Housing Bubble Economics

There has been a lot of main stream media coverage of the Housing Bubble lately.  I've been covering it for over two years.  A lot of the discussions over at Patrick.net recently have focused on buyers, sellers and prices.  So, let's talk some basic economics...

Price Stickiness:

Prices in residential real estate tend to move rapidly on the way up but sticky on the way back down.  That is in a normal market.  The market has been anything but normal over the past few years -- it has been an out and out Bubble in many parts of the country, including the San Francisco Bay Area.

The Bubble started to come apart when upwards stickiness, or Buyer Stickiness, started to kick in.  This caused there to be an ever smaller pool of willing buyers at sellers' prices.  In short, buyers ceased to be price takers.

Now that we've hit the top, we enter the period where there are essentially no buyers at sellers' prices and no sellers' at buyers' prices.  Both sides are trying to be price setters.

What happens next is some sellers start accepting the market and accept buyers' prices.  Over time, more and more sellers convert to this understanding of the market and eventually equilibrium is returned, where neither side holds a commanding price setting advantage.

The Inflating Bubble:

• Pre-Bubble prices started out at the equilibrium price of P(-1)
• As the Bubble expanded, a few buyers started opting to refuse sellers' prices, and are represented by the right-portion of D(0).  There are not yet enough buyers on this portion of the curve to affect sellers' prices.
• However, most buyers continued to participate in the market, and are represented by the left-portion of D(0).  The vertical span on D(0) is the break between buyers who buy into the bubble and those who refuse.
• Sellers, who started early in the Bubble on S(0), selling at P(0), quickly adjust their expectations.  Sellers shift to S(1), requiring even higher prices.
• As sellers shift to S(1), buyers shift to D(1), but the number of unwilling buyers continues to grow.  Sellers are still able to get fairly consistent prices even though there are less buyers because there are still enough to support a definable market.  (This is represented by the sloped, upper left portion of D(1).
• Eventually the pool of buyers at D(1) dries up enough to shift to D(2).  This curve shows that most buyers are sitting out the market.
• Sellers on S(1) start finding erratic prices, represented by P(1).  This is because they are selling into the disconnect between buyers.  This period was marked by a wide variation in prices of homes sold compared to one another.
• Eventually sellers climb all the way to S(2).  Note that now sellers are unwilling to accept any price below somewhere in the range of P(1).
• A few buyers on D(2) are left, even at the sky high prices P(2).  Eventually those buyers dry up.

Popping the Bubble:

• There are now no buyers left willing to take sellers' prices.  The upper portion of buyer demand is gone.  The buyer curve is represented by D(0).
• D(0) comes in under S(0), so sellers cannot even get their minimum price of P(0).
• Sellers start shifting their prices slightly, and minimum prices fall somewhat as the quantity of unsold homes grows.  But even at S(1), there are still no buyers on D(0) willing to pay sellers' prices.  The lower leg of S(1) are sellers who have moderated their prices slightly.
• Some sellers defect, and the curve moves to S(2).  The few sellers willing to drop their prices substantially find buyers on D(0), but the prices they get are erratic.  This is again because of the vertical portion of the curve, this time in S(2).
• Buyers, sensing declining prices, shift down to D(1) and the few willing to buy at higher prices purchase the available supply on S(2).
• More sellers realize that lowering their prices is necessary.  Their curve shifts to S(3), which clearly delineates price-taking sellers from price-setting sellers.  Ironically, some sellers may actually be expected to raise their prices during this phase, even in the face of adversarial demand.
• Buyers purchase the lower end of S(3) and shift a bit further to D(2) as they still expect prices to fall, seeing the large inventory of overpriced, unsold homes.
• Between S(3) and S(4) sellers make a critical decision to either sell or not sell.  Sellers either leave the market indefinitely or convert to price-takers.
• Finally, equilibrium is reached at P(3), where S(4) and D(3) intersect.  Neither sellers nor buyers control prices.

Conclusions:
Price stickiness is a fact of life in the residential real estate market.  It occurred upwards near the end of the Bubble's inflation period, and it is occurring now, early in the Bubble's deflation.

This analysis is basically of a hypothetical set of comparable homes at an equivalent price, within a homogeneous region.  In real-reality, of course, real estate is heavily affected by hyper-local factors.  Therefore this analysis is merely a mental exercise to help in understanding general price movements in the face of sellers and buyers.

Comments

[DinOR]

Randy H,

Wow! That says it all. "The difference between "price taking" sellers and "price SETTING" sellers! Dorsey Wright and Assoc. performs similar charting w/stocks.

[Robert Coté]

Randy H,

Aren't you assuming volluntary participation by a homogenous group of sellers and the same for buyers? Additionally I see no consideration of time value decay in a falling market or time value pressure in a rising market.

I'm sure sellers would always be sticky when they have that choice but prices are set at the margin often by people who have no choice. Then there's different classes of sellers unlike your assumed homegenous class. REOs have very little sticky anywhere above outstand loan balances in their portfolio. What little sticky there is gets wiped out by time pressure to get a nonperforming asset off the books by the next reporting period. The same arguement goes for the buyers. At current levels there are no rational investors, only irrational investors and a core of owner/occupiers that are always in the market for lifestyle reasons. Your analysis works for a rational contigous market. There's are reasonable arguement that the current market is neither.

[randolfe]

Robert,

The definition of "rational" here is more along the lines of behavioral economics. That is, rational is as people do. It matters not that the actors are objectively rational. Only that they believe themselves to be behaving rationally, even if that behavior is objectively irrational. I contend that in bubbles panic buyers are indeed acting rationally to the best of their own capacity. They are attempting to avoid greater losses they perceive they will suffer by not buying.

I am assuming a generic set of buyers and sellers for a homogeneous set of comparable homes, in goring micromarket factors (which can override many other variables in many cases). But I am not assuming homogeneous buyers or sellers. To the contrary. I am assuming a dynamic whereby buyers and later sellers represent a fragmented, discontinuous set. Nonetheless, buyers are still on the demand curve because they are actively demanding houses, just not within any intersection of sellers' prices. Vice versa on the way down.

I am purposefully ignoring all non-owner occupier units, including foreclosed properties and investor speculative inventory. These do not constitute a very large absolute percentage of the overall set of inventory.

As to marginal prices, I contend this is false for housing. It is true only as a mathematical abstraction. That is because a realistic buyer does not have a large enough set of potential inventory (in most cases) to successfully approximate integral prices. If there are a mere 15 homes in buyer's target set, and 10 of those sellers are holding out refusing to drop prices, then his choices are 5 homes, which often are priced significantly differently, especially during growing seller defection. If the low home is $419K and the high is $625K in the lower set and $799K-$1029K in the upper set, we cannot say the price of all these homes is $419K.

This is because to say homes are priced at the margin assumes that this market is always a price-taking market for sellers. That is objectively false. As evidence for this contention I offer the very dynamic occurring today: growing inventory. In a marginal price market prices would always be set at market clearing price.

I do realize, however, that buyer psychology is one whereby most hopeful buyers strongly wish to believe that market prices are set on the margins. Anyone's who's ever bought a home knows that this is, in reality, almost never the case.

[Gavin]

For a detailed study on illiquidity in a declining real estate market see this paper from the Federal Reserve

http://www.frbsf.org/publications/economics/papers/2004/wp04-16bk.pdf

[Editor: Added local document link here]

[randolfe]

Thanks Gavin. Having just skimmed through it, there is a lot of great meat in that Fed study.

[Robert Coté]

First; I stand in awe of your professional accumen.

Second; thanks for the forebearance of my amateur observations.

Alright; Please just consider this; you are saying that you ignore 40% of the 2005 purchases in the Phoenix area, for example, in your model. And I still don't see any consideration of a sale by a 1st note holder who ignores all other potential debtholders be they 2nds, HELOCs or investor/speculator.

The best part of this disagreement is that witihn a year or 6 quaters at longest we'll see without question whether this time will be sticky like every time before or whether this time is different in this one respect.

[randolfe]

Robert,

We shall indeed know for certain in a few short months. I would only contend that we are already well into a sticky period now. The question probably isn't one of whether prices will be sticky or not; but instead whether they'll be significantly more or less sticky than previous episodes.

Your observations are far from amateur. Please feel welcome to add your knowledge to the pool here anytime.

[TN]

Can the sellers afford to be as sticky in relation to previous downtrend episodes? If you classify all the factors that influence the stickiness coefficient (savings,leverage, income, interest rates, prices...etc) normalized historically, it seems that the 'its going to be a lot less sticky scenario' is an overwhelming favorite.

[Robert Coté]

Randy H,
Thanks for the kind welcome but I know when I'm swimming way over my head. When you extend the proposition; "whether they'll [prices] be significantly more or less sticky than previous episodes" I accept the life preserver. That is indeed what we are discussing. I'm more of a visual thinker. I imagine the current conditions to be one of a wave cresting. There's solid water real close to the highest prices and alot of momentum behind them but go a little deeper and you hit free air with no support. This rough analogy explains the flight to quality and market bifurcation we are seeing a little bit of. I'm still concerned that a lot of 80/20 loans will turn into 80/screwed loans and take your vertical pricing notches far lower than you model.

[randolfe]

There are certainly many fundamentals aligned suggesting that prices may actually be less sticky this time around. I am not ruling this out. I am just coming up short finding a coherent model that would support that outcome.

It may ultimately be a problem of framing. If there is a hard landing, then whether there was stickiness or not depends largely on when one believes that "rational" sellers should have first cut prices. Maybe the sticky-camp is simply putting too much faith in sellers' ability to behave rationally.

I still hold that the debt overhang model proposed in Gavin's contribution above will gum things up as sellers refuse to take nominal losses, even if that decision ultimately exposes them to even worse risk of foreclosure.

In the end, everyone thinks they'll be the lucky one to escape disaster.

[randolfe]

I found this in Mankiw, Macroeconomics 5th ed. It doesn't entirely apply to real estate, but much of it does.

Ironically, I find myself arguing with many practical believers over at Patrick.net about whether real estate prices are/will be sticky during this correction. Why ironic? Because academic types were criticized for *not* considering price stickiness in their models a lot in the mid-90s. Business practitioners knew that prices were essentially almost always sticky everywhere, and that the nice equations and graphs economics professors drew were not reflective of the real world.

--

Theories of Price Stickiness:

1. Coordination failures. A lot of this has to do with game theory. "Who goes first" when lower prices, etc.

2. Cost-based pricing with lags. Applies to home builders more than existing home sellers. Tendency to wait for costs to rise/fall before raising/lowering prices.

3. Service & transaction lags. This is prevalent in real estate. Deals take time, so the market will tend to be backwards looking in regards to prices.

4. Implicit contracts. This is happening now with home builders. It is also game theory. Price coordination. They try real hard to prevent a price war, thus all the complicated incentives.

5. Nominal contracts. Explicit contracts exist and must be either completed at price, or discharged for option-exit exercise cost. That exercise cost creates price lags.

6. Costs of price adjustments. There are costs associated with changing prices, either direction, but usually more so downward. Again, the price adjustments will happen less frequently and be bigger in order to minimize these costs.

7. Pro-cyclical elasticity. Demand curves become less elastic as they shift in. This is true of many things, including real estate.

8. Pricing points. Psychology. This is a huge factor in real estate. It also affects (12) below.

9. Inventories. Suppliers who can manage inventory levels will do so before shifting prices, since it has the same effect on their bottom line. Home builders are doing this now.

10. Constant marginal costs. A lot of bubble-believers who are anti-sticky-prices like to often refer to how sellers cannot/will not be able to get away with "needs based pricing". Nonetheless, there is an absolute marginal cost to both home builders and existing home sellers alike. Neither will accept a price below their marginal cost. At this point, foreclosure and/or bankruptcy are preferable exits costing less money.

11. Hierarchical delays. Not so applicable to real estate except for some bigger home builders. This is the bureaucratic effect.

12. Judging quality by price. Another psychological factor. In real estate we'd call it the "prime area premium", or the extra amount over fair value that a prime area can command. Prices are supported by a psychological perception that higher prices translate into better houses/neighborhoods/schools/etc.

**"On Sticky Prices", N. G. Mankiw, Chicago: University of Chicago Press, 1994), 117-154.

[Boston]

I have a thought regarding the "price set on the margins" debate. Your argument seems to rest on the fact that many sellers do not drop prices to the market clearing price. This makes intuitive sense and I'm sure most people would agree that it reflects reality.

On the other hand, I think most casual observers of the market are thinking of the frequent use of comps in real estate. While many sellers do not meet market clearing prices in a falling market, one could argue that the "value" of their homes has fallen nonetheless, based on the comps of the homes around them. Just because sellers do not acknowledge this doesn't change the situation. Maybe it's just an issue of semantics, but it might make sense to say "the price of a home is not set on the margins, but the *value* of a home is."

What do you think?

[randolfe]

Boston,

While many sellers do not meet market clearing prices in a falling market, one could argue that the "value" of their homes has fallen nonetheless

The problem I see is the reflectivity in that thinking. For real-estate, it is exceedingly difficult to discern the value outside of an actual sale price settlement, because the market is not liquid and the comps are usually not direct substitutes. Add to that the fact that when a seller refuses to drop her price to market-clearing price, she is in fact affecting her own price as a function of value because she has effectively restricted supply.

The really only consistent way to value real estate is to value at book and only mark-to-market when sold. The same is true of any illiquid market.

I would definitely agree that the *perceived* value of a home is set on the margins by comps. Add in mental accounting psychology, and this would go a long way towards explaining non-optimizing behaviors of sellers overpricing their homes. They *think* they are more valuable because other people think the same about their own homes.

[Boston]

Randy,

Thanks for the reply. I have another question. Does the prices-not-set-on-the-margins theory rely on some buyers making a "sub-optimal" purchase decision? If as you say each seller has his own supply curve, isn't it critical that on occasion a buyer jumps up to one of those higher supply curves? If this never happens, then it seems those homes are not really on the market.

When your hypothetical seller refuses to drop her price you say it restricts supply—maybe it would be useful to have an inventory tracking measure that excludes homes priced out of the market range? (Obviously this would be difficult to measure.) I guess this leads to the phantom inventory debate on Patrick.net.

I also have a comment about your last post. You say it is difficult to find direct substitutes for homes in many cases. I agree with Robert Cote this is true most (but not all) of the time. Where I live in downtown Boston, it is true that every unit is truly different. But a couple I know purchased a townhome in New Hampshire without ever having seen it in person. It was not an investment—they are still living in it. How were they able to do this? Well, they had toured other, virtually identical units. When an offer fell through on one of those, they were able to re-bid on a different unit. To Robert’s point, I think this is common for SFH in some subdivisions as well. I think that taking townhomes, tract homes, and some high-rises into account, this is a non-negligible percentage of the housing stock. Perhaps state-of-the-art pricing theory could be modified slightly in such situations?

Thanks for all your insights.

[randolfe]

Boston,

You bring up some great points. To your second point, I definitely think that pricing theory could be more readily applied to mass-produced homes. Focusing on the segment that demands new(er) homes/units, especially in homogeneous, CCR/association neighborhoods/complexes, there is a strong argument that reasonable substitutes exist. I had not considered this angle before.

As to seller & buyer curves, I definitely see buyers on different "non-optimizing" curves as well. This is largely due to the psychological factors which loom large in home purchases. In fact, the buyer is not irrational if they optimize factors other than pure financial value. And even an individual buyer will "jump" a different curve. A big raise, having a first child, having surprise twins, getting married, are just a few common events which can cause a buyer's utility optimization function to change shifting them to a higher price curve.

Referring back to a big Patrick.net debate I had a while ago, this was the point I was trying to make vis-a-vis "women working". It really is irrelevant whether one believes women in the work force affects some value-based notion of affordability. The fact that women do work, and they contribute their income to the family, means that a large number of potential buyers shift to a higher price demand curve. This soaks up supply at those levels soaking up potential supply at lower levels. It's like arguing that inflation isn't important because a penny used to buy a lollipop.

**It is important to note that when I keep saying "different curves" I really mean there is one distinct demand and supply curve, but they are discontinuous functions, almost step-functions. This is important because it is when an intersection occurs across a discontinuity that prices become erratic, information is misleading, and buyers & sellers are uncertain and apprehensive.

I think this is the lesson I've really taken from Patrick.net. Most people there seem to be waiting for sellers to become "eager". I do not see this happening for some time, if ever. It is quite likely that sellers and buyers both remain apprehensive for months or years while prices vacillate sideways with a downward drift (meaning inflation does much of the work of correcting).