The difference between Median Price and Average Price

The difference between Median Price and Average Price

Median Price and Average Price are commonly reported metrics in the real estate market, but they are very different. Let’s step back to college statistics class to revisit them both, and learn why Median is preferred for housing prices.
Average: the sum of all the values divided by the total number of values; also known as the Mean
Median: the midpoint in a given range of values, taken as the average of the two middle values when the sequence has an even number of values
A typical market has a few low-priced homes, lots of mid-priced homes, many medium-high-priced homes, and some very-high-priced homes. The natural shape of the housing price data (known as the distribution), illustrates the two primary reasons why Average fails:
  1. Very high priced homes (outliers) overwhelmingly overpower the Average, artificially inflating the market performance even though the market hasn’t drastically changed, and
  2. Asymmetrical distribution of prices (skewness) causes the Average to overestimate the center.

Average can over-estimate the market by 15-20% or more than the Median! As data sizes gets smaller—through filter selections, etc.—the average degenerates even more poorly. 

 

Let’s look at a sample market. The blue line represents a distribution of all listings by sold price for a time period. The orange vertical line is the Median Sold Price, and the gray vertical line is the Average Sold Price. That’s a huge difference!  In this example, 68% of all sold listings are below the average price. Yikes! In short, Median automatically accounts for variability in sales performance at the high and low ends of the market, giving a true economic picture of the market.



 

So why do some market reports publish Average Price when it’s simply wrong?

  1. They don’t know it’s wrong, it’s just “always been done that way.”
  2. A vocal minority, not knowing better, insist on it.
  3. Calculating Average is simple; and many databases do not support the code to calculate Median. 
For other metrics, like Percent of Ask Price Received (PAR), or Price per Square Foot (PPSF), using Average may be fine. We typically test and validated this for each market, and revisit it occasionally as economics change over time.
    • Related Articles

    • Choosing Median vs Average for metrics

      Some stats programs give the user the choice between using Median or Average aggregations for metrics.  Domus typically does not offer that choice, and the reason is simple.  The decision to choose between Median and Average for any aggregation is ...
    • What are Seasonal metrics?

      Some metrics are discussed in terms of being "Seasonal" or not, so what does that mean?  Put simply, a seasonal metric has some type of observable and consistent pattern that repeats over several time periods. The best examples of this tend to be ...
    • How the overall median close price can be up when individual area prices are down

      Have you come across a situation where the overall median sold price is up, but the sold prices by geographic areas are down? Well, sometimes data looks weird, even when it's correct. This is one of those cases. The issue is rooted in the definition ...
    • The Difference between Year to Date and Annualized metrics

      In Market Insights, choosing the 'By Year' time period gives you the additional option of choosing 'Year to Date' or 'Annualized' as the aggregation against which to compare the current year results. Note that the current year on the chart AND the ...
    • The Lifecycle of Listing Metrics

      There are lots of metrics that can be used to evaluate the real estate market, and they're all related to specific dates in the listing lifecycle.  For example "Number of New Listings" are counted based on the Listing Date, and "Number of Closed ...