Supporting a market conditions adjustment

Why this line is different

Every other adjustment on the grid compares two properties. This one compares two dates. It is the only line that touches all of your comparables at once, and the only one where a mistake pushes every indication the same way instead of scattering them.

It's also the line most often carried over from the last report. A rate derived in the spring gets reused in the fall because the market feels about the same. It might be. That still isn't support.

Decide which sales you're measuring

A trend belongs to a segment, not to a county. The right set is the one your subject competes in: same kind of property, same price range, same part of town. Cast the net too wide and you average the trend away. Draw it too tight and you're reading noise.

A quick test: would you accept a sale from this set as a comparable on some other assignment? If not, it doesn't belong in the trend either.

• Property type and style. Detached with detached. Don't mix attached product in.
• Price range. A market can move at one speed at the low end and another at the top, and often does.
• Geography. The smallest area that still gives you enough sales. Usually wider than the subject's subdivision and narrower than the metro area.
• Time window. Long enough to show a direction, short enough that the direction hasn't reversed inside it. Twelve to twenty-four months is common. Let the data decide.

Write down what you put in and what you left out. That sentence carries much of the support, and it is the first thing a reviewer looks for.

Use more than one indicator

No single statistic is the market. Each of the usual ones gets thrown off by something, and they don't all get thrown off the same way. That's why you run several.

• Median sale price. Stands up to outliers, but it moves when the mix of what sold changes, even if no single property changed in value.
• Median price per square foot. Handles some of that mix problem, though it drifts if the typical house in your set is getting bigger or smaller.
• Sale price to list price. Turns sooner than closed prices, because it picks up negotiation before settlement.
• Days on market and months of supply. These move before price does, which makes them good corroboration rather than the measurement itself.
• Regression on sale date. Gives you a rate directly and can hold other variables still. It needs a real number of sales to mean anything, and on a thin set it will hand you a confident slope fitted to noise.

When several indicators point the same way and land near the same size, you have a finding. When they contradict each other, that tells you the trend isn't established in this data. The right answer may be a smaller adjustment, or none.

Turn the trend into a rate

"Roughly four percent over the last year" still has to become something you can multiply a sale price by. Two conventions are in ordinary use.

• A monthly percentage, applied to each comparable for the months between its contract date and your effective date. Simple, and the easiest to explain.
• An index. Each period sits relative to a base period, and each comparable is adjusted by the ratio between its period and yours. Truer when the rate of change is itself changing, and worth the extra step in a market that sped up or turned inside your window.

Either way, don't claim more precision than you have. If you measured the market to within a point, you can't adjust to two decimal places. A rate written as 0.37% per month claims a confidence the data almost never supports.

Apply it from the contract date

The price is set when the parties agree, not when the deal closes. A sale that closed in June at a price agreed in March reflects the March market. Adjust it from June and you understate the movement by however long escrow ran. That's routinely thirty to sixty days, and longer on new construction.

There's more on this in contract date, not closing date. It's the most common reason two appraisers measure the same trend and end up with different adjustments.

Test the result

A correct time adjustment should bring your comparables closer together than they were. Compare the range of the adjusted indications to the range before adjustment. If the spread got wider, the rate is probably wrong, or your set is really two markets, or another line on the grid is doing the same work.

The test costs nothing. It catches sign errors, decimal errors and segment errors while the report is still on your desk.

What the support has to show

A reviewer looking at this line is asking four questions. Answer them in order and the conversation usually ends there.

1. Which sales. The data set, with your filters stated.
2. Which method. What you measured, and what backed it up.
3. What range. The spread the data supports, not only the point you chose from it.
4. How applied. Per comparable, from which date, at what rate.

None of this requires the market to have moved. A documented conclusion of no measurable change is a supported adjustment of zero, and it beats a small number picked because an empty cell looks unfinished. See sometimes the right adjustment is no adjustment.

Common questions