When the Comparables Aren’t There
Some properties have no close comparables among the sales of the past twelve months: the only Victorian on a street of ranches, a house on twenty acres in a subdivision town, a market area in which nothing has closed in six months. How the analysis responds to that situation separates an analysis from a guess, and it is one of the places where method is most visible in the finished analysis.
There are two ways to find more evidence: go back further in time, and expand the area. Neither always comes first. The choice depends on which one produces sales that compete with the subject, and the analysis records what was searched, what was found and why each sale was used. The aim does not change when sales are scarce: the most recent and most proximate similar sales are the ones relied upon. In a thin market the work is finding the sales that remain both similar and as recent and as close as the market allows.
Going back in time. A sale from nine or twelve months ago, in the subject’s own competitive market area, is often better evidence than a recent sale from outside the subject’s market area. An older sale is used with a reading of how the market has moved since it went under contract, and the direction and rate of that movement are stated and applied.
Example: going back in time. The subject is a 1,750-square-foot cape in a neighborhood of about sixty similar houses. Nothing comparable has closed there in six months. Two capes in the neighborhood closed nine and eleven months ago, at $389,000 and $382,000, and both were verified as typical sales. The MLS data for the subject’s segment, capes and colonials of 1,500 to 2,000 square feet in the town, shows median sale prices up about 2.5 percent since the newer sale went under contract and about 3 percent since the older one did. Applied from the contract dates, the two sales indicate roughly $399,000 ($389,000 × 1.025) and $393,000 ($382,000 × 1.03) in today’s market. Where the segment is too thin to show a trend on its own, it is widened, for example to the town’s single-family sales in the same price tier, and the sale-to-list ratio over the same months is read alongside it as support. Two sales from the subject’s own neighborhood, read forward with a stated trend, describe the subject’s market more faithfully than a recent sale from a different market area.
Expanding the area. Fannie Mae’s Selling Guide permits an appraiser to use sales from a competing market area when they are the best available, identified as such and with the differences addressed. The method this course teaches for a CMA is narrower. A buyer may consider more than one town, but the comparable evidence comes from the subject’s town, because a neighboring town has its own taxes, services and assessment, so its sales reflect that town’s market; the search leaves the town only as the exception described below. When the subject’s neighborhood has produced too little, the search widens to the neighborhoods in town whose sales show them selling as one market with the subject’s: the same house type, era, lot size and price tier, with price histories that move together. A neighborhood that meets that test is part of the subject’s competitive market area, and the analysis states the wider boundary with the sales evidence for it. A sale from a different market area is ordinarily set aside rather than adjusted, because a location adjustment does not make it a property that competes with the subject. Reaching into a different town for a sale that supports a price is not the same thing as establishing that the subject competes there.
Example: expanding the area. The subject is a 1,400-square-foot ranch in a 1950s neighborhood bounded by a state road and a brook. The neighborhood has produced one sale in eight months, a house that needed a full renovation. Across town, a second neighborhood of ranches built in the same decade, on similar lots and in the same price tier, has produced four sales in the same period, and the price history of the two neighborhoods has moved together for years. The two neighborhoods meet the test of one market, so the analysis states the wider market area with that evidence, and two of the second neighborhood’s sales, closed within the last four months, are used. The difference the analysis addresses is a site difference, not a location one: the second neighborhood is served by public sewer and the subject by a septic system, and the effect of that difference is read from sales in the town rather than assumed. A ranch in the neighboring town that closed last month at a higher price is not used. The comparables stay in the subject’s town, and that sale, made under another town’s taxes, services and assessment, shows what a different market paid.
In many thin markets both are needed: an older sale from the subject’s own neighborhood and a recent sale from elsewhere in the widened market area, each supplying part of the picture and read together. The older sale carries the subject’s immediate surroundings and the newer sale carries the current market, and where the two agree after the time reading and any measured site difference are applied, each confirms the other.
Example: acreage, where both are needed. The subject is a 2,200-square-foot farmhouse on twelve acres in a small town, with a 36-by-48-foot horse barn and fenced paddocks. The town records few single-family sales in a year, and in the past twelve months only one had more than five acres. Acreage alone does not take the search out of town. Sales in the town show houses on acreage with and without barns selling in the same range, so the search includes acreage properties with or without a barn; on that evidence, a house on land with a barn is not the kind of equestrian property for which the search leaves the town. The search goes back eighteen months, with a time reading applied to the older sales, and widens across the town to every area that holds houses on acreage, and it yields four in-town sales of houses on eight to twenty acres. Two have small sheds, one has a detached garage with a workshop, and none has a horse barn.
When the town has no sales of the subject’s kind. Staying in the subject’s town is the best choice and the default, and it is not always possible. Some properties have buyers who would consider only houses of their own kind: an antique home, a log home, an equestrian property. When the town has no sales of that kind that serve as good substitutes, sales of the same kind in a nearby town may be better evidence than an older in-town sale or a sale of a different kind of house down the street. Leaving the town is the exception, used only when staying in town cannot produce suitable comparables.
Example: an antique house. The subject is a center-chimney colonial built in 1750, with 2,100 square feet, wide-board floors, original fireplaces and exposed beams, on an acre and a half. Its buyers are looking for an antique house, and a 1990s colonial of the same size down the street is not an alternative they would weigh. The search goes back twenty-four months and across the whole town and finds no sale of an antique house; the town’s closed sales of that size are colonials built after 1960. A nearby town with a similar stock of eighteenth-century houses has two closed sales of antique colonials in the past year, and they are used. The CMA notes why each sale was chosen and which town it came from, and the differences between the towns are taken into account: the taxes, services and assessment of each town, and whether a sale lies in a local historic district. One of the two sales is in that town’s local historic district, where a house carries restrictions on exterior changes; the subject is not in such a district, and that difference is noted and considered in placing the subject between the two sales.
Active and pending listings as support. Active and pending listings are not comparable sales, and they do not become evidence of value when closed sales are scarce. They serve as support and guidance. They show what buyers are being offered and how quickly listings at a given list price are going under contract, they indicate the direction of the market when the closed sales are old, and they help place the subject within the range the closed sales produce. The range itself rests on closed sales.
Example. In the cape example above, the two closed sales read forward indicate roughly $393,000 to $399,000 before any other adjustment. Two capes in the neighborhood are now pending, listed at $399,900 and $409,000, and a third has been on the market for seventy days at $425,000. The pendings, listed at or somewhat above the top of what the closed sales indicate, went under contract, which is consistent with the rising trend in the segment data; their contract prices will be known at closing. The active listing, after seventy days without a contract at $425,000, is at a price buyers have not yet accepted. None of the three moves the range. They inform where the list price is placed within it, and they help explain to a seller why a list price well above it may not go under contract.
The cost approach as support. The cost approach estimates the value of the land and adds the cost to build the improvements new, less depreciation. It does not measure what buyers pay, and it does not replace sales in a CMA. It is acceptable as support, particularly for improvements the available sales cannot isolate, such as a barn, a large outbuilding or another improvement with no counterpart among the sales. The support holds only when the figures are correct: a current cost to build that improvement in that market, and depreciation that reflects its actual age, condition and usefulness to the next buyer. A barn that would cost $150,000 to build today but is forty years old and suited to a specialized use contributes far less than its cost to build. A contractor’s replacement figure with no depreciation, or a national cost-per-square-foot figure, overstates the contribution and pulls the analysis away from the sales. Where the sales do show how the market responds to an improvement, the sales govern. If they show no market reaction to it, the improvement is not adjusted, whatever it would cost to build; cost is a check, not a substitute for the market’s reaction.
Example: checking the barn. In the farmhouse example, a local builder of agricultural buildings puts the current cost to build a comparable barn at about $150,000. The barn is forty years old, and its roof and sills show their age; physical depreciation of about two-thirds of that cost, or $100,000, leaves about $50,000. Its layout, box stalls with low headroom and no bay a vehicle can enter, suits keeping horses and little else, and a further deduction of $15,000 for that limited usefulness leaves a depreciated cost of about $35,000. That figure is a check, not the adjustment. The sales govern: houses on acreage in the town sell within one range with and without barns, and the in-town acreage sales show no added price for horse facilities. Because the sales show no market reaction to the barn, it is not adjusted, and the $35,000 is neither added to the range nor substituted for what the sales show. The check is still useful. It shows that even a correctly depreciated cost figure is far below the $150,000 cost to build, and the record states the builder’s figure, the basis for each deduction and the sales evidence that governs. Had the town’s sales shown a higher price for acreage with a usable barn, the depreciated cost would have served as support for the size of that adjustment.
For other improvements the sales reflect, cost is likewise a check rather than the adjustment. A finished basement, an addition or a kitchen that the sales do reflect is read from the sales, and a cost figure serves only to test whether that reading is reasonable.
Condominiums and new construction. A condominium complex with few resales presents the same choice between time and area. Older sales within the complex are often better evidence than recent sales from another complex, because the complex itself, its fees, amenities and management, is part of what is being sold. Where a sale from another complex is used, the differences in fees and amenities are addressed. In a new subdivision whose builder still controls pricing, the sales inside the project are read together with sales from outside it, so that the analysis tests whether the wider market supports the prices inside the project.
A CMA with few comparables is presented as what it is: a thin market with a wider range, rather than a narrow range the evidence does not support. In the cape example, two older sales read forward support a range of roughly $390,000 to $405,000, wider than three recent sales from the neighborhood would have produced. The seller conversation addresses why: fewer sales, older sales and a trend reading applied to them. Narrowing that to a single figure would claim more than the evidence holds.
How it applies. A CMA for a hard-to-match property reads as a record of the search: the market area, the window, what was found, how far the search went back in time or out in area and why, what the actives and pendings indicate, and where any cost check or cost support came from. A reader can follow the reasoning and see that each sale competes with the subject. A CMA that leads with a sale from two towns away, when older sales from the subject’s own neighborhood were available, reads as evidence chosen for the price it supports rather than for its relevance.