Selecting Comparable Sales
A list price generally begins with a set of closed sales that an agent selected. The selection determines the range the analysis produces, shapes the seller’s understanding of that range, and is a major influence on whether the price is supported when the property is appraised on a later contract. For that reason, the method by which comparable sales are selected is worth understanding thoroughly, and it is ordinarily the first thing examined in a CMA.
What makes a sale comparable. A comparable sale is a closed sale of a property that a buyer of the subject would have considered as an alternative to the subject. It is not simply a recent sale that resembles the subject. The test is competition for the same buyer. Two properties are comparable when the people shopping for one would have seriously considered the other. The competitive market area, established before the search, supplies the candidates; everything that follows is the process of deciding which of those candidates a buyer of the subject would realistically have considered.
The selection criteria. Each criterion below is the same question, competition for the same buyer, applied to one characteristic of the sale.
Location within the competitive market area. A sale outside the market area is ordinarily set aside, whatever its distance. A sale on the far side of a parkway that separates two market areas fails this criterion at four-tenths of a mile as clearly as it would at four miles.
Same property type, and a similar style. Property type describes how the property is owned and, for two- to four-family houses, how many dwelling units it holds: a condominium and a single-family house, or a two-family and a single-family, are different property types, and one is ordinarily not a comparable for the other. Style can be part of what draws a buyer to a property; a buyer looking for a ranch is generally not considering a colonial. The comparable sales are similar in style when possible. Where there are too few such sales, the difference is addressed in the analysis.
Gross living area close to the subject’s. This is a guide rather than a fixed limit. As a difference in size grows, the size adjustment grows with it, and the sale provides less reliable evidence than it appears to; a difference beyond roughly 15 to 20 percent is a warning sign that the sale may not have been competing for the subject’s buyer. A 1,500-square-foot subject and a 2,100-square-foot sale, 40 percent apart, rarely compete for the same buyer.
Date of sale: the most recent similar sales, within the past twelve months. Fannie Mae’s Selling Guide (B4-1.3-08) looks to comparable sales closed within the last twelve months and recognizes that the best and most appropriate sale may not always be the most recent. The most recent sale generally reflects the current market most closely, and an older sale that supports a higher price does not ordinarily outweigh more recent sales of similar properties. An older sale is usable when it is the best indicator available, for example the only true substitute in the subject’s market area, with the reason written into the analysis and the market’s movement since its contract date applied.
Similar condition and similar site. An updated house and a dated house on the same street sell at different prices; they are in the same market area, and the analysis adjusts for the condition difference, which is smaller the closer the condition. The same is true of a half-acre lot and a quarter-acre lot in a market where the sales show a price difference for lot size.
A sale between unrelated parties, on the open market. The parties were unrelated, neither was under pressure, and the property was exposed to the open market. An estate sale and a bank-owned sale are genuine closings that need a closer look: the motivation behind the sale and how the property was marketed are verified, and the sale is set aside, with the reason noted, only if those affected the price. A transfer between family members that was never offered on the open market is set aside with the reason noted, or, when no better evidence exists, retained with the reason stated.
A sale that fails two or more of these criteria is a warning sign rather than a sale to rely on: each miss adds a difference the analysis has to measure, and together they usually mean the sale was not competing for the subject’s buyer. Such a sale is ordinarily set aside unless no better evidence exists, with the reason stated. A difference in condition or site within the market area is generally an adjustable difference rather than a failed criterion.
Two further questions. The criteria produce candidates. Two additional questions determine whether each candidate belongs in the analysis. First, is there a closer and more similar sale that has been left off the list? If so, the analysis has a gap, and a sale that is closer and more similar than those selected needs a stated reason for its absence. A lower price is not such a reason: a sale that is otherwise highly comparable is not left off because of the price it supports. Second, is the sale on the list because it is the best available comparison, or because of the price it supports? An analysis assembled to determine a price and an analysis assembled to support one read differently. A seller may not detect the difference. The sales themselves show it.
Sales that miss one criterion. Many candidate sets contain sales that satisfy every criterion and sales that miss one. A single miss on size, date, style, condition or site does not ordinarily remove a sale (a sale outside the market area or of a different property type is ordinarily set aside, as above); it identifies the difference the analysis has to account for, either with an adjustment the market supports or by pairing the sale with another that differs in the opposite direction. A sale fourteen months old that is the best indicator in a market with little sales activity, or 300 square feet larger in a subdivision where house sizes vary, remains useful evidence with that difference stated. Among the sales that qualify, the most recent and most proximate similar sales are relied upon most, because they carry the fewest differences from the subject and from one another.
How the criteria apply by property type. For a condominium unit, the first candidates are resales in the same project, because the association, the fees, the amenities and the condition of the common elements are shared; a sale from a competing project is used when the subject’s project has too few, with the differences in fees and amenities identified. For a two- to four-unit building, the unit count is part of property type, and a three-family is ordinarily compared with three-families first; the unit mix and the rents follow. For new construction, the builder’s own closed sales and resales of similar houses nearby are both candidates, and each shows something the other cannot: whether the builder’s prices are holding, and what the wider market pays for a house of that kind once it is no longer new. For property on acreage, the market area within the town is often a distance rather than a neighborhood, and acreage, frontage and outbuildings join gross living area as the characteristics the sales are compared on. For a property whose buyers would consider only its own kind, such as an antique house built in 1750, the first candidates are other antique houses in the town, searched back in time and across the town; where the town has none that serve as good substitutes, sales of antique houses in a nearby town may be better evidence than a newer colonial down the street, with the town of each sale noted and the differences between the towns, including whether a sale lies in a local historic district, taken into account.
How it applies. A CMA for a 1,650-square-foot colonial contains three sales: a 1,620-square-foot colonial two streets away that closed in April; a 2,050-square-foot colonial on the opposite side of the highway that closed the previous month; and a 1,700-square-foot ranch in the same subdivision that closed in February. The first sale satisfies every criterion. The second is about 24 percent larger, beyond the size guide, and in all likelihood outside the market area; it should not be relied on, and if it stays, the analysis needs a reason other than its price. The third satisfies location, size and date and differs in style; whether it remains depends on whether there are enough colonial sales and on whether the subdivision’s sales history shows ranches and colonials selling in the same range, which is a matter of evidence rather than assumption, and if it is used, the difference is addressed in the analysis. A fourth candidate, a 1,580-square-foot colonial in the same subdivision that closed eight months ago in a market that has held level, is older than the other sales; it is retained with the reason stated, because it is closer in location, style and size than any sale from outside the subdivision, and the note records that the market has not moved since its contract date. A check of that kind takes a few minutes and takes place before the seller sees the analysis. The same check supplies the answer when a seller asks why the house across the highway is not on the list. Understanding how the sales were chosen is what makes it possible to answer directly.
Downloadable: The Comp Selection Checklist (two-sided; Section A search area, Section B the selection criteria as yes/no tests, Section C the two further questions), reissued for this course with a notes column for the reason each sale is kept or set aside.