (2) probably some investor could build or renovate profitably (which would tend to increase supply and bring prices down). (3) probably some potential buyers will decide to get the same apartment for less by renting (which would tend to reduce demand and prices). (4) possibly some buyers and employers will notice that other cities or neighborhoods offer almost the same urban experience for less money (which will tend to decrease demand and prices).

Of course, there are lots of stories you could tell about why the rule of thumb is giving too low a number, including (1) the government could increase subsidies for owner-occupants, (2) rents could unexpectedly jump or costs could unexpectedly drop more than you or the consensus expect, (3) the bubble could return and prospective buyers could conclude that overpaying is ok because someone else will overpay more when they want to sell, or that it is always better to write checks to banks/maintenance/taxes than to landlords (4) banks could be willing to lend at rates that do not reflect their actual risk because they are being subsidized or think they have shifted the risk to someone else, so borrowing will allow you to earn more money than the rule of thumb assumes (and lose more if your projections are wrong) or (5) this particular unit has unusually low costs, so the rule of thumb is underestimating profits. The question is whether any of those are plausible enough to gamble on.

Finally, remember that if NYC prices get high enough, there might be political pressure to bring our transit systems up to the level of Japan and double the commutable area of NYC. Or employers might discover the wonders of Philadelphia/DC/London/Phoenix.

FG, read the executive summary:

By comparing home prices to rents, as suggested by basic economic theory, this paper finds that
while most of the nation’s metropolitan housing bubbles have deflated and many markets never had
one to contend with, there is the possibility of a persistent housing slump in the years ahead. An
appropriate response to this problem involves:

1) Stimulating the fundamental demand for housing through acting to lower unemployment
and raise wages;

2) Recognizing a leading role for rental housing in federal foreclosure mitigation and
neighborhood stabilization policy, including allowing foreclosed homeowners to remain in
their homes as renters; and

3) Adequately funding the National Housing Trust Fund to capitalize on current low prices,
ensure long-term affordability in a recovery, absorb excess housing, and stimulate
employment.

Sure sounds like an agenda to me.

FG, thanks for the comprehensive answer to a vague question!

a) "Historically, rents in successful cities go up with inflation"

Wrong. Historically, rents are 100% correlated to income.

b) "You can ignore any gain or loss when you sell"

Wrong. Gains and losses are real - ask Las Vegas.

c) "Try to estimate what it would cost to create a comparable unit"

Completely irrelevant to price. The only thing that is relevant to price are rents, and leverage.

d) "most run of the mill apts should sell for between 100 x monthly rent and 10 x annual rent."

Wrong. First, it depends on the place, the interest rate, and the supply. But 10x annual rent is not sustainable in most instances as it would make the capitalized stream of rent payments that purchasing is some 20% less than market-rate rentals, giving an extraordinarily high return on capital. The real figure, depending on where you are in the country and other factors, is between 12 and 15.

Sheesh! I sound more like JuiceMan in that last post than I ever thought possible. :)

Steve:
1. If by "agenda" you mean that Baker has a political proposal to deal with the collapse of the bubble, that's correct. It just happens to be almost the precise reverse of the one you ascribed to him.
2. Rents correlate to income? Prices have nothing to do with costs? That makes little sense theoretically and I've never seen any evidence of it empirically. What is your basis for these claims?
3. 10% gross returns before expenses is "extraordinarily high return on capital"? What are you comparing it to?
4. as for (b), I suggest you reread what I wrote.

"If by "agenda" you mean that Baker has a political proposal to deal with the collapse of the bubble, that's correct."

And that is to support housing prices.

"Rents correlate to income?"

They absolutely do, as they must. Income is what you pay your rent with. If income goes up rents go up, and vice versa. It's almost what you say about "successful cities," but inflation has nothing to do with the price of rents. Rather, rents have to do with the rate of inflation - about 40% of inflation is rent and owners' equivalent rent. Therefore, your nonsense statement that rents go up with inflation is just plain dumb: rents are an input into inflation, the single largest one, in fact. They MUST be correlated.

You really don't know what you're talking about.

"Prices have nothing to do with costs?"

The price to build a place to live today has nothing to do with the cost of purchasing (or renting) an existing dwelling. NOTHING. Prices do have an effect on whether one will choose to build, but not the cost of the building itself.

Input costs have NOTHING to do with price. Output value does. Yesterday's prices have nothing to do with tomorrow's. These are fundamental tenets of modern economic theory.

You should stop spewing so much bombast lest someone believe that you know what you're talking about. Like when you disregarded perfect markets AND used Miller-Mogdaliani (or however you spell it), and when you applied the theory of manufacturing costs to the market price for housing.

Like these posts today, pure sh*t.

Happy to debate, Steve, but debates are more useful if you actually make arguments or cite facts.

Baker is calling for affordable housing, not higher prices. Read a little more carefully.

Income has gone up much faster than inflation over the last 100 and 300 years, as I'm sure you are aware. Thus, since inflation and income are not correlated, your claim that rents "must" correlate with both is somewhat implausible. Could you explain either the logic or the evidence?

I think you will find the theory that in competitive markets prices tend to converge to the marginal cost of production in the first or second chapter of every college textbook. If you have some reason to think that this market isn't competitive or the standard theory is wrong or doesn't apply, I'm happy to be educated. But to simply assert that prices and costs have "NOTHING" to do with each other? What's the point?

If you really think that housing is not manufactured, or that MM has no importance except in purely imaginary perfect markets, or that -- contrary to daily experience -- "yesterday's prices have nothing to do with tomorrow's," you should do your readers the courtesy of explaining why, not just spewing insults.

Finally, I spell Modigliani correctly. How do you spell it?

Equilibrium is a bit of a misnomer.

It is best to think of the equilibrium price, where price equals marginal cost, as a kind of gravitational force, counteracted by the quite different force of self-reinforcing trends. Just as the planets are always pulled towards the sun, but momentum means that they orbit around it rather than crashing into it, so to prices are always pulled towards equilibrium, but rarely reach it. Indeed, again like a planet in orbit, half the time they are moving away from it...

So using equilibrium to predict future prices is a lot like predicting the position of the earth by examining where the sun is. You are going to be wrong. But if the earth is heading away from the sun, you are going to be less wrong than those who simply extrapolate from recent trends and assume the earth is heading towards Alpha Centuri ("real estate prices always go up").

Efficient markets hypothesis is the theory that the earth is always in the sun. It's wrong: it ignores momentum (and the evidence). But that doesn't mean that gravity doesn't exist.

"since inflation and income are not correlated"

Stop, FG. Can't call you "financeguy" as it's obviously not true. Did you ever hear of "wage inflation"?

Steve, I give up. Do you really think that real wages were flat over the last century?

Or are you just in practice for talk-radio, deliberately misrepresenting and shouting slogans louder to avoid any actual engagement?

FG, I'm just copying the silly things you write. Wages have been flat for about the last 25 or 30 years. Over the last century wages have risen at at real rate of just over 1% per year, the same as housing prices. "Wage inflation" occurs when wage increases exceed productivity increases, which can (though may not) cause price inflation.

There are no "slogans" by me - you just spout stupid things, and I repeat them for the edification of all. Your pseudo-intellectual style belies the truth of what you write.

Steve -- If you keep the abuse down to a reasonable level, it might be possible to learn something. I write to learn, so if you have something to teach, I'm more than happy to hear it.

I don't see the relevance of wage inflation.

The question is whether real prices for housing track real wage growth.

I think real wages have gone up in the last century: we've had significant economic growth in that period and no one seems to worry about darning socks or stockpiling food and wood for the winter any more.

You very assuredly say I'm very wrong. You add (not quite correctly but close enough), that wages have been flat since the late-1970s. So I guess you think that living standards for median workers are the same today as they were during Reconstruction. Perhaps you also mean that in Amsterdam, where Shiller went back 300 years and found that real estate prices were basically flat, median living conditions have barely changed since 1700.

These claims are startling to me, but I'm here to learn. Please explain. Maybe I just don't understand the point you are trying to make or you have data I don't know about.

What is your source for claiming that real wages have risen at the same rate as housing or vice versa?

What is your economic model that would make housing prices a function of income, so that this correlation, if it exists, isn't an accident? You wrote above that "income is what you pay your rent with" but that's not enough: income is what you pay for everything with. Clothes have gotten much cheaper over the last century as incomes have gone up.

Usually, we -- or at least I and my teachers -- expect that as productivity increases, prices drop. Productivity certainly seems to have increased in building: I don't see construction workers using lath and plaster or shaping siding with hand tools much. So why is housing different?