What is ARV in real estate?

ARV is after repair value — what a property will be worth once the work is done.

It is the anchor of every flip. Purchase price, loan size, repair budget and profit are all derived from it, which is why an ARV that is out by ten percent puts everything downstream out by more than ten percent. It is also the number most often estimated backwards from the deal someone already wants to do.

What ARV is, and is not#

ARV is a forecast of a future sale price. Specifically: the price the property would achieve if it were on the market today, finished, competing with other finished properties.

It is not:

• The current value. That is what it is worth in its present condition, and it is usually much lower.
• The assessed value. Assessors value on a cycle for tax purposes, often with a lag and a methodology that has nothing to do with market price.
• An automated estimate. Those are trained on normal transactions and are weakest on distressed and recently renovated property — the exact population that matters here.
• What you need it to be for the deal to work. That is the failure mode the rest of this article is about.

How to calculate it#

Define the finished product first. ARV depends entirely on what the property will be when you are done. A three-bed with one bath finishing as a three-bed with two baths is a different product from the one that exists now, and it must be compared against other three-bed two-baths.

Find comparable sales of finished properties. The tightest criteria you can afford:

Use three comps minimum, and check they agree. If three comps give you $185, $210 and $260 per square foot, the area is not homogeneous enough for this method and your ARV is far less certain than a single number suggests.

Work in price per square foot, then adjust. Take the per-foot figure the comps agree on, apply it to the subject's finished square footage, then adjust for real differences — a garage, an extra bathroom, a finished basement, a corner lot on a busy road.

Then sanity-check against the top of the market. If your ARV is above the highest recent sale in the neighbourhood, you are betting on setting a new ceiling. Sometimes that is right. Usually it means the number is wrong.

The 70% rule and what it is actually for#

The standard screening formula:

Maximum offer = (ARV × 0.70) − repair costs

On a property with an ARV of $300,000 and $50,000 of work:

$300,000 × 0.70 = $210,000
$210,000 − $50,000 = $160,000 maximum offer

What the 30% covers:

It is a filter, not a valuation. Its purpose is to decide in thirty seconds whether a property is worth an hour of analysis. Experienced buyers in competitive markets often work at 75% or higher, and in slow markets or on long renovations 70% is not enough.

The rule assumes the ARV is right. Every criticism of the 70% rule is really a criticism of the ARV it was applied to.

Build your ARV in the calculator — it works out the price per square foot from three comps, flags it when they disagree with each other, and warns when the result lands above the local ceiling.

Where ARV estimates fail#

Unrenovated comps. The most common error by a distance. A distressed sale two doors down tells you what your property is worth now. It says nothing about the finished value, and using it produces an ARV close to the current value — which makes every deal look bad, or worse, gets adjusted upward by feel until it looks good.

Stretching the criteria. When three good comps do not support the number, the temptation is to widen the radius, extend the date range, or accept a different property type. Each stretch adds error, and the direction of the stretch is always toward the answer that was wanted.

Over-improving for the area. Quartz worktops and a wet room in a neighbourhood where every comparable sale has laminate and a bath does not raise the ceiling. It raises your cost. The finished product should be at the top of the local range, not above it.

Ignoring days on market. An ARV is a price at some speed. If comparable finished properties sit for 120 days, your holding costs are four months, not six weeks, and the 30% buffer absorbs it.

Using a stale market. Comps from nine months ago describe a different market. In a moving market a three-month window matters more than a tight radius.

Assuming the repair estimate. ARV and repair cost are two separate estimates and both carry error. A 10% ARV miss and a 20% repair overrun on the same deal compound into something much larger than either.

Why automated estimates are worst exactly here#

Automated valuation models work by finding statistical patterns in normal sales. Two things make them unreliable for this use.

Distressed property is out of distribution. The training data is dominated by arms-length sales of maintained homes. A property with a failed roof, no kitchen and a foreclosure history does not resemble that population, and the model has no reliable way to price it.

They cannot see the plan. ARV depends on what the property will become. No automated model knows you are adding a bathroom.

There is a broader point worth making, and we have tested it directly. An automated valuation model built on Minnesota sales data was trained three times on a growing dataset and lost to a one-line county-year median every time — 12.53% median error against 11.32% — with the gap widening as data grew. Five approaches failed to beat a simple grouped median.

That is not an argument against modelling in general. It is an argument that a transparent method you can check beats a sophisticated one you cannot, and it is the reason this article recommends comparable sales you can read rather than an estimate you have to trust.

Finding sale prices to build comps from#

You need what properties actually sold for. Where that lives depends on your state, and in some it is not public at all.

Disclosure states record the sale price on the deed or in a filed transfer document. In Minnesota the electronic Certificate of Real Estate Value captures price on most transfers, and extracts are available from the Department of Revenue.

Non-disclosure states do not require prices to be public. Alaska, Idaho, Kansas, Louisiana, Mississippi, Missouri, Montana, New Mexico, North Dakota, Texas, Utah and Wyoming. In those, a deed may show a nominal consideration such as "$10 and other valuable consideration", and assessor sites often show no sale price at all.

This is why valuation is harder in Texas than in Minnesota, and why national estimates are notably less accurate there — the underlying prices are not in the record. In a non-disclosure state the MLS becomes the practical source, which means an agent relationship is not optional.

ARV in distressed acquisitions specifically#

Three complications apply when the property is coming from a foreclosure, an auction or a tax sale.

You often cannot see inside before committing. At a foreclosure auction there is no interior access, so the repair estimate is a guess and the ARV sits on top of it. Widen the contingency accordingly.

Comparable sales may themselves be distressed. In a neighbourhood with a concentration of foreclosures, the recent sales are foreclosure sales, which are not the market you will be selling into. Filter them out or the ARV is depressed.

In redemption states you may not keep the property. In about half of states, including Minnesota, the former owner can reclaim a property after the auction. Across 326 tracked Minnesota redemption windows, 33.4% ended with the owner redeeming — and where the winning bid was under half the assessed value, 58.1% did.

That last figure is directly an ARV problem. A property that looks like an outstanding flip — a low bid against a high finished value — is the profile most likely to be redeemed out from under the buyer. The ARV was right; the buyer never got to realise it.

Estimating repairs, which is the other half#

ARV gets the attention and repair cost causes the losses. Both feed the same formula and an error in either lands in the same place.

Per-square-foot rules of thumb are for screening only. A figure like "$30 a foot for a light cosmetic refresh" is fine for deciding whether to walk a property. It is not a budget, because it does not know your roof is failing.

The items that blow budgets are the ones you cannot see:

On distressed property, add vacancy damage. A house empty through a Minnesota winter with the heat off has burst pipes, and the damage is not visible until the water is turned on. Frozen and split supply lines behind finished walls are among the most expensive surprises in cold-state flips.

Get quotes before committing where you can. On an auction purchase you cannot, which is the argument for the widest contingency you can afford. A 20% overrun on a $60,000 renovation is $12,000, and on a deal underwritten at a $30,000 profit that is 40% of the upside gone.

Permits and time. Work requiring permits takes longer than the work itself. Inspection scheduling, correction notices and re-inspections add weeks, and weeks are holding cost. A budget that prices materials and labour but not the calendar is incomplete.

A worked example#

A three-bed, two-bath, 1,450 square foot house, built 1962, needing a kitchen, one bathroom, flooring throughout and a roof.

Comps — renovated three-bed two-baths within half a mile, sold in the last four months:

Tight agreement — $224 to $226. That consistency is what makes the estimate usable.

ARV = 1,450 × $225 = $326,250, call it $325,000.

Repairs, priced from contractor quotes rather than a per-foot rule of thumb: $58,000.

Maximum offer = ($325,000 × 0.70) − $58,000 = $227,500 − $58,000 = $169,500.

Then check it. Is $325,000 above the highest recent sale in the area? If comp B at $340,000 is the ceiling, you are inside it. If the highest sale is $328,000 and you are projecting $325,000, you are pricing at the very top and the buffer is doing more work than it appears.

The short version#

1. Define what the property will be when finished.
2. Find at least three renovated comps that agree with each other.
3. Work in price per square foot, then adjust for real differences.
4. Sanity-check against the neighbourhood ceiling.
5. Price repairs separately and from quotes, not rules of thumb.
6. Apply the 70% rule as a filter, then do the real analysis.
7. In redemption states, price the possibility that you do not keep it.

The number is only as good as the comps. Everything else is arithmetic.