What evidence actually works in a reconsideration of value?

The appraiser can only move on what they can check

A reconsideration of value (ROV) is a request, sent through the lender, asking the appraiser to look again. On Fannie Mae and Freddie Mac loans the borrower gets one ROV per appraisal, it has to come before closing, and it can include up to five alternative comparable sales with their data sources.

Here is the part that decides whether it works. Fannie Mae says the appraiser must analyze the relevance of what was sent and deliver a revised report that describes the points in dispute and the outcome, whether the value changes or not. Freddie Mac says the revised report has to include specific commentary explaining the conclusion.

Read that again from the appraiser's side. Whatever you send, I have to write down why it did or did not matter. Evidence I can verify gives me something to write. A feeling gives me one sentence: "The information provided does not change the opinion of value." That sentence takes eight seconds to type, and nobody enjoys reading it.

Three kinds of evidence that move a value

1. A factual error, with proof. The report says two baths and the permit, plans and photos show three. The report lists 1,640 square feet and the floor plan shows 1,820. These are strong because they are yes-or-no. Attach the document and name the page of the report it corrects.
2. A better closed sale the report missed, with the reason. Any higher sale won't do. It has to be more like the house than one the appraiser used: same model, same street or pocket, similar condition, closer to the effective date. One sentence on why it fits beats a paragraph on how nice it is.
3. An adjustment that lowered the value without support. Fannie Mae says adjustments must reflect the market's reaction to the difference between the properties, and that time adjustments, or the lack of them, must be supported by evidence. If a comp was adjusted down for a feature the market does not pay for, or a rising market got no time adjustment, show the paired sales (two similar homes that differ mainly in that one feature) that prove what the market actually did.

Notice the pattern. Each one points at a specific line in the report and brings proof from outside the report. That is the whole recipe.

What gets read, then set aside

I read everything that comes in on an ROV. Some of it just can't carry weight:

• The contract price. The appraisal is the test of that price. Citing it as proof is like a student grading their own exam.
• A Zestimate or any automated estimate. It is a computer's guess made without seeing the house. I am being asked to replace my evidence with less evidence.
• The agent's opinion. Agents know their markets, and I respect that. But "I've sold 40 homes here and this is worth more" is a résumé, not a sale.
• "The buyer loves it." Wonderful. Love is not on the grid.
• A list of sales with no reasons. Twelve addresses pasted from a search, including the one with the pool and the view, tells me someone ran a search. It doesn't tell me which sale is better or why.
• Active listings as proof of value. An asking price is a hope with a lockbox. Listings can show a trend. They can't show a value.

None of this is personal. If I changed values for the strongest feelings, the seller's mother would win every time.

Write it like a desk review: findings first

When I do a desk review, which is a second opinion on someone else's appraisal checked against the evidence, I use one structure, and it works just as well for an ROV.

Findings. This list holds only the items that lowered the value without support. Each one gets its own number. Each one names the line in the report, says what is wrong, and backs it with named sales: the address, sale date, price and why it applies. If a statistic helps, it goes here too.

Other errors. Everything else goes in a separate short list of bullets: the misspelled street, the wrong school district, the photo labeled as the wrong room. Those may be worth fixing, but they do not change the number, so they do not get to crowd the findings.

Here is why the split matters. The lender's reviewer has to confirm the request has enough support before it ever reaches the appraiser. A reviewer who sees three numbered findings, each with proof, knows exactly what to send. A reviewer who sees nineteen mixed complaints sees a wall.

A hypothetical finding might read: "Finding 2. Comparable 1 was adjusted down for its pool. Two pairs of similar sales in the same neighborhood, one with a pool and one without, sold within a small range of each other. The adjustment is not supported by the local market." Specific line, outside proof, no adjectives.

A checklist before the loan officer sends it

Loan officers are the ones who hold the pen at the end, so here is what I would check before anything goes to the lender's review desk:

1. Every item points to a line in the report. Page number and field. No general complaints.
2. Every comp is closed and sourced. MLS number or public record, sale date and price. The rules require a data source.
3. Every comp has a reason. One or two sentences on why it is more similar than one already used.
4. Five or fewer. That is the cap, and three strong sales beat five average ones.
5. Findings separated from nitpicks. Value-moving items numbered first, small corrections listed after.
6. No target value. Fannie Mae says its ROV policy does not change the Appraiser Independence Requirements. "We need it at $850,000" turns a fair request into a problem.
7. Sent early. The lender has to finish its own review and screen the request before the appraiser sees it, and the window closes at closing.

And plan for both outcomes. A well-built ROV can still confirm the value. That is not a loss. It means the number held up when someone looked hard at it, which is what everyone in the deal should want to know anyway.