Variation in price is a fact of real estate. And no one knows it better than the appraiser. Each week, as we sift the market data that make up the meat of our appraisals, we encounter variation: the house sold by an estate and sold again, unchanged, within months for another $100,000; the defunct nursing home, bought by an investor and transferred again immediately to a college for an extra million; the church land sold to a builder and "flipped" on the same day for twice the price. The question is not whether these things occur, because they do. In some markets, they are common. The question for appraisers is, what do we do with the knowledge?

Appraisers make use of "paired sales" - sales and resales of the same property or of similar properties over a time span - to measure the rate of price change. On the theory that real estate markets are near-perfect and that changes in the price of one property over time provide an accurate reflection of change in the market, appraisal texts advise that appraisers read change from paired sales. The advice ignores variation. The gas station that is sold by a major oil company in 1997 to an investor for $400,000 and that is sold by the investor to a station operator in 1998 for $600,000 likely did not appreciate by 50% in a year. Rather, the investor found an opportunity for a below-market purchase and bought for that reason.

The real estate market that most closely approaches the "perfect market" model is houses. In a neighborhood of standard three-bedroom Colonials similar in size and maintenance, if one sells for $300,000, we can likely predict a selling price in a range from $290,000 to $310,000 for any similar house. The language of variation is standard deviations and confidence intervals. One standard deviation is the range that encompasses roughly two thirds of the likely outcomes for a house sale. In the example, we might say that the 68% confidence interval for the house is within about $10,000, or 3%, of $300,000. For other market sectors, a general range for the 68% confidence interval is as follows:

- 3% to 4% for leased retail property
- 4% to 6% for leased offices (which experience more changes in rent rate)
- 15% for single-occupant industrials
- 20% or more for development land and special purpose property

For those whose interest in committing funds to real estate is conservative, variation introduces an unwelcome element of uncertainty. For investors, variation creates opportunity. The investor's goal is to buy low. Among the sources of underpriced real estate are these:

- assets disposed by institutions whose business is not real estate
- assets disposed by government agencies
- properties labeled "contaminated"
- poorly marketed properties

On the other end, investors do well to seek out buyers with special needs and buyers in overheated markets. In 2000 and early 2001, developers of luxury housing and of internet data centers were "overpayers."

Awareness of variation is second nature to an appraiser. To our clients - the courts, lenders, tax assessors - however, certainty concerning value is a goal, and talk of variation is unsettling. Variation is fact. The question for appraisers is whether to make that fact plain or to tell the audience what it wants to hear. The question is no question at all. Fact is fact. Our job is to make it plain.

Eric T. Reenstierna, MAI

The Reenstierna Associates Report is published as a service to the clients of Eric Reenstierna Associates and other real estate professionals. The views expressed are those of the articles' authors and do not necessarily reflect those of other members of the organization. Copyright 2002. All rights reserved.

Eric Reenstierna Associates
24 Thorndike Street
Cambridge, Massachusetts 02141
(617) 577-0096