What is the difference between median cap rate and average cap rate?

The average cap rate is the sum of all cap rates divided by the number of comps. The median cap rate is the middle value when the rates are sorted. The median is resistant to outliers; the average is not.

When should I use the median cap rate instead of the average?

Use the median whenever the comparable-property set contains a sale that is clearly unusual, such as a distressed sale, a property with a below-market lease, or a sale with non-standard financing. The median gives you the rate that the bulk of normal transactions support.

How many comparable sales do I need for a reliable median cap rate?

A minimum of five comps is necessary for either the mean or the median to be meaningful. Ten comps is better. With fewer than five, neither measure is reliable, and you should expand your search or split the property type.

What happens if the comp set has an even number and the two middle values are far apart?

The median becomes the average of those two values, which may not represent any actual sale. In that case, examine the distribution, look for a second outlier, and consider adding more comps or splitting the set.

Does the median ignore informative outliers?

Yes. If an outlier signals a genuine market shift, such as a new rent trend or a change in investor demand, the median may understate the direction of change. Cross-check the median against the most recent two or three sales to catch a trend.

Can I use the average cap rate if I remove the outliers first?

You can, but removing outliers requires a defensible rule, such as the interquartile range method, not a subjective decision. The median avoids that judgment call entirely.

What is the primary authority for using the median in appraisal?

The Appraisal Institute. The Appraisal of Real Estate and The Dictionary of Real Estate Appraisal both define the median as the more robust measure for central tendency in a comp set with outliers.

Median Cap Rate or Average Cap Rate for Market Comps?

The common assumption is that averaging all comparable-sale cap rates gives you the market rate. That is wrong when the set contains an outlier. A single distressed sale or a property with a below-market lease can pull the average away from where most deals trade. The safer center measure for a distorted comparable-property set is the median cap rate or average cap rate for market comps? The median, because it ignores the size of the extreme values. You use the mean only after you have confirmed no outliers exist. For a set of five comps at 5.0%, 5.5%, 6.0%, 6.5% and 12.0% (the last one a distressed sale), the mean is 7.0% while the median is 6.0%. The 6.0% better represents what a buyer and seller negotiating on normal terms would agree to.

A Worked Example

Work through a concrete set. You have seven closed sales of similar multi-tenant office buildings in a suburban submarket. The cap rates, ranked from low to high, are: 4.8%, 5.1%, 5.3%, 5.4%, 5.6%, 5.8%, and 9.2%.

The 9.2% sale was a 40-year-old building with a single tenant who had just filed for bankruptcy; the lease was at 60% of market rent. That is an unusual sale that distorts the comparable-property set. The mean cap rate of those seven numbers is (4.8+5.1+5.3+5.4+5.6+5.8+9.2)/7 = 36.2/7 = 5.17%. The median cap rate is the fourth value, 5.4%.

If you used the mean 5.17% to underwrite a normal building in that market, you would overvalue the property by roughly 4.5% relative to the median-based benchmark. That error is enough to lose a bid or overpay. Now run the same set through a median calculator to confirm the center: seven values, so the middle is the fourth. The median holds at 5.4% regardless of how high the outlier goes.

Calculation Steps

Calculating the median follows a fixed protocol. Sort the cap rates in ascending order. For an odd number of comps, the median is the (n+1)/2th value. For an even number, the median is the average of the two middle values.

The limitation appears when you have an even number of comps and the two middle values are far apart. A set of six comps at 4.5%, 5.0%, 5.0%, 10.0%, 10.5%, and 11.0% has a median of (5.0+10.0)/2 = 7.5%. That 7.5% represents no actual sale, and the distribution is clearly bimodal. The median has told you that the center is uninformative, but it has not told you why. You must then examine the distribution and consider adding more comps or splitting the set into two submarkets.

Why the Method Works

The median cap rate is the more robust measure when a few unusual sales distort a comparable-property set. The Appraisal Institute's The Appraisal of Real Estate, 15th edition, Chapter 11 on Statistical Analysis, directs appraisers to the median precisely for this reason. The mean is sensitive to every extreme value; the median is resistant.

But the median has a failure case. If the comp set is smaller than five sales, neither the mean nor the median is reliable. With four comps, any single outlier shifts the median by at least one position. With three comps, the median is just the middle value of three, which may itself be the outlier. The Appraisal Institute recommends a minimum of five to ten comps for a meaningful central tendency. Below five, you are not measuring a market; you are guessing.

Who This Subject Suits And Who Should Skip

This method suits small landlords and first-time rental investors who need to set a purchase offer without being misled by one odd sale. It also suits commercial real estate analysts and brokers who must defend their cap rate benchmark in a negotiation: the median is easier to explain and harder to attack than a mean that an opposing broker can show was inflated by a distressed comp.

Skip the median if you are a day-trader or stock investor looking for a quick yield metric on REITs; you want FFO or AFFO, not a property-level cap rate. Skip it if you are a homeowner trying to value your primary residence; cap rates apply to income-producing property, not owner-occupied housing. Skip it if you are seeking a single correct cap rate for a property without a market survey; the number is only meaningful relative to a specific time and place.

The single thing that most often goes wrong: using the mean without checking whether the comp set contains a single sale that is not comparable, and then underwriting a property at a rate that no normal deal would support.