Reversion to mean

[Mexico Resident]

There has been a lot of discussion about reversion to the mean and it appears there is a loose concensus here that this is likely to occur. But what is the “mean”. If you take 1997 as the base year, you have included the real estate bull markets of 1979 and 1989. Why not unwind the last 30 years? Then real estate in San Diego will be priced like the rest of the country – something like $220K is the SFH mean in the entire USA for 2006. I feel that there is a point to be made in saying that San Diego is a highly constrained market. This implies it has a price premium but also more volatility in the price. It’s sort of like the situation with oil or any other commodity. If people feel there isn’t enough, then the price can climb (or fall) rapidly due to minor market condition changes.

[4plexowner]

Here is how I think of reversion to the mean:

Over time prices oscillate around a trendline. Sometimes prices are above the trendline – sometimes they are below. The trendline may be flat or it may be upward or downward sloping.

When prices reach an extreme on either side of the trendline there is a tendency for them to head back towards the trendline or revert to the mean. Oftentimes, prices overshoot the trendline before stabilizing on or near the trendline.

That describes the behavior of reverting to the mean.

The only question left is, “What is the mean?”, and that is the question you are asking.

I think it is 1998 prices because, IMO, that is when the price of real estate and the rent that could be obtained for that real estate started to diverge dramatically (ie, from an investor’s viewpoint, property prices relative to rents no longer made sense).

Other people have different opinions about what the mean is and some people want to complicate the matter by trying to account for inflation.

I’m curious to see what others have to say on this topic.

[poorgradstudent]

The mean that I think of is the mean value for the ratio between housing prices and incomes. If real incomes rise, you expect housing prices to rise at similar rates, and the converse is true.

Because of inflation, you’re not going to see a return to nominal 1998 prices. Also, according to the data I saw, real, inflation adjusted incomes are up, roughly 36% since 2000.

Nominal prices are going to continue to fall. Just don’t expect the median to fall back to the low $200s. From my crude math, based on current incomes somewhere around $400 for nominal prices seems reasonable, but I’ll be the first to admit my calculations were quite quick and dirty.

[Anonymous]

I agree with your logic 4plexowner. I also think 1998-99 prices are the more sustainable mean, and comparing to rent is the most logical and common sense measure.

Poorgrad, not sure how you calculated that 400K is a reasonable price based on incomes. The median household income in SD is 63K. When lending standards return to normal, this will only get you into a 200-250K house.

[blahblahblah]

Just say “200-250K in 1998 dollars” and you’ll take care of the issue. That’s where we’re headed.

[(former)FormerSanDiegan]

Nominal 1998/99 prices are not the mean. It’s the ratio of prices to incomes in 1998 that matter. I’m with poorgradstudent. You need to account for increase in income to come to the nominal price.

deadzone – Median income never buys a median house in SD. COnsider that the homeownership rate is 50-60% histornically in SD. Significantly below national averages. Since we have a large fraction of renters (more so in the past than now), I would suggest that median incomes are buying the bottom 20th percentile or so of homes, not the 50th percentile (median).

[(former)FormerSanDiegan]

Just say “200-250K in 1998 dollars” and you’ll take care of the issue. That’s where we’re headed.

That’s around 320-340K by my calculations in today’s dollars.

[powayseller]

There are 3 ways to calculate what the housing prices should be today. Draw a trendline to 1997, and keep going from there. Where do you end up? I’ve done this for SD and US, and am now working on the other So CA counties, and it will be for fee on my website ($10 or less per month). So I won’t give you my results, but here are the 3 methods:
1) historic prices
2) price/income ratio
3) price/rent ratio

So look at the data until 1997, and then extrapolate the trendline. Why 1997? That’s when the tax law changed to allow you great tax benefits for owning real estate. You no longer had to pay capital gains tax on huge profits. Shortly after 1997, the real estate prices started a rapid acceleration.

This is my opinion, but the charts are powerful.

Look also at the great charts Rich and jg have made.

[qcomer]

I don’t agree with calculating the mean baased on historical median prices for last 80 years. Median or Mean are statistical estimators of a trend. However, it is common practice in statistics to ignore data beyond a certain sample window as that old data is considered “stale” and may skew the median/mean. The reason is because wars, taxes, economy,foreign investment, etc all make the dynamics of each cycle different.

A housing boom and bust cycle on avergae spans 13-15 years or 50-60 quarters. So I calculate mean based on median prices for each quarter, for the last 50 quarters. e.g. the tax changes PS described, benefited home ownership and added a premium over the mean for the next boom/bust cycle. This added premium bumped up the median for thuc cycle from the last cycle. Extrapolating a straight line from 97 would underestimate the true trend, as it ignores the tax premium existsed.

I think median will be around 400K in 2 years and that will be true reversion to mean. That is probably 370K in today’s dollar terms. But the downward momentum may push the median prices further below for sometime onwards.

[powayseller]

Excellent points.

[kev374]

Historically I believe RE prices have exceeded local inflation by about 1% but you also have to take into account income which has pretty much stagnated and even in some cases receded in the past 5 yrs.

Income is the only fundamental that drives demand. People keep harping on the fact that prices are high because everyone “wants” to live in SoCal etc. which may be true but everyone wants a Ferrari as well so are Ferraris selling in the thousands? I think not. The median household income in SoCal is around $60ish K. That doesn’t support a very high price point, perhaps $200k or so.

Infact per 2005 Census estimates only 25% of SoCal makes 100k or over in household income. My guess is that the vast majority of that 25% are older and already homeowners are not creating any new demand for homes…which means that the demand at this price point in the absence of exotic financing schemes is close to nil!! We are seeing this lack of demand in the market right now which will help adjust prices to demand driven by fundamentals.

[barnaby33]

Everybody keeps throwing around wage increase percentages. kev374, you say wage growth has been stagnate the last five years. A previous poster said wage growth has been 36% over that period, which is it?

Furthermore, wages are not the only fundamental that drive prices, interest rates do as well. There could be others but I am too hung over to think about what they might be.

Josh

[no_such_reality]

I like HSBC’s economic analysis of the bubble and one of the ratios they use: presnet value of the future debt load to income.

Essentially, it accounts for the low interest rate effect in flattening the curve, however it’s still outrageous.

[(former)FormerSanDiegan]

qcomer – The concept to which you are referring to in statistics is called “stationarity”. In general terms this means that the statistical measures apply over a time frame (or set of samples) for which the system remained stable, with no fundamental underlying changes. If there are external forces which change the system, the previous statistics do not apply.

PS suggests that she wants to project future prices by only considering data prior to what was a fundamental underlying change to the system in 1997 (i.e. change in tax treratment of the sale of personal residences).

wiki “stationary process” for more insight.

[unlawflcombatnt]

The mean it will revert to will be the mean relationship between home prices and income. San Diego is far above that relationship in price-to-income. So is most of the rest of the country. A reversion to the mean price-to-income ratio means a price drop of over 50%.

unlawflcombatnt

EconomicPopulistCommentary

Economic Patriot Forum

[Author]

Posts