bostonbubble.com

[admin]

Possibly one of the most important fundamentals in determining what a property is worth is the income of those living in the area. Employment and business opportunities vary geographically and people would logically be willing to pay more to live in areas with better income opportunities. Income is also the source of the money used to purchase housing and as such limits the extent to which housing prices may rise over the long term. While such a limitation may be overpowered for periods of time by factors such as aggressive financing and temporarily low interest rates, in the long term prices must be sufficiently supported by income.

A simple method of determining how well housing prices are supported by incomes is to calculate the ratio of sale prices to incomes and compare the ratio from different time periods. While this does not provide an intricate model detailing exactly how incomes affect prices, it does illuminate how much income has been necessary in the past to support housing prices as well as when prices rose beyond what was sustainable for the long term. A plausible hypothesis is that there is a correct ratio at which one would be able to purchase a home with a certain number of years worth of labor and if the market were always appropriately priced that ratio would be nearly constant over time.

Below is a graph of the average sale price of single family homes and condominiums in Massachusetts divided by the median income in Massachusetts for 1984 - 2005. The blue line is the price to income ratio. The yellow line is the running average, which is the average of the ratio for all years up to that point. The red and green lines are each one running standard deviation from the running average and show the range within which the ratio could have been expected to vary in general at that point in time. The muted red lines show additional standard deviations above the average, which typically correspond with extreme values.



From the above graph, current prices as of 2005 are clearly above historical norms. Prices in 2005 were at 7.54 times incomes, which is over two standard deviations from the ending average of 5.12. Furthermore, the average for most of the recorded data was hovering a little above 4.5 and only reached 5.12 due to the recent, dramatic upswing. In the past, homes have even been purchased as cheaply as for 3.39 times income.

Notably, 1999 was the first time that the ratio of prices to incomes deviated far enough from the running average to go beyond the range set by the standard deviation. The previous boom and bust stayed entirely within one standard deviation of the running average. This may signal that the real estate bubble started to inflate in earnest in Massachusetts in 1999. Prices began heading back toward the normal range in 2000 but then shot skyward for the next few years. Probably not coincidentally, the ascent coincided with the aggressive slashing of interest rates by the Federal Reserve at first in order to clean up the mess left by the dot-com bubble and then to aid the economy in recovering from the attacks of September 11, 2001. Rates have only recently returned to historically comparable levels and the temporary boost that very low rates may have provided to housing prices could be expected to reverse as well, over time.

This also offers a useful tool for gauging how far prices may correct in Massachusetts' current downturn. There may be considerably above ordinary risk involved in the market until prices are back within one standard deviation of their relationship to income. Furthermore, the previous downturn saw prices dip well below the running average and almost reach a full standard deviation below the average, so even once prices return to the normal range, additional declines should not be unexpected. Prices in 2005 were 41% above the historical average relative to income - a corresponding reversion should not be surprising. This can be achieved through the continued decline in Massachusetts housing prices coupled with the gradual increase in nominal incomes which generally occurs.

Another very useful comparison would be to examine the ratio of housing prices to rents over time. This might be the future subject of a Boston Bubble Report - pointers to references for yearly rent data in Massachusetts would be welcome.

Pointers to references with better housing and income data would also be welcome. The income data used for this comparison unfortunately only goes back to 1984 whereas the housing price data goes all the way back to 1968. Also, the housing price data is the average selling price whereas the median would be preferable.

Double checking of the data presented here by others would be greatly appreciated in order to identify and correct any errors. The following sources were used:

• The US Census Bureau's Table H-8. Median Household Income by State: 1984 to 2005
  
• The Massachusetts Association of Realtors' Statistical Review of Massachusetts Single-Family Housing Market (PDF). Many thanks to The Massachusetts Housing Market blog for pointing out this data series.
  

The text of this post and the associated graphs are Copyright 2006 by bostonbubble.com with all rights reserved, except as stated here. You may reproduce the graph or the text of the entire post as a whole (including the graph) under the Creative Commons Attribution-NoDerivs 2.5 License. You may additionally scale the graph to fit your work. Alternatively, if you remove the bostonbubble.com signature from the bottom left hand corner of the image within this post, that modified image (and only that modified image) can then be distributed under the Create Commons Attribution 2.5 License. In all cases, attribution should be made via a hyperlink to http://www.bostonbubble.com/forums/viewtopic.php?t=141 or http://www.bostonbubble.com/ Quoting excerpts of the text is also allowed provided that the quotes would normally fall under fair use. To request other terms for reproduction, please post your request in the original thread at http://www.bostonbubble.com/forums/viewtopic.php?t=141

- admin

[F5F]

This is an interesting analysis, but does not account for the more relevant ratio that homebuyers must consider: expense ratio (the ratio between the buyers' monthly income and the monthly mortage payment). Plotting expense ratio factors the affect of interest rate into the plot. I am sure that expense ratios are also out-of-whack, though the fact that interest rates are far lower now than ever before will have a big impact on the conclusion as far a just how big this bubble REALLY is (but I do agree this IS a bubble).

[john p]

I would love to see another line which charted the trend of the typical monthly payment amount. For instance if the average price at a specific date had a corresponding interest rate, one could run the typical 30 year mortgage monthly payment trend. I think it will dilute the effect to some degree almost like how property taxes fluxuate based on the ups and downs of property assessment value.

Lastly, the typical trend is for people to start off with condos or starter homes and trade up. I wonder how much money people pay in the long term in doing so. They have to pay transaction costs multiple times, sometimes carry two mortgages, and they usually start off again with a new 30 year term so most of what they are paying out year to year in the early years is interest and they post pone traction in biting into principal.

I'd like to see the trend line I mentioned earlier overlayed with the mortgage payment lifecycle of the typical strategies: one for renting for the "single" chapter of ones life, then condo, starter home, and final home (assuming realistic investment savings); second for the guy who buys a condo right out of school and goes to the final house, etc. I guess I'd like to test drive the different strategies to see if "stretching" to make a mortgage payment at a younger age really plays out. Is now a good time to flip houses and does trading up on a more frequent basis pay out better? My guess is that it may be in certain times and not so in others. It would be cool to see what strategy would make someone starting out more wealthy in the long run today... Although this exercise would have way too many variables, attempting to do them would I think, make one much more savvy.

Again, great work Admin.

[admin]

[john p]

My goal is to create a virtural physical environment that has the mathematical behavioral properties of all of the elements in the Periodic Table. Then, I study the behavior when I add or subtract certain elements. I then plan to understand how to promote or inhibit behavior, cure global diseases and retrace the common growth fractal back to the big bang, give my Nobel Peace Prize speech and then retire in Vegas. Let me know when you're finished and you can help me reach my dream. I also want to have a sandwich and mixed drink named after me.

[gadgetmaniac]

Here is a quick go at figuring out the historic percentage of income which is spent on mortgage payments. It is based on the assumption of a zero down mortgage. It takes into account interest rates, median family income and median house prices for Massachusetts.

http://spreadsheets.google.com/ccc?key=pFv2oZYywvl5gURQh8_qxHw

I didnt know how to present the data as a graph in this post, but plotting the year vs. percent of income reveals interesting trends. Before the last housing downturn (1991) people were spending 50% of their income on housing, with the percentage dropping down to 35% after the crash. Since that time, the percentage has been steadily increasing back up to 50% level.

The links in the discussion above only present median income as far back as 1984. To arrive at numbers for 1980-1983, I used data from the median income for 4-member families from the link below. For the years 1984-1987 the 4-person income was a multiple of 1.37 on the median family income so I just divided the 4-person income by 1.37 to arrive at estimates for median family income from 1980-1983.
http://www.census.gov/hhes/www/income/4person.html

Mortgage rates for each year were averaged from the following site:
http://www.census.gov/hhes/www/income/4person.html

[gadgetmaniac]

Historic mortgage rates are from the site below.

http://www.erate.com/mortgage_rates_history.htm

[admin]

gadgetmaniac,

Thanks for putting that together. I would love to see your graph, and I'm sure others would too. To post a graph, you can follow these steps:

• Save the graph as an image (gif, jpeg, or png).
  
• Put the image on a web server.
  
• Link to the image using the "img" BBCode tag.
  

If you don't have your own web server, you may want to give the ImageShack image hosting service a try at http://imageshack.us/ I have never used this myself or seen it in action (at least that I am aware of), but it looks like it could be worth checking out. It is free and it looks like it's pretty easy to use. Right after you upload an image, they even give you the BBCode that you can cut and paste to display the image in a forum (including this one). That should make the third step above a bit easier if you aren't already familiar with BBCode.

- admin

[gadgetmaniac]

[john p]

What is the "percentage on housing?"

Is it gross family income? I'm sure there were more 1 breadwinner households back in the 80's than today...

Does the cost of housing include: mortgage, pmi (if any), home insurance, property taxes? (this was my assumption).

I read another article recently that lumped in utilities and landscaping/plowing etc. in the "cost of housing".

So for all of these "rules of thumb" are there standard definitions that go along with them? Does OFHEO define them?

Does anyone know what the source is that the FED uses to gauge the real estate market?

Thank you gadgetmaniac for your chart, the results were a bit different than I would have assumed and I'm still trying to reconfigure my understanding based on this data.

[gadgetmaniac]

This a very simple analysis which does not include property taxes, landscaping costs, etc. If you follow the link to the google spreadsheet, I have provided all the numbers I used.

All I did was figure out a monthly mortgage payment (with the PMT function in Excel) using the average mortgage interest rate and average home purchase price for each year assuming no down payment. The monthly mortgage payment was then divided by the median monthly household income to give a value of "monthly payment as a percentage of income"

For me, the interesting point from this simple analysis is that the last housing downturn looks similar to the current one. Before both downturns, housing prices appreciated much more than incomes so that people were spending more and more of their income on housing. Both downturns started when the mortgage payment for an average priced house reached about 50% of median household income.

During the last downturn, prices dropped from 1990 until 1994 until this percentage dropped to 35%. From 1994-2005 this percentage rose steadily back up to 50%, the same level as before the 1990 downturn.

[john p]

This is really great to see something that manifolds so many major influences. I think what's interesting is how this may overlay against the Dow Jones Average. Further, when the market is rising, a family that is trading up is drawing capital gains from their starter home/ condo so they will have a greater amount to put down on their new home. If they buy a starter home for $250k and sell it 8 years later for $450k they will walk away with around $175k at closing. So if they can swing another thousand or so a month in mortgage, and they've most likely refinanced and brought down the rate, they can get a house in the mid to high $500's (maybe more). Your assumptions are based on no money down (which is the the most controlled approach).

I would like to see a trend of the percentage of home as a down payment overlayed to see this tradeup/ capital gains net effect.

Awesome idea and work gadgetmaniac; very though provoking.

[admin]

gadgetmaniac,

I've added a running average and running standard deviations to your data:



That's quite a plunge from 1990 to 1993. The percentage went from more than one standard deviation above the running average to practically two standard deviations below, and it only took three years.

Also of note, 2004 had the highest percentage on record, so the recent run-up was historically predominant even using this metric favored by the NAR.

- admin

[JCK]

[admin]