Thursday, June 1, 2006 | When real estate analysts want to quantify price movements in the housing market, the metric they invariably use is the median home price. It’s certainly true that the median conveys some useful information about the housing market, but its accuracy as a gauge of pricing power is often overstated.
What is the median? Given an ordered set of numbers, the median is the number that appears in the very middle of the list (or, in the case of a list with an even number of elements, the average of the two numbers at the midpoint).
Let’s take a hypothetical example of five homes sold in a given neighborhood. The homes were sold at $400,000, $500,000, $600,000, $700,000, and $2,000,000. (The residents of this neighborhood like round numbers).
The median home price in our neighborhood would be $600,000 – the number that appears in the middle of the list, and a fairly good representation of home prices in the area.
You can immediately see the advantage of using the median price instead of the average price, as the latter can be thrown off by just a few deep outliers. In our pretend neighborhood, the super-sized house with the gilded fountains and topiary maze has brought the average price up to $840,000 – higher than four out of the five homes sold in the neighborhood.
The median price, then, is an excellent measure of the typical amount being paid to buy a home.
The misinterpretation enters the picture when people equate the typical amount being paid to buy homes with the market value of a specific home. Thus, if the median price increased by 10 percent, many people would assume that the market prices of their own homes have increased by 10 percent as well. But this is a questionable assumption, because while the median accurately reflects the price tag of typical home sale, it does not measure changes in the character of homes being purchased.
Put another way, the median price shows how much the typical person is willing to spend on a house – but it does not show what that person is getting for the money.
Consider, for instance, the enormous amount of home improvement dollars that have been spent in recent years. The ubiquity of dumpsters in residential neighborhoods has been a testament to San Diegans’ multi-year attempt to turn the city into one giant granite countertop. Home improvements will increase the resale value of homes involved, and thus increase the median sale price. But the portion of the median price increase caused by home improvements does not reflect an increase in the market price of non-renovated homes.
Another factor that muddies the water is the recent preponderance of buyer incentives. We’ve probably all heard about builders who are giving away new cars to condo purchasers. Other such deal-sweeteners include free unit upgrades or, common among resale transactions, having the seller pay closing costs and add them to the purchase price.
If someone purchases a $400,000 condo and gets a “free” $20,000 car, that buyer is in reality paying $380,000 for the condo. The median calculations, however, show the purchase price at $400,000. As with improvements, the rise in the median caused by incentives does not reflect a rise in the resale value of a given home.
There are factors that can cause the median price to understate pricing power, as well. As San Diego home affordability has gotten really bad in recent years, buyers have often compensated by purchasing units that were smaller or in a less desirable location than they would have liked. This trend has culminated in the recent popularity of low-cost (and often low-quality) condo conversions.
In this case, the increased purchases of smaller or otherwise less valuable homes exerted a downward pressure on the median price that was not reflective of a given home’s decline in pricing power.
The attached graph displays two measures of pricing movement for single-family homes in San Diego: the median price, and the Case-Shiller index value. The Case-Shiller home price indexes attempt to abstract out all the confounding factors we’ve been discussing, and to ascertain the resale value of a given, unrenovated home. They are thus very useful in determining measuring the effectiveness of the median as an indicator of home price movements.
What we can see from the graph is that from 2001 to 2003, the median price overstated resale pricing power. My guess is that this was due mostly to the fact that people were becoming very emboldened about the housing boom and were opting to stretch further in order to buy bigger houses with more improvements. Buyers raised their price range even faster than home prices were rising.
From 2003 to 2005, however, things headed in the opposite direction, and the growth in the median price started to understate growth in the market price of an individual home. At this point, I would surmise that the desire to spring for as much house as possible was being eclipsed by the crushing expense of doing so. At this point, people started to buy units that were smaller or otherwise less desirable, a trend that was especially pronounced in 2005. The buyer price range grew more slowly than the actual underlying home prices, causing the median to understate the latter. (Incidentally, movements in the total median price including both condos and detached homes were almost identical to those of the detached-only median price shown in the graph.)
The home improvement craze and the trend towards buyer incentives may have made a difference. However, both of these factors were more prevalent in the later period than in the earlier period. And, despite the fact that both home improvements and incentives would help to inflate the median price, we find that the median price was actually understating home pricing power during this period. What that tells us is that the buyers’ target price range is most important factor in determining whether the median price exaggerates or understates resale prices.
It seems that the trend is now heading in the opposite direction once again. Some analysts have noted that the pool of first-time buyers is shrinking, and that the resulting decrease in low-end home sales is causing the median price to overstate home price levels. And with the increased competition for buyers resulting from the steep rise in inventory, at least some buyers are said to be getting more bang for the buck than they could have last year. The net effect, if this anecdotal is correct, is that the target price range is still fairly high, but the actual homes being purchased with that money are flat or dropping in price.
In general, as the housing market returns to normalcy, the median price will most probably artificially overstate the actual movement of individual home prices.
The median home price certainly has its uses as a statistic. It is very easy to ascertain and can be sliced and diced by area, price level, property type, and any number of characteristics. And it is a fairly accurate assessment of what people are, in general, willing to pay for a home at any given time.
But the median only gives us a rough idea of the changes to the market price of a given home. For this purpose, it fails to tell the whole story. In the years to come, as everyone attempts to interpret what’s going on with the housing market, it will be important to keep the limitations of the median price in mind.
Rich Toscano is an independent real estate analyst residing in Hillcrest and working in La Jolla. He writes extensively about San Diego housing at http://piggington.com “target=”_blank”>Piggington’s Econo-Almanac.Send a letter to the editor.